Search Results

Search found 3459 results on 139 pages for 'if modified since'.

Page 79/139 | < Previous Page | 75 76 77 78 79 80 81 82 83 84 85 86  | Next Page >

  • How mod_cache working with "must-revalidate" and "max-age"?

    - by Dmitriy Sosunov
    Quick question before I will explain my flow: ?an mod_cache perform revalidate with if-none-match only if max-age is expired in case if it configured in reverse proxy mode? My goal is to reduce a number of revalidation requests to our the origin server. For instance: The first request goes to the origin server and then mod_cache save a response in to the cache according to header cache-control: max-age. And only when max-age is expired then mod_cache will revalidate with if-none-match. Currently, mod_cache revalidate each request, regardless that max-age is defined or not. My configuration of Apache 2.4.3 (Windows), on linux I see the same behavior that I will show below. ServerName proxy.lo ProxyRequests Off ProxyPreserveHost Off Header set Vary "Accept, Content-Type, Content-Encoding, Accept-Language" RequestHeader set X-Forwarded-Proto "http" # modify header for user agent's Header set Cache-Control "private, no-cache, no-store, no-transform" CacheQuickHandler off CacheDefaultExpire 300 # the origin server do not provide last-modified CacheIgnoreNoLastMod On CacheIgnoreCacheControl On # the origin server define cache-control: private, no-store only for user agents # Therefore, I would like ignore those headers on the proxy server. CacheStorePrivate On CacheStoreNoStore On CacheEnable disk / CacheRoot "C:/Apache.Cache" CacheDirLevels 5 CacheDirLength 4 CacheMinExpire 15 CacheDetailHeader on CacheHeader on KeepAlive Off ProxyPass / http://origin.lo/ ProxyPassReverse / http://origin.lo/ Also, I have turned on debug log level to see how mod_cache handles a content for caching: I provided this to show that mod_proxy always decides that a content isn't fresh. Why?I provided this to show that mod_proxy always decide that a content isn't fresh. Why? max-age was provided (see below). [Sun Nov 04 11:58:42.899890 2012] [cache:debug] [pid 6492:tid 1400] cache_storage.c(624): [client 192.168.1.100:63741] AH00698: cache: Key for entity /testpage?(null) is http://proxy.lo/testpage? [Sun Nov 04 11:58:42.899890 2012] [cache_disk:debug] [pid 6492:tid 1400] mod_cache_disk.c(569): [client 192.168.1.100:63741] AH00709: Recalled cached URL info header http://proxy.lo/testpage? [Sun Nov 04 11:58:42.899890 2012] [cache_disk:debug] [pid 6492:tid 1400] mod_cache_disk.c(865): [client 192.168.1.100:63741] AH00720: Recalled headers for URL http://proxy.lo/testpage? [Sun Nov 04 11:58:42.899890 2012] [cache:debug] [pid 6492:tid 1400] cache_storage.c(320): [client 192.168.1.100:63741] AH00695: Cached response for /testpage isn't fresh. Adding/replacing conditional request headers. [Sun Nov 04 11:58:42.899890 2012] [cache:debug] [pid 6492:tid 1400] mod_cache.c(414): [client 192.168.1.100:63741] AH00757: Adding CACHE_SAVE filter for /testpage [Sun Nov 04 11:58:42.899890 2012] [cache:debug] [pid 6492:tid 1400] mod_cache.c(448): [client 192.168.1.100:63741] AH00759: Adding CACHE_REMOVE_URL filter for /testpage [Sun Nov 04 11:58:42.899890 2012] [proxy:debug] [pid 6492:tid 1400] mod_proxy.c(1068): [client 192.168.1.100:63741] AH01143: Running scheme http handler (attempt 0) [Sun Nov 04 11:58:42.899890 2012] [proxy:debug] [pid 6492:tid 1400] proxy_util.c(1976): AH00942: HTTP: has acquired connection for (origin.lo) [Sun Nov 04 11:58:42.899890 2012] [proxy:debug] [pid 6492:tid 1400] proxy_util.c(2029): [client 192.168.1.100:63741] AH00944: connecting http://origin.lo/testpage to origin.lo:80 [Sun Nov 04 11:58:42.901890 2012] [proxy:debug] [pid 6492:tid 1400] proxy_util.c(2151): [client 192.168.1.100:63741] AH00947: connected /testpage to origin.lo:80 [Sun Nov 04 11:58:42.901890 2012] [proxy:debug] [pid 6492:tid 1400] proxy_util.c(2554): AH00962: HTTP: connection complete to 192.168.1.100:80 (origin.lo) [Sun Nov 04 11:58:42.903890 2012] [proxy:debug] [pid 6492:tid 1400] proxy_util.c(1991): AH00943: http: has released connection for (origin.lo) [Sun Nov 04 11:58:42.903890 2012] [headers:debug] [pid 6492:tid 1400] mod_headers.c(800): AH01502: headers: ap_headers_output_filter() [Sun Nov 04 11:58:42.903890 2012] [cache:debug] [pid 6492:tid 1400] mod_cache.c(1190): [client 192.168.1.100:63741] AH00769: cache: Caching url: /testpage [Sun Nov 04 11:58:42.903890 2012] [cache:debug] [pid 6492:tid 1400] mod_cache.c(1196): [client 192.168.1.100:63741] AH00770: cache: Removing CACHE_REMOVE_URL filter. [Sun Nov 04 11:58:42.904890 2012] [cache_disk:debug] [pid 6492:tid 1400] mod_cache_disk.c(1318): [client 192.168.1.100:63741] AH00737: commit_entity: Headers and body for URL http://proxy.lo/testpage? cached. The first request to the origin server without mod_proxy to http://origin.lo/ GET http://origin.lo/testpage HTTP/1.1 Host: origin.lo Connection: keep-alive User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.4 (KHTML, like Gecko) Chrome/22.0.1229.94 Safari/537.4 Accept: application/json Accept-Encoding: gzip,deflate,sdch Accept-Language: en-US,en;q=0.8 Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.3 The first response from the origin without mod_proxy HTTP/1.1 200 OK Cache-Control: must-revalidate, proxy-revalidate, max-age=30 Content-Type: application/json; charset=utf-8 ETag: "7cf651e2-176f-4ac1-808e-0e0c17cfd0a2" Server: Microsoft-IIS/7.5 X-AspNet-Version: 4.0.30319 X-Powered-By: ASP.NET Date: Sun, 04 Nov 2012 10:11:01 GMT Content-Length: 1877 So, I assumed that revalidation must be occur only in 30 seconds after the success response. Is't right? Let's check it:) Within 30 sec, the Google Chrome didn't perform any requests to the origin server to revalidate a request and has return the response from local cache. When max-age is expired, the Google Chrome perform a request to revalidate: GET http://origin.lo/testpage HTTP/1.1 Host: origin.lo Connection: keep-alive Cache-Control: max-age=0 User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.4 (KHTML, like Gecko) Chrome/22.0.1229.94 Safari/537.4 Accept: application/xml If-None-Match: "7cf651e2-176f-4ac1-808e-0e0c17cfd0a2" Accept-Encoding: gzip,deflate,sdch Accept-Language: en-US,en;q=0.8 Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.3 and response: HTTP/1.1 304 Not Modified Cache-Control: must-revalidate, proxy-revalidate, max-age=30 ETag: "7cf651e2-176f-4ac1-808e-0e0c17cfd0a2" Server: Microsoft-IIS/7.5 X-AspNet-Version: 4.0.30319 X-Powered-By: ASP.NET Date: Sun, 04 Nov 2012 10:16:20 GMT As you can see, all works as expected. User agent revalidates request only when max-age is expired. Let's now try perform the folling flow though mod_proxy (see configuration above). The first request: GET http://proxy.lo/testpage HTTP/1.1 Host: proxy.lo Connection: keep-alive User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.4 (KHTML, like Gecko) Chrome/22.0.1229.94 Safari/537.4 Accept: application/json Accept-Encoding: gzip,deflate,sdch Accept-Language: en-US,en;q=0.8 Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.3 and the response was: HTTP/1.1 200 OK Date: Sun, 04 Nov 2012 10:23:36 GMT Server: Apache Cache-Control: private, no-cache, no-store, no-transform Content-Type: application/json; charset=utf-8 ETag: "7cf651e2-176f-4ac1-808e-0e0c17cfd0a2" Content-Length: 1932 Vary: Accept,Content-Type,Content-Encoding,Accept-Language X-Cache: MISS from proxy.lo X-Cache-Detail: "cache miss: attempting entity save" from proxy.lo Connection: close Ok, let's see to the disk cache and try to see how request and response was stored. (I cut binary data) http://proxy.lo/testpage? Cache-Control: private, no-cache, no-store, no-transform Content-Type: application/json; charset=utf-8 ETag: "7cf651e2-176f-4ac1-808e-0e0c17cfd0a2" Date: Sun, 04 Nov 2012 10:27:15 GMT Content-Length: 1932 Vary: Accept, Content-Type, Content-Encoding, Accept-Language Host: proxy.lo User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.4 (KHTML, like Gecko) Chrome/22.0.1229.94 Safari/537.4 Accept: application/json Accept-Encoding: gzip,deflate,sdch Accept-Language: en-US,en;q=0.8 Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.3 X-Forwarded-Proto: http Cache-Control: max-age=300, must-revalidate X-Forwarded-For: 192.168.1.100 X-Forwarded-Host: proxy.lo X-Forwarded-Server: origin.lo Ok, what we see? We see that the first request was performed with max-age=300 & must-revalidate Ok, looks good, as for me, lets perform the next call: GET http://proxy.lo/testpage HTTP/1.1 Host: proxy.lo Connection: keep-alive User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.4 (KHTML, like Gecko) Chrome/22.0.1229.94 Safari/537.4 Accept: application/json Accept-Encoding: gzip,deflate,sdch Accept-Language: en-US,en;q=0.8 Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.3 and the second response from mod_proxy: HTTP/1.1 200 OK Date: Sun, 04 Nov 2012 10:31:58 GMT Server: Apache Cache-Control: private, no-cache, no-store, no-transform ETag: "7cf651e2-176f-4ac1-808e-0e0c17cfd0a2" Content-Length: 1932 Vary: Accept,Content-Type,Content-Encoding,Accept-Language X-Cache: REVALIDATE from proxy.lo X-Cache-Detail: "conditional cache hit: entity refreshed" from proxy.lo Connection: close Content-Type: application/json; charset=utf-8 SO, MY QUESTION IS: WHY mod_proxy perform revalidation on each request regardless that max-age is defined? N.B. Apache 2.4.3 Thanks, I would be grateful for any help.

    Read the article

  • Unity3D draw call optimization : static batching VS manually draw mesh with MaterialPropertyBlock

    - by Heisenbug
    I've read Unity3D draw call batching documentation. I understood it, and I want to use it (or something similar) in order to optimize my application. My situation is the following: I'm drawing hundreds of 3d buildings. Each building can be represented using a Mesh (or a SubMesh for each building, but I don't thing this will affect performances) Each building can be textured with several combinations of texture patterns(walls, windows,..). Textures are stored into an Atlas for optimizaztion (see Texture2d.PackTextures) Texture mapping and facade pattern generation is done in fragment shader. The shader can be the same (except for few values) for all buildings, so I'd like to use a sharedMaterial in order to optimize parameters passed to the GPU. The main problem is that, even if I use an Atlas, share the material, and declare the objects as static to use static batching, there are few parameters(very fews, it could be just even a float I guess) that should be different for every draw call. I don't know exactly how to manage this situation using Unity3D. I'm trying 2 different solutions, none of them completely implemented. Solution 1 Build a GameObject for each building building (I don't like very much the overhead of a GameObject, anyway..) Prepare each GameObject to be static batched with StaticBatchingUtility.Combine. Pack all texture into an atlas Assign the parent game object of combined batched objects the Material (basically the shader and the atlas) Change some properties in the material before drawing an Object The problem is the point 5. Let's say I have to assign a different id to an object before drawing it, how can I do this? If I use a different material for each object I can't benefit of static batching. If I use a sharedMaterial and I modify a material property, all GameObjects will reference the same modified variable Solution 2 Build a Mesh for every building (sounds better, no GameObject overhead) Pack all textures into an Atlas Draw each mesh manually using Graphics.DrawMesh Customize each DrawMesh call using a MaterialPropertyBlock This would solve the issue related to slightly modify material properties for each draw call, but the documentation isn't clear on the following point: Does several consecutive calls to Graphic.DrawMesh with a different MaterialPropertyBlock would cause a new material to be instanced? Or Unity can understand that I'm modifying just few parameters while using the same material and is able to optimize that (in such a way that the big atlas is passed just once to the GPU)?

    Read the article

  • Make Your 64 bit Computer Look like a Commodore 64

    - by Matthew Guay
    The Commodore 64 was one of the bestselling home computers ever, and many geeks got their first computing experience on one of these early personal computers. Here’s an easy way to revisit the early years of personal computing with a theme for Windows 7. With only 64Kb of ram and an 8 bit processor, the Commodore 64 is light-years behind today’s computers.  But with a Windows 7 themepack, you can turn back the years and give your computer a quick overhaul to look more like its ancient predecessor. Age Windows 7 with a click Download the Commodore 64 theme from PC World (link below), and unzip the files. Now, double-click on the Themepack file to apply the theme. This will open your Personalization panel and will automatically change your system fonts, window style, background, and more. Your desktop will go from your Windows 7 look… to a modified Windows 7 look that is reminiscent of the Commodore 64. Open an application to see all the changes … notice the old-style font in the Window boarder and menus. This theme also changes your Computer, Recycle Bin, and User folder icons to Commodore 64-inspired icons. And, if you want to go back to the standard Windows 7 look and feel, it’s only a click away in the Personalization dialog.  Right-click on your desktop, select Personalize, and then choose the theme you want.   Conclusion Although this doesn’t give you the real look and feel of the Commodore 64, it is still a fun way to experience a bit of computer nostalgia.  There are tons of excellent themes available for Windows 7, so check back for more exciting ways to customize your desktop! Link Download the Commodore 64 theme for Windows 7 Similar Articles Productive Geek Tips Make MSE Create a Restore Point Before Cleaning MalwareMake Ubuntu Automatically Save Changes to Your SessionMake Windows Vista Shut Down Services QuickerChange Your Computer Name in Windows 7 or VistaMake Windows 7 or Vista Log On Automatically TouchFreeze Alternative in AutoHotkey The Icy Undertow Desktop Windows Home Server – Backup to LAN The Clear & Clean Desktop Use This Bookmarklet to Easily Get Albums Use AutoHotkey to Assign a Hotkey to a Specific Window Latest Software Reviews Tinyhacker Random Tips DVDFab 6 Revo Uninstaller Pro Registry Mechanic 9 for Windows PC Tools Internet Security Suite 2010 Dark Side of the Moon (8-bit) Norwegian Life If Web Browsers Were Modes of Transportation Google Translate (for animals) Out of 100 Tweeters Roadkill’s Scan Port scans for open ports

    Read the article

  • Disable Opera Thumbnail Previews on Windows 7 Taskbar

    - by Asian Angel
    If you are one of the people who does not care for the Taskbar Thumbnail Previews in Windows 7 then we have a quick and easy way for you to turn them off in Opera Browser. Before Here is our Opera Browser with four tabs full of HTG Network goodness… Hovering the mouse over the Taskbar Icon gives a nice preview of each tabs content. Looking closer you can see the fanned edge on the Taskbar Icon indicating that there are multiple tabs open. This is all good but what if you just want something simpler? Disabling the Previews If you want to disable the Taskbar Thumbnail Previews in Opera you will need to type opera:config in the Address Bar and press Enter. Once you have done that, you will see a condensed listing for all of Opera’s preferences. There is one Preference Category that we need to look for…User Prefs. Note: While a Quick Find Search could be conducted for the entry that needs to be modified, we have chosen to show the full method here. After scrolling down and finding the User Prefs category you will need to expand the section. Notice the size of the scrollbar in comparison with the screenshot above…there is quite a lot that you can look at and finesse in Opera if desired. Scroll down until you find the Use Windows 7 Taskbar Thumbnails entry. Uncheck the box but do not close the opera:config Tab yet…or your changes will not take effect. Scroll down once more until you reach the end of the User Prefs category and click Save. With this particular modification you will need to restart Opera after clicking OK. After restarting Opera the Taskbar Icon and Taskbar Thumbnail Preview will revert to the minimal Windows 7 default as shown here. You can see Opera’s Tab Bar in the thumbnail and the Taskbar Icon no longer has a “fanned edge”. Conclusion If you want to disable Opera’s Taskbar Thumbnail Previews on your Windows 7 system, then this quick modification will help get it sorted out in just a few moments. Similar Articles Productive Geek Tips Disable IE 8 Thumbnail Previews on Windows 7 TaskbarIncrease the size of Taskbar Preview Thumbnails in Windows 7Vista Style Popup Previews for Firefox TabsEnable Thumbnail Previews for Firefox in Windows 7 TaskbarWorkaround for Vista Taskbar Thumbnail Previews Not Showing Correctly TouchFreeze Alternative in AutoHotkey The Icy Undertow Desktop Windows Home Server – Backup to LAN The Clear & Clean Desktop Use This Bookmarklet to Easily Get Albums Use AutoHotkey to Assign a Hotkey to a Specific Window Latest Software Reviews Tinyhacker Random Tips All My Movies 5.9 CloudBerry Online Backup 1.5 for Windows Home Server Snagit 10 VMware Workstation 7 10 Superb Firefox Wallpapers OpenDNS Guide Google TV The iPod Revolution Ultimate Boot CD can help when disaster strikes Windows Firewall with Advanced Security – How To Guides

    Read the article

  • The broken Promise of the Mobile Web

    - by Rick Strahl
    High end mobile devices have been with us now for almost 7 years and they have utterly transformed the way we access information. Mobile phones and smartphones that have access to the Internet and host smart applications are in the hands of a large percentage of the population of the world. In many places even very remote, cell phones and even smart phones are a common sight. I’ll never forget when I was in India in 2011 I was up in the Southern Indian mountains riding an elephant out of a tiny local village, with an elephant herder in front riding atop of the elephant in front of us. He was dressed in traditional garb with the loin wrap and head cloth/turban as did quite a few of the locals in this small out of the way and not so touristy village. So we’re slowly trundling along in the forest and he’s lazily using his stick to guide the elephant and… 10 minutes in he pulls out his cell phone from his sash and starts texting. In the middle of texting a huge pig jumps out from the side of the trail and he takes a picture running across our path in the jungle! So yeah, mobile technology is very pervasive and it’s reached into even very buried and unexpected parts of this world. Apps are still King Apps currently rule the roost when it comes to mobile devices and the applications that run on them. If there’s something that you need on your mobile device your first step usually is to look for an app, not use your browser. But native app development remains a pain in the butt, with the requirement to have to support 2 or 3 completely separate platforms. There are solutions that try to bridge that gap. Xamarin is on a tear at the moment, providing their cross-device toolkit to build applications using C#. While Xamarin tools are impressive – and also *very* expensive – they only address part of the development madness that is app development. There are still specific device integration isssues, dealing with the different developer programs, security and certificate setups and all that other noise that surrounds app development. There’s also PhoneGap/Cordova which provides a hybrid solution that involves creating local HTML/CSS/JavaScript based applications, and then packaging them to run in a specialized App container that can run on most mobile device platforms using a WebView interface. This allows for using of HTML technology, but it also still requires all the set up, configuration of APIs, security keys and certification and submission and deployment process just like native applications – you actually lose many of the benefits that  Web based apps bring. The big selling point of Cordova is that you get to use HTML have the ability to build your UI once for all platforms and run across all of them – but the rest of the app process remains in place. Apps can be a big pain to create and manage especially when we are talking about specialized or vertical business applications that aren’t geared at the mainstream market and that don’t fit the ‘store’ model. If you’re building a small intra department application you don’t want to deal with multiple device platforms and certification etc. for various public or corporate app stores. That model is simply not a good fit both from the development and deployment perspective. Even for commercial, big ticket apps, HTML as a UI platform offers many advantages over native, from write-once run-anywhere, to remote maintenance, single point of management and failure to having full control over the application as opposed to have the app store overloads censor you. In a lot of ways Web based HTML/CSS/JavaScript applications have so much potential for building better solutions based on existing Web technologies for the very same reasons a lot of content years ago moved off the desktop to the Web. To me the Web as a mobile platform makes perfect sense, but the reality of today’s Mobile Web unfortunately looks a little different… Where’s the Love for the Mobile Web? Yet here we are in the middle of 2014, nearly 7 years after the first iPhone was released and brought the promise of rich interactive information at your fingertips, and yet we still don’t really have a solid mobile Web platform. I know what you’re thinking: “But we have lots of HTML/JavaScript/CSS features that allows us to build nice mobile interfaces”. I agree to a point – it’s actually quite possible to build nice looking, rich and capable Web UI today. We have media queries to deal with varied display sizes, CSS transforms for smooth animations and transitions, tons of CSS improvements in CSS 3 that facilitate rich layout, a host of APIs geared towards mobile device features and lately even a number of JavaScript framework choices that facilitate development of multi-screen apps in a consistent manner. Personally I’ve been working a lot with AngularJs and heavily modified Bootstrap themes to build mobile first UIs and that’s been working very well to provide highly usable and attractive UI for typical mobile business applications. From the pure UI perspective things actually look very good. Not just about the UI But it’s not just about the UI - it’s also about integration with the mobile device. When it comes to putting all those pieces together into what amounts to a consolidated platform to build mobile Web applications, I think we still have a ways to go… there are a lot of missing pieces to make it all work together and integrate with the device more smoothly, and more importantly to make it work uniformly across the majority of devices. I think there are a number of reasons for this. Slow Standards Adoption HTML standards implementations and ratification has been dreadfully slow, and browser vendors all seem to pick and choose different pieces of the technology they implement. The end result is that we have a capable UI platform that’s missing some of the infrastructure pieces to make it whole on mobile devices. There’s lots of potential but what is lacking that final 10% to build truly compelling mobile applications that can compete favorably with native applications. Some of it is the fragmentation of browsers and the slow evolution of the mobile specific HTML APIs. A host of mobile standards exist but many of the standards are in the early review stage and they have been there stuck for long periods of time and seem to move at a glacial pace. Browser vendors seem even slower to implement them, and for good reason – non-ratified standards mean that implementations may change and vendor implementations tend to be experimental and  likely have to be changed later. Neither Vendors or developers are not keen on changing standards. This is the typical chicken and egg scenario, but without some forward momentum from some party we end up stuck in the mud. It seems that either the standards bodies or the vendors need to carry the torch forward and that doesn’t seem to be happening quickly enough. Mobile Device Integration just isn’t good enough Current standards are not far reaching enough to address a number of the use case scenarios necessary for many mobile applications. While not every application needs to have access to all mobile device features, almost every mobile application could benefit from some integration with other parts of the mobile device platform. Integration with GPS, phone, media, messaging, notifications, linking and contacts system are benefits that are unique to mobile applications and could be widely used, but are mostly (with the exception of GPS) inaccessible for Web based applications today. Unfortunately trying to do most of this today only with a mobile Web browser is a losing battle. Aside from PhoneGap/Cordova’s app centric model with its own custom API accessing mobile device features and the token exception of the GeoLocation API, most device integration features are not widely supported by the current crop of mobile browsers. For example there’s no usable messaging API that allows access to SMS or contacts from HTML. Even obvious components like the Media Capture API are only implemented partially by mobile devices. There are alternatives and workarounds for some of these interfaces by using browser specific code, but that’s might ugly and something that I thought we were trying to leave behind with newer browser standards. But it’s not quite working out that way. It’s utterly perplexing to me that mobile standards like Media Capture and Streams, Media Gallery Access, Responsive Images, Messaging API, Contacts Manager API have only minimal or no traction at all today. Keep in mind we’ve had mobile browsers for nearly 7 years now, and yet we still have to think about how to get access to an image from the image gallery or the camera on some devices? Heck Windows Phone IE Mobile just gained the ability to upload images recently in the Windows 8.1 Update – that’s feature that HTML has had for 20 years! These are simple concepts and common problems that should have been solved a long time ago. It’s extremely frustrating to see build 90% of a mobile Web app with relative ease and then hit a brick wall for the remaining 10%, which often can be show stoppers. The remaining 10% have to do with platform integration, browser differences and working around the limitations that browsers and ‘pinned’ applications impose on HTML applications. The maddening part is that these limitations seem arbitrary as they could easily work on all mobile platforms. For example, SMS has a URL Moniker interface that sort of works on Android, works badly with iOS (only works if the address is already in the contact list) and not at all on Windows Phone. There’s no reason this shouldn’t work universally using the same interface – after all all phones have supported SMS since before the year 2000! But, it doesn’t have to be this way Change can happen very quickly. Take the GeoLocation API for example. Geolocation has taken off at the very beginning of the mobile device era and today it works well, provides the necessary security (a big concern for many mobile APIs), and is supported by just about all major mobile and even desktop browsers today. It handles security concerns via prompts to avoid unwanted access which is a model that would work for most other device APIs in a similar fashion. One time approval and occasional re-approval if code changes or caches expire. Simple and only slightly intrusive. It all works well, even though GeoLocation actually has some physical limitations, such as representing the current location when no GPS device is present. Yet this is a solved problem, where other APIs that are conceptually much simpler to implement have failed to gain any traction at all. Technically none of these APIs should be a problem to implement, but it appears that the momentum is just not there. Inadequate Web Application Linking and Activation Another important piece of the puzzle missing is the integration of HTML based Web applications. Today HTML based applications are not first class citizens on mobile operating systems. When talking about HTML based content there’s a big difference between content and applications. Content is great for search engine discovery and plain browser usage. Content is usually accessed intermittently and permanent linking is not so critical for this type of content.  But applications have different needs. Applications need to be started up quickly and must be easily switchable to support a multi-tasking user workflow. Therefore, it’s pretty crucial that mobile Web apps are integrated into the underlying mobile OS and work with the standard task management features. Unfortunately this integration is not as smooth as it should be. It starts with actually trying to find mobile Web applications, to ‘installing’ them onto a phone in an easily accessible manner in a prominent position. The experience of discovering a Mobile Web ‘App’ and making it sticky is by no means as easy or satisfying. Today the way you’d go about this is: Open the browser Search for a Web Site in the browser with your search engine of choice Hope that you find the right site Hope that you actually find a site that works for your mobile device Click on the link and run the app in a fully chrome’d browser instance (read tiny surface area) Pin the app to the home screen (with all the limitations outline above) Hope you pointed at the right URL when you pinned Even for you and me as developers, there are a few steps in there that are painful and annoying, but think about the average user. First figuring out how to search for a specific site or URL? And then pinning the app and hopefully from the right location? You’ve probably lost more than half of your audience at that point. This experience sucks. For developers too this process is painful since app developers can’t control the shortcut creation directly. This problem often gets solved by crazy coding schemes, with annoying pop-ups that try to get people to create shortcuts via fancy animations that are both annoying and add overhead to each and every application that implements this sort of thing differently. And that’s not the end of it - getting the link onto the home screen with an application icon varies quite a bit between browsers. Apple’s non-standard meta tags are prominent and they work with iOS and Android (only more recent versions), but not on Windows Phone. Windows Phone instead requires you to create an actual screen or rather a partial screen be captured for a shortcut in the tile manager. Who had that brilliant idea I wonder? Surprisingly Chrome on recent Android versions seems to actually get it right – icons use pngs, pinning is easy and pinned applications properly behave like standalone apps and retain the browser’s active page state and content. Each of the platforms has a different way to specify icons (WP doesn’t allow you to use an icon image at all), and the most widely used interface in use today is a bunch of Apple specific meta tags that other browsers choose to support. The question is: Why is there no standard implementation for installing shortcuts across mobile platforms using an official format rather than a proprietary one? Then there’s iOS and the crazy way it treats home screen linked URLs using a crazy hybrid format that is neither as capable as a Web app running in Safari nor a WebView hosted application. Moving off the Web ‘app’ link when switching to another app actually causes the browser and preview it to ‘blank out’ the Web application in the Task View (see screenshot on the right). Then, when the ‘app’ is reactivated it ends up completely restarting the browser with the original link. This is crazy behavior that you can’t easily work around. In some situations you might be able to store the application state and restore it using LocalStorage, but for many scenarios that involve complex data sources (like say Google Maps) that’s not a possibility. The only reason for this screwed up behavior I can think of is that it is deliberate to make Web apps a pain in the butt to use and forcing users trough the App Store/PhoneGap/Cordova route. App linking and management is a very basic problem – something that we essentially have solved in every desktop browser – yet on mobile devices where it arguably matters a lot more to have easy access to web content we have to jump through hoops to have even a remotely decent linking/activation experience across browsers. Where’s the Money? It’s not surprising that device home screen integration and Mobile Web support in general is in such dismal shape – the mobile OS vendors benefit financially from App store sales and have little to gain from Web based applications that bypass the App store and the cash cow that it presents. On top of that, platform specific vendor lock-in of both end users and developers who have invested in hardware, apps and consumables is something that mobile platform vendors actually aspire to. Web based interfaces that are cross-platform are the anti-thesis of that and so again it’s no surprise that the mobile Web is on a struggling path. But – that may be changing. More and more we’re seeing operations shifting to services that are subscription based or otherwise collect money for usage, and that may drive more progress into the Web direction in the end . Nothing like the almighty dollar to drive innovation forward. Do we need a Mobile Web App Store? As much as I dislike moderated experiences in today’s massive App Stores, they do at least provide one single place to look for apps for your device. I think we could really use some sort of registry, that could provide something akin to an app store for mobile Web apps, to make it easier to actually find mobile applications. This could take the form of a specialized search engine, or maybe a more formal store/registry like structure. Something like apt-get/chocolatey for Web apps. It could be curated and provide at least some feedback and reviews that might help with the integrity of applications. Coupled to that could be a native application on each platform that would allow searching and browsing of the registry and then also handle installation in the form of providing the home screen linking, plus maybe an initial security configuration that determines what features are allowed access to for the app. I’m not holding my breath. In order for this sort of thing to take off and gain widespread appeal, a lot of coordination would be required. And in order to get enough traction it would have to come from a well known entity – a mobile Web app store from a no name source is unlikely to gain high enough usage numbers to make a difference. In a way this would eliminate some of the freedom of the Web, but of course this would also be an optional search path in addition to the standard open Web search mechanisms to find and access content today. Security Security is a big deal, and one of the perceived reasons why so many IT professionals appear to be willing to go back to the walled garden of deployed apps is that Apps are perceived as safe due to the official review and curation of the App stores. Curated stores are supposed to protect you from malware, illegal and misleading content. It doesn’t always work out that way and all the major vendors have had issues with security and the review process at some time or another. Security is critical, but I also think that Web applications in general pose less of a security threat than native applications, by nature of the sandboxed browser and JavaScript environments. Web applications run externally completely and in the HTML and JavaScript sandboxes, with only a very few controlled APIs allowing access to device specific features. And as discussed earlier – security for any device interaction can be granted the same for mobile applications through a Web browser, as they can for native applications either via explicit policies loaded from the Web, or via prompting as GeoLocation does today. Security is important, but it’s certainly solvable problem for Web applications even those that need to access device hardware. Security shouldn’t be a reason for Web apps to be an equal player in mobile applications. Apps are winning, but haven’t we been here before? So now we’re finding ourselves back in an era of installed app, rather than Web based and managed apps. Only it’s even worse today than with Desktop applications, in that the apps are going through a gatekeeper that charges a toll and censors what you can and can’t do in your apps. Frankly it’s a mystery to me why anybody would buy into this model and why it’s lasted this long when we’ve already been through this process. It’s crazy… It’s really a shame that this regression is happening. We have the technology to make mobile Web apps much more prominent, but yet we’re basically held back by what seems little more than bureaucracy, partisan bickering and self interest of the major parties involved. Back in the day of the desktop it was Internet Explorer’s 98+%  market shareholding back the Web from improvements for many years – now it’s the combined mobile OS market in control of the mobile browsers. If mobile Web apps were allowed to be treated the same as native apps with simple ways to install and run them consistently and persistently, that would go a long way to making mobile applications much more usable and seriously viable alternatives to native apps. But as it is mobile apps have a severe disadvantage in placement and operation. There are a few bright spots in all of this. Mozilla’s FireFoxOs is embracing the Web for it’s mobile OS by essentially building every app out of HTML and JavaScript based content. It supports both packaged and certified package modes (that can be put into the app store), and Open Web apps that are loaded and run completely off the Web and can also cache locally for offline operation using a manifest. Open Web apps are treated as full class citizens in FireFoxOS and run using the same mechanism as installed apps. Unfortunately FireFoxOs is getting a slow start with minimal device support and specifically targeting the low end market. We can hope that this approach will change and catch on with other vendors, but that’s also an uphill battle given the conflict of interest with platform lock in that it represents. Recent versions of Android also seem to be working reasonably well with mobile application integration onto the desktop and activation out of the box. Although it still uses the Apple meta tags to find icons and behavior settings, everything at least works as you would expect – icons to the desktop on pinning, WebView based full screen activation, and reliable application persistence as the browser/app is treated like a real application. Hopefully iOS will at some point provide this same level of rudimentary Web app support. What’s also interesting to me is that Microsoft hasn’t picked up on the obvious need for a solid Web App platform. Being a distant third in the mobile OS war, Microsoft certainly has nothing to lose and everything to gain by using fresh ideas and expanding into areas that the other major vendors are neglecting. But instead Microsoft is trying to beat the market leaders at their own game, fighting on their adversary’s terms instead of taking a new tack. Providing a kick ass mobile Web platform that takes the lead on some of the proposed mobile APIs would be something positive that Microsoft could do to improve its miserable position in the mobile device market. Where are we at with Mobile Web? It sure sounds like I’m really down on the Mobile Web, right? I’ve built a number of mobile apps in the last year and while overall result and response has been very positive to what we were able to accomplish in terms of UI, getting that final 10% that required device integration dialed was an absolute nightmare on every single one of them. Big compromises had to be made and some features were left out or had to be modified for some devices. In two cases we opted to go the Cordova route in order to get the integration we needed, along with the extra pain involved in that process. Unless you’re not integrating with device features and you don’t care deeply about a smooth integration with the mobile desktop, mobile Web development is fraught with frustration. So, yes I’m frustrated! But it’s not for lack of wanting the mobile Web to succeed. I am still a firm believer that we will eventually arrive a much more functional mobile Web platform that allows access to the most common device features in a sensible way. It wouldn't be difficult for device platform vendors to make Web based applications first class citizens on mobile devices. But unfortunately it looks like it will still be some time before this happens. So, what’s your experience building mobile Web apps? Are you finding similar issues? Just giving up on raw Web applications and building PhoneGap apps instead? Completely skipping the Web and going native? Leave a comment for discussion. Resources Rick Strahl on DotNet Rocks talking about Mobile Web© Rick Strahl, West Wind Technologies, 2005-2014Posted in HTML5  Mobile   Tweet !function(d,s,id){var js,fjs=d.getElementsByTagName(s)[0];if(!d.getElementById(id)){js=d.createElement(s);js.id=id;js.src="//platform.twitter.com/widgets.js";fjs.parentNode.insertBefore(js,fjs);}}(document,"script","twitter-wjs"); (function() { var po = document.createElement('script'); po.type = 'text/javascript'; po.async = true; po.src = 'https://apis.google.com/js/plusone.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(po, s); })();

    Read the article

  • SQL SERVER – WRITELOG – Wait Type – Day 17 of 28

    - by pinaldave
    WRITELOG is one of the most interesting wait types. So far we have seen a lot of different wait types, but this log type is associated with log file which makes it interesting to deal with. From Book On-Line: WRITELOG Occurs while waiting for a log flush to complete. Common operations that cause log flushes are checkpoints and transaction commits. WRITELOG Explanation: This wait type is usually seen in the heavy transactional database. When data is modified, it is written both on the log cache and buffer cache. This wait type occurs when data in the log cache is flushing to the disk. During this time, the session has to wait due to WRITELOG. I have recently seen this wait type’s persistence at my client’s place, where one of the long-running transactions was stopped by the user causing it to roll back. In the future, I will see if I could re-create this situation once again on my machine to validate the relation. Reducing WRITELOG wait: There are several suggestions to reduce this wait stats: Move Transaction Log to Separate Disk from mdf and other files. Avoid cursor-like coding methodology and frequent committing of statements. Find the most active file based on IO stall time based on the script written over here. You can also use fn_virtualfilestats to find IO-related issues using the script mentioned over here. Check the IO-related counters (PhysicalDisk:Avg.Disk Queue Length, PhysicalDisk:Disk Read Bytes/sec and PhysicalDisk :Disk Write Bytes/sec) for additional details. Read about them over here. There are two excellent resources by Paul Randal, I suggest you understand the subject from those videos. The links to videos are here and here. Note: The information presented here is from my experience and there is no way that I claim it to be accurate. I suggest reading Book OnLine for further clarification. All the discussion of Wait Stats in this blog is generic and varies from system to system. It is recommended that you test this on a development server before implementing it to a production server. Reference: Pinal Dave (http://blog.SQLAuthority.com) Filed under: Pinal Dave, PostADay, SQL, SQL Authority, SQL Query, SQL Scripts, SQL Server, SQL Tips and Tricks, SQL Wait Stats, SQL Wait Types, T SQL, Technology

    Read the article

  • Umbraco Permissions Script - Secure Version

    - by Vizioz Limited
    Back in May I blogged about how to set Permissions for Umbraco using SetACL to set the appropriate directory permissions based on the installation recommendations.Recently I have been working on a site for a client who wanted every security item to be locked down as tightly as possible. And so I modified the script based on the Umbraco security best practices, I thought I'd share it with everyone, if I have missed anything, or if anyone has any suggestions on how to improve this, please let me know :)Please refer to my previous post regarding the SetAcl command line application that you will need.I suggest you save the following into a batch file called: umbPermSecure.batecho offREM Script to setup the Security Permissions for an Umbraco siteREM This script will give your machine Network Service the minimum rights requiredREM for Umbraco to workREM I suggest you update this script to also remove any users who do not need REM access to the web foldersREM **** Pre-requisites ****REM You will need to download - http://setacl.sourceforge.net/REM It is assumed that you have stored SetACL in a directory called, C:\SetACL ifREM not, you will need to modify the script.REM **** Usage ****REM You need to pass in the path for the root of your Umbraco directoryREM E.g. umbPermSecure.bat C:\inetpub\umbracoroot@echo umbPermSecure.bat - Script to set Umbraco File and Directory Permissions@echo based on the Umbraco Security Best Practices Document (13th March 2009)@echo Published by Chris Houston - 19th October 2009@echo http://blog.vizioz.com@echo Adding READ only access SetACL.exe -on "%1" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\web.config" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\bin" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\umbraco" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"@echo Adding READ and EXECUTE access SetACL.exe -on "%1\app_code" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read_ex" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\usercontrols" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read_ex" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"@echo Adding READ, WRITE and MODIFY access SetACL.exe -on "%1\config" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -ace "n:%computername%\NETWORK SERVICE;p:change" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\css" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -ace "n:%computername%\NETWORK SERVICE;p:change" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\data" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -ace "n:%computername%\NETWORK SERVICE;p:change" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\masterpages" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -ace "n:%computername%\NETWORK SERVICE;p:change" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\media" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -ace "n:%computername%\NETWORK SERVICE;p:change" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\python" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -ace "n:%computername%\NETWORK SERVICE;p:change" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\scripts" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -ace "n:%computername%\NETWORK SERVICE;p:change" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"SetACL.exe -on "%1\xslt" -ot file -actn ace -ace "n:%computername%\NETWORK SERVICE;p:read" -ace "n:%computername%\NETWORK SERVICE;p:change" -actn clear -clr "dacl,sacl" -log "c:\setacl\log.txt"

    Read the article

  • TFS Auto Shelve - New Visual Studio 2010 / TFS 2010 Extension

    - by MikeParks
    We've been working with the Visual Studio 2010 SDK and the TFS 2010 SDK a lot recently to create new Visual Studio Extensions. You can find these extensions in the Visual Studio Gallery. If you're a developer/programmer, you should check it out, they have some pretty cool tools out there. I'd be surprised if you told me you went there and couldn't find any tools that could help you. One of the new extensions Cory and I made is called TFS Auto Shelve. Check out the description and read about it below. If you're interested and you have VS 2010 w/TFS 2010, feel free to try it out and let us know what you think. You can download it here: http://visualstudiogallery.msdn.microsoft.com/en-us/080540cb-e35f-4651-b71c-86c73e4a633d   Here's a description and screenshots of what it does: Automatically shelves the latest version of all pending changes from local TFS workspaces to the TFS Server every "x" number of minutes when solutions are opened.   ·         Purpose o    Created for Team Foundation Server 2010 and Visual Studio 2010 o    This tool is mainly aimed at the Programmer/Developer audience so they can always have the latest copy of their pending changes backed up to the TFS Server while coding ·         Functionality o    Menu options become active and automatic shelving begins when a solution that mapped to a TFS Workspace is opened in Visual Studio o    In Tools > TFS Auto Shelve (Running/NotRunning):  Automatic shelving can be turned on/off o    In Tools > TFS Auto Shelve Now : Shelve all code can be manually triggered o    Each TFS workspace has its own shelveset which is re-used to save the latest version of pending changes o    Shelvesets are named as Base Name + Workspace Name o    Shelveset comment contains item count o    If there are no pending changes, no shelvesets will be created/updated o    If a solution is opened that is not mapped to a TFS Workspace, menu options are disabled since shelving only works for mapped workspaces. ·         Configuration o    In Tools > Options > TFS Auto Shelve Options: Base Name is configurable o    In Tools > Options > TFS Auto Shelve Options: "x" number of minutes is configurable in options ·         Logging o    Custom Visual Studio Activity Logging is implemented. If you run into any errors, please startup Visual Studio with the /log switch, re-create the error, then close Visual Studio. You can browse to “%AppData%\Microsoft\VisualStudio\10.0\ActivityLog.XML” to view the log. Please feel free to inform us of any errors you see and we can work it out via email. ·         Other Helpful Information o    To view shelvesets, open Source Control Explorer, click on File > Source Control > Unshelve Pending Changes o    Workspaces can be modified by opening the Source Control Explorer > Clicking on Workspaces drop down > Click Workspaces… > Click Add / Edit / Removed   Thanks! - Mike

    Read the article

  • On configuring GC 10.2.0.5 to monitor LISTENER SCAN using UDMs ...

    - by [email protected]
    Hi,Looks like Grid Control 10.2.0.5 is not fully prepared for monitoringthe Grid Infrastructure (11gR2).Even I'm pretty sure the upcoming version of GC (11g) will of course support all the new features of 11gR2, some customersare asking for some "hand-made" procedures for monitoring all the new stuff.I think one of the most critical components that cant be monitored are the LISTENER SCAN, so I have developed a little script for doing sousing the GC User Defined Metrics ( at host level )I am more than happy to share with you:#!/bin/ksh   ###    NAME###     monitor_scan.sh######    DESCRIPTION###      SCAN Listener monitoring######    RETURNS######    NOTES######    MODIFIED           (DD/MM/YY)###      Oracle            25/03/10     - Creation###export ORACLE_HOME=/opt/oracle/soft/11.2/gridRSC_KEY=$1AWK=/sbin/awk   LISTENER_DOWN_COUNT=$(${ORACLE_HOME}/bin/crsctl status resource -w 'TYPE = ora.scan_listener.type' | grep OFFLINE | wc -l)if [ ${LISTENER_DOWN_COUNT} != 0 ]; then  SCAN_DOWN_LIST=$(${ORACLE_HOME}/bin/crsctl status resource  -w 'TYPE = ora.scan_listener.type' | $AWK \ 'BEGIN { FS="="; state = 0; }  $1~/NAME/ && $2~/'$RSC_KEY'/ {appname = $2; state=1};  state == 0 {next;}  $1~/TARGET/ && state == 1 {apptarget = $2; state=2;}  $1~/STATE/ && state == 2 {appstate = $2; state=3;}  state == 3 {printf "%-45s %-10s %-18s\n", appname, apptarget, appstate; state=0;}' | grep OFFLINE | awk '{ print $1 }')  echo em_result=ALERT  echo em_message=There are LISTENER SCAN with down status: [${SCAN_DOWN_LIST}]else  echo em_result=NORMAL  echo em_message=All SCAN Listener are UPfiHope it helpsL

    Read the article

  • OWB 11gR2 &ndash; Flexible and extensible

    - by David Allan
    The Oracle data integration extensibility capabilities are something I love, nothing more frustrating than a tool or platform that is very constraining. I think extensibility and flexibility are invaluable capabilities in the data integration arena. I liked Uli Bethke's posting on some extensibility capabilities with ODI (see Nesting ODI Substitution Method Calls here), he has some useful guidance on making customizations to existing KMs, nice to learn by example. I thought I'd illustrate the same capabilities with ODI's partner OWB for the OWB community. There is a whole new world of potential. The LKM/IKM/CKM/JKMs are the primary templates that are supported (plus the Oracle Target code template), so there is a lot of potential for customizing and extending the product in this release. Enough waffle... Diving in at the deep end from Uli's post, in OWB the table operator has a number of additional properties in OWB 11gR2 that let you annotate the column usage with ODI-like properties such as the slowly changing usage or for your own user-defined purpose as in Uli's post, below you see for the target table SALES_TARGET we can use the UD5 property which when assigned the code template (knowledge module) which has been modified with Uli's change we can do custom things such as creating indices - provides The code template used by the mapping has the additional step which is basically the code illustrated from Uli's posting just used directly, the ODI 10g substitution references also supported from within OWB's runtime. Now to see whether this does what we expect before we execute it, we can check out the generated code similar to how the traditional mapping generation and preview works, you do this by clicking on the 'Inspect Code' button on the execution units code template assignment. This then  creates another tab with prefix 'Code - <mapping name>' where the generated code is put, scrolling down we find the last step with the indices being created, looks good, so we are ready to deploy and execute. After executing the mapping we can then use the 'Audit Information' panel (select the mapping in the designer tree and click on View/Audit Information), this gives us a view of the execution where we can drill into the tasks that were executed and inspect both the template and the generated code that was executed and any potential errors. Reflecting back on earlier versions of OWB, these were the kinds of features that were always highly desirable, getting under the hood of the code generation and tweaking bit and pieces - fun and powerful stuff! We can step it up a bit here and explore some further ideas. The example below is a daisy-chained set of execution units where the intermediate table is a target of one unit and the source for another. We want that table to be a global temporary table, so can tweak the templates. Back to the copy of SQL Control Append (for demo purposes) we modify the create target table step to make the table a global temporary table, with the option of on commit preserve rows. You can get a feel for some of the customizations and changes possible, providing some great flexibility and extensibility for the data integration tools.

    Read the article

  • Tips on Migrating from AquaLogic .NET Accelerator to WebCenter WSRP Producer for .NET

    - by user647124
    This year I embarked on a journey to migrate a group of ASP.NET web applications developed to integrate with WebLogic Portal 9.2 via the AquaLogic® Interaction .NET Application Accelerator 1.0 to instead use the Oracle WebCenter WSRP Producer for .NET and integrated with WebLogic Portal 10.3.4. It has been a very winding path and this blog entry is intended to share both the lessons learned and relevant approaches that led to those learnings. Like most journeys of discovery, it was not a direct path, and there are notes to let you know when it is practical to skip a section if you are in a hurry to get from here to there. For the Curious From the perspective of necessity, this section would be better at the end. If it were there, though, it would probably be read by far fewer people, including those that are actually interested in these types of sections. Those in a hurry may skip past and be none the worst for it in dealing with the hands-on bits of performing a migration from .NET Accelerator to WSRP Producer. For others who want to talk about why they did what they did after they did it, or just want to know for themselves, enjoy. A Brief (and edited) History of the WSRP for .NET Technologies (as Relevant to the this Post) Note: This section is for those who are curious about why the migration path is not as simple as many other Oracle technologies. You can skip this section in its entirety and still be just as competent in performing a migration as if you had read it. The currently deployed architecture that was to be migrated and upgraded achieved initial integration between .NET and J2EE over the WSRP protocol through the use of The AquaLogic Interaction .NET Application Accelerator. The .NET Accelerator allowed the applications that were written in ASP.NET and deployed on a Microsoft Internet Information Server (IIS) to interact with a WebLogic Portal application deployed on a WebLogic (J2EE application) Server (both version 9.2, the state of the art at the time of its creation). At the time this architectural decision for the application was made, both the AquaLogic and WebLogic brands were owned by BEA Systems. The AquaLogic brand included products acquired by BEA through the acquisition of Plumtree, whose flagship product was a portal platform available in both J2EE and .NET versions. As part of this dual technology support an adaptor was created to facilitate the use of WSRP as a communication protocol where customers wished to integrate components from both versions of the Plumtree portal. The adapter evolved over several product generations to include a broad array of both standard and proprietary WSRP integration capabilities. Later, BEA Systems was acquired by Oracle. Over the course of several years Oracle has acquired a large number of portal applications and has taken the strategic direction to migrate users of these myriad (and formerly competitive) products to the Oracle WebCenter technology stack. As part of Oracle’s strategic technology roadmap, older portal products are being schedule for end of life, including the portal products that were part of the BEA acquisition. The .NET Accelerator has been modified over a very long period of time with features driven by users of that product and developed under three different vendors (each a direct competitor in the same solution space prior to merger). The Oracle WebCenter WSRP Producer for .NET was introduced much more recently with the key objective to specifically address the needs of the WebCenter customers developing solutions accessible through both J2EE and .NET platforms utilizing the WSRP specifications. The Oracle Product Development Team also provides these insights on the drivers for developing the WSRP Producer: ***************************************** Support for ASP.NET AJAX. Controls using the ASP.NET AJAX script manager do not function properly in the Application Accelerator for .NET. Support 2 way SSL in WLP. This was not possible with the proxy/bridge set up in the existing Application Accelerator for .NET. Allow developers to code portlets (Web Parts) using the .NET framework rather than a proprietary framework. Developers had to use the Application Accelerator for .NET plug-ins to Visual Studio to manage preferences and profile data. This is now replaced with the .NET Framework Personalization (for preferences) and Profile providers. The WSRP Producer for .NET was created as a new way of developing .NET portlets. It was never designed to be an upgrade path for the Application Accelerator for .NET. .NET developers would create new .NET portlets with the WSRP Producer for .NET and leave any existing .NET portlets running in the Application Accelerator for .NET. ***************************************** The advantage to creating a new solution for WSRP is a product that is far easier for Oracle to maintain and support which in turn improves quality, reliability and maintainability for their customers. No changes to J2EE applications consuming the WSRP portlets previously rendered by the.NET Accelerator is required to migrate from the Aqualogic WSRP solution. For some customers using the .NET Accelerator the challenge is adapting their current .NET applications to work with the WSRP Producer (or any other WSRP adapter as they are proprietary by nature). Part of this adaptation is the need to deploy the .NET applications as a child to the WSRP producer web application as root. Differences between .NET Accelerator and WSRP Producer Note: This section is for those who are curious about why the migration is not as pluggable as something such as changing security providers in WebLogic Server. You can skip this section in its entirety and still be just as competent in performing a migration as if you had read it. The basic terminology used to describe the participating applications in a WSRP environment are the same when applied to either the .NET Accelerator or the WSRP Producer: Producer and Consumer. In both cases the .NET application serves as what is referred to as a WSRP environment as the Producer. The difference lies in how the two adapters create the WSRP translation of the .NET application. The .NET Accelerator, as the name implies, is meant to serve as a quick way of adding WSRP capability to a .NET application. As such, at a high level, the .NET Accelerator behaves as a proxy for requests between the .NET application and the WSRP Consumer. A WSRP request is sent from the consumer to the .NET Accelerator, the.NET Accelerator transforms this request into an ASP.NET request, receives the response, then transforms the response into a WSRP response. The .NET Accelerator is deployed as a stand-alone application on IIS. The WSRP Producer is deployed as a parent application on IIS and all ASP.NET modules that will be made available over WSRP are deployed as children of the WSRP Producer application. In this manner, the WSRP Producer acts more as a Request Filter than a proxy in the WSRP transactions between Producer and Consumer. Highly Recommended Enabling Logging Note: You can skip this section now, but you will most likely want to come back to it later, so why not just read it now? Logging is very helpful in tracking down the causes of any anomalies during testing of migrated portlets. To enable the WSRP Producer logging, update the Application_Start method in the Global.asax.cs for your .NET application by adding log4net.Config.XmlConfigurator.Configure(); IIS logs will usually (in a standard configuration) be in a sub folder under C:\WINDOWS\system32\LogFiles\W3SVC. WSRP Producer logs will be found at C:\Oracle\Middleware\WSRPProducerForDotNet\wsrpdefault\Logs\WSRPProducer.log InputTrace.webinfo and OutputTrace.webinfo are located under C:\Oracle\Middleware\WSRPProducerForDotNet\wsrpdefault and can be useful in debugging issues related to markup transformations. Things You Must Do Merge Web.Config Note: If you have been skipping all the sections that you can, now is the time to stop and pay attention J Because the existing .NET application will become a sub-application to the WSRP Producer, you will want to merge required settings from the existing Web.Config to the one in the WSRP Producer. Use the WSRP Producer Master Page The Master Page installed for the WSRP Producer provides common, hiddenform fields and JavaScripts to facilitate portlet instance management and display configuration when the child page is being rendered over WSRP. You add the Master Page by including it in the <@ Page declaration with MasterPageFile="~/portlets/Resources/MasterPages/WSRP.Master" . You then replace: <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" > <HTML> <HEAD> With <asp:Content ID="ContentHead1" ContentPlaceHolderID="wsrphead" Runat="Server"> And </HEAD> <body> <form id="theForm" method="post" runat="server"> With </asp:Content> <asp:Content ID="ContentBody1" ContentPlaceHolderID="Main" Runat="Server"> And finally </form> </body> </HTML> With </asp:Content> In the event you already use Master Pages, adapt your existing Master Pages to be sub masters. See Nested ASP.NET Master Pages for a detailed reference of how to do this. It Happened to Me, It Might Happen to You…Or Not Watch for Use of Session or Request in OnInit In the event the .NET application being modified has pages developed to assume the user has been authenticated in an earlier page request there may be direct or indirect references in the OnInit method to request or session objects that may not have been created yet. This will vary from application to application, so the recommended approach is to test first. If there is an issue with a page running as a WSRP portlet then check for potential references in the OnInit method (including references by methods called within OnInit) to session or request objects. If there are, the simplest solution is to create a new method and then call that method once the necessary object(s) is fully available. I find doing this at the start of the Page_Load method to be the simplest solution. Case Sensitivity .NET languages are not case sensitive, but Java is. This means it is possible to have many variations of SRC= and src= or .JPG and .jpg. The preferred solution is to make these mark up instances all lower case in your .NET application. This will allow the default Rewriter rules in wsrp-producer.xml to work as is. If this is not practical, then make duplicates of any rules where an issue is occurring due to upper or mixed case usage in the .NET application markup and match the case in use with the duplicate rule. For example: <RewriterRule> <LookFor>(href=\"([^\"]+)</LookFor> <ChangeToAbsolute>true</ChangeToAbsolute> <ApplyTo>.axd,.css</ApplyTo> <MakeResource>true</MakeResource> </RewriterRule> May need to be duplicated as: <RewriterRule> <LookFor>(HREF=\"([^\"]+)</LookFor> <ChangeToAbsolute>true</ChangeToAbsolute> <ApplyTo>.axd,.css</ApplyTo> <MakeResource>true</MakeResource> </RewriterRule> While it is possible to write a regular expression that will handle mixed case usage, it would be long and strenous to test and maintain, so the recommendation is to use duplicate rules. Is it Still Relative? Some .NET applications base relative paths with a fixed root location. With the introduction of the WSRP Producer, the root has moved up one level. References to ~/ will need to be updated to ~/portlets and many ../ paths will need another ../ in front. I Can See You But I Can’t Find You This issue was first discovered while debugging modules with code that referenced the form on a page from the code-behind by name and/or id. The initial error presented itself as run-time error that was difficult to interpret over WSRP but seemed clear when run as straight ASP.NET as it indicated that the object with the form name did not exist. Since the form name was no longer valid after implementing the WSRP Master Page, the likely fix seemed to simply update the references in the code. However, as the WSRP Master Page is external to the code, a compile time error resulted: Error      155         The name 'form1' does not exist in the current context                C:\Oracle\Middleware\WSRPProducerForDotNet\wsrpdefault\portlets\legacywebsite\module\Screens \Reporting.aspx.cs                51           52           legacywebsite.module Much hair-pulling research later it was discovered that it was the use of the FindControl method causing the issue. FindControl doesn’t work quite as expected once a Master Page has been introduced as the controls become embedded in controls, require a recursion to find them that is not part of the FindControl method. In code where the page form is referenced by name, there are two steps to the solution. First, the form needs to be referenced in code generically with Page.Form. For example, this: ToggleControl ctrl = new ToggleControl(frmManualEntry, FunctionLibrary.ParseArrayLst(userObj.Roles)); Becomes this: ToggleControl ctrl = new ToggleControl(Page.Form, FunctionLibrary.ParseArrayLst(userObj.Roles)); Generally the form id is referenced in most ASP.NET applications as a path to a control on the form. To reach the control once a MasterPage has been added requires an additional method to recurse through the controls collections within the form and find the control ID. The following method (found at Rick Strahl's Web Log) corrects this very nicely: public static Control FindControlRecursive(Control Root, string Id) { if (Root.ID == Id) return Root; foreach (Control Ctl in Root.Controls) { Control FoundCtl = FindControlRecursive(Ctl, Id); if (FoundCtl != null) return FoundCtl; } return null; } Where the form name is not referenced, simply using the FindControlRecursive method in place of FindControl will be all that is necessary. Following the second part of the example referenced earlier, the method called with Page.Form changes its value extraction code block from this: Label lblErrMsg = (Label)frmRef.FindControl("lblBRMsg" To this: Label lblErrMsg = (Label) FunctionLibrary.FindControlRecursive(frmRef, "lblBRMsg" The Master That Won’t Step Aside In most migrations it is preferable to make as few changes as possible. In one case I ran across an existing Master Page that would not function as a sub-Master Page. While it would probably have been educational to trace down why, the expedient process of updating it to take the place of the WSRP Master Page is the route I took. The changes are highlighted below: … <asp:ContentPlaceHolder ID="wsrphead" runat="server"></asp:ContentPlaceHolder> </head> <body leftMargin="0" topMargin="0"> <form id="TheForm" runat="server"> <input type="hidden" name="key" id="key" value="" /> <input type="hidden" name="formactionurl" id="formactionurl" value="" /> <input type="hidden" name="handle" id="handle" value="" /> <asp:ScriptManager ID="ScriptManager1" runat="server" EnablePartialRendering="true" > </asp:ScriptManager> This approach did not work for all existing Master Pages, but fortunately all of the other existing Master Pages I have run across worked fine as a sub-Master to the WSRP Master Page. Moving On In Enterprise Portals, even after you get everything working, the work is not finished. Next you need to get it where everyone will work with it. Migration Planning Providing that the server where IIS is running is adequately sized, it is possible to run both the .NET Accelerator and the WSRP Producer on the same server during the upgrade process. The upgrade can be performed incrementally, i.e., one portlet at a time, if server administration processes support it. Those processes would include the ability to manage a second producer in the consuming portal and to change over individual portlet instances from one provider to the other. If processes or requirements demand that all portlets be cut over at the same time, it needs to be determined if this cut over should include a new producer, updating all of the portlets in the consumer, or if the WSRP Producer portlet configuration must maintain the naming conventions used by the .NET Accelerator and simply change the WSRP end point configured in the consumer. In some enterprises it may even be necessary to maintain the same WSDL end point, at which point the IIS configuration will be where the updates occur. The downside to such a requirement is that it makes rolling back very difficult, should the need arise. Location, Location, Location Not everyone wants the web application to have the descriptively obvious wsrpdefault location, or needs to create a second WSRP site on the same server. The instructions below are from the product team and, while targeted towards making a second site, will work for creating a site with a different name and then remove the old site. You can also change just the name in IIS. Manually Creating a WSRP Producer Site Instructions (NOTE: all executables used are the same ones used by the installer and “wsrpdev” will be the name of the new instance): 1. Copy C:\Oracle\Middleware\WSRPProducerForDotNet\wsrpdefault to C:\Oracle\Middleware\WSRPProducerForDotNet\wsrpdev. 2. Bring up a command window as an administrator 3. Run C:\Oracle\Middleware\WSRPProducerForDotNet\uninstall_resources\IISAppAccelSiteCreator.exe install WSRPProducers wsrpdev "C:\Oracle\Middleware\WSRPProducerForDotNet\wsrpdev" 8678 2.0.50727 4. Run C:\Oracle\Middleware\WSRPProducerForDotNet\uninstall_resources\PermManage.exe add FileSystem C:\Oracle\Middleware\WSRPProducerForDotNet\wsrpdev "NETWORK SERVICE" 3 1 5. Run C:\Oracle\Middleware\WSRPProducerForDotNet\uninstall_resources\PermManage.exe add FileSystem C:\Oracle\Middleware\WSRPProducerForDotNet\wsrpdev EVERYONE 1 1 6. Open up C:\Oracle\Middleware\WSRPProducerForDotNet\wsdl\1.0\WSRPService.wsdl and replace wsrpdefault with wsrpdev 7. Open up C:\Oracle\Middleware\WSRPProducerForDotNet\wsdl\2.0\WSRPService.wsdl and replace wsrpdefault with wsrpdev Tests: 1. Bring up a browser on the host itself and go to http://localhost:8678/wsrpdev/wsdl/1.0/WSRPService.wsdl and make sure that the URLs in the XML returned include the wsrpdev changes you made in step 6. 2. Bring up a browser on the host itself and see if the default sample comes up: http://localhost:8678/wsrpdev/portlets/ASPNET_AJAX_sample/default.aspx 3. Register the producer in WLP and test the portlet. Changing the Port used by WSRP Producer The pre-configured port for the WSRP Producer is 8678. You can change this port by updating both the IIS configuration and C:\Oracle\Middleware\WSRPProducerForDotNet\[WSRP_APP_NAME]\wsdl\1.0\WSRPService.wsdl. Do You Need to Migrate? Oracle Premier Support ended in November of 2010 for AquaLogic Interaction .NET Application Accelerator 1.x and Extended Support ends in November 2012 (see http://www.oracle.com/us/support/lifetime-support/lifetime-support-software-342730.html for other related dates). This means that integration with products released after November of 2010 is not supported. If having such support is the policy within your enterprise, you do indeed need to migrate. If changes in your enterprise cause your current solution with the .NET Accelerator to no longer function properly, you may need to migrate. Migration is a choice, and if the goals of your enterprise are to take full advantage of newer technologies then migration is certainly one activity you should be planning for.

    Read the article

  • Backup Meta-Data

    - by BuckWoody
    I'm working on a PowerShell script to show me the trending durations of my backup activities. The first thing I need is the data, so I looked at the Standard Reports in SQL Server Management Studio, and found a report that suited my needs, so I pulled out the script that it runs and modified it to this T-SQL Script. A few words here - you need to be in the MSDB database for this to run, and you can add a WHERE clause to limit to a database, timeframe, type of backup, whatever. For that matter, I won't use all of the data in this query in my PowerShell script, but it gives me lots of avenues to graph: SELECT distinct t1.name AS 'DatabaseName' ,(datediff( ss,  t3.backup_start_date, t3.backup_finish_date)) AS 'DurationInSeconds' ,t3.user_name AS 'UserResponsible' ,t3.name AS backup_name ,t3.description ,t3.backup_start_date ,t3.backup_finish_date ,CASE WHEN t3.type = 'D' THEN 'Database' WHEN t3.type = 'L' THEN 'Log' WHEN t3.type = 'F' THEN 'FileOrFilegroup' WHEN t3.type = 'G' THEN 'DifferentialFile' WHEN t3.type = 'P' THEN 'Partial' WHEN t3.type = 'Q' THEN 'DifferentialPartial' END AS 'BackupType' ,t3.backup_size AS 'BackupSizeKB' ,t6.physical_device_name ,CASE WHEN t6.device_type = 2 THEN 'Disk' WHEN t6.device_type = 102 THEN 'Disk' WHEN t6.device_type = 5 THEN 'Tape' WHEN t6.device_type = 105 THEN 'Tape' END AS 'DeviceType' ,t3.recovery_model  FROM sys.databases t1 INNER JOIN backupset t3 ON (t3.database_name = t1.name )  LEFT OUTER JOIN backupmediaset t5 ON ( t3.media_set_id = t5.media_set_id ) LEFT OUTER JOIN backupmediafamily t6 ON ( t6.media_set_id = t5.media_set_id ) ORDER BY backup_start_date DESC I'll munge this into my Excel PowerShell chart script tomorrow. Script Disclaimer, for people who need to be told this sort of thing: Never trust any script, including those that you find here, until you understand exactly what it does and how it will act on your systems. Always check the script on a test system or Virtual Machine, not a production system. Yes, there are always multiple ways to do things, and this script may not work in every situation, for everything. It’s just a script, people. All scripts on this site are performed by a professional stunt driver on a closed course. Your mileage may vary. Void where prohibited. Offer good for a limited time only. Keep out of reach of small children. Do not operate heavy machinery while using this script. If you experience blurry vision, indigestion or diarrhea during the operation of this script, see a physician immediately. Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • Backup Meta-Data

    - by BuckWoody
    I'm working on a PowerShell script to show me the trending durations of my backup activities. The first thing I need is the data, so I looked at the Standard Reports in SQL Server Management Studio, and found a report that suited my needs, so I pulled out the script that it runs and modified it to this T-SQL Script. A few words here - you need to be in the MSDB database for this to run, and you can add a WHERE clause to limit to a database, timeframe, type of backup, whatever. For that matter, I won't use all of the data in this query in my PowerShell script, but it gives me lots of avenues to graph: SELECT distinct t1.name AS 'DatabaseName' ,(datediff( ss,  t3.backup_start_date, t3.backup_finish_date)) AS 'DurationInSeconds' ,t3.user_name AS 'UserResponsible' ,t3.name AS backup_name ,t3.description ,t3.backup_start_date ,t3.backup_finish_date ,CASE WHEN t3.type = 'D' THEN 'Database' WHEN t3.type = 'L' THEN 'Log' WHEN t3.type = 'F' THEN 'FileOrFilegroup' WHEN t3.type = 'G' THEN 'DifferentialFile' WHEN t3.type = 'P' THEN 'Partial' WHEN t3.type = 'Q' THEN 'DifferentialPartial' END AS 'BackupType' ,t3.backup_size AS 'BackupSizeKB' ,t6.physical_device_name ,CASE WHEN t6.device_type = 2 THEN 'Disk' WHEN t6.device_type = 102 THEN 'Disk' WHEN t6.device_type = 5 THEN 'Tape' WHEN t6.device_type = 105 THEN 'Tape' END AS 'DeviceType' ,t3.recovery_model  FROM sys.databases t1 INNER JOIN backupset t3 ON (t3.database_name = t1.name )  LEFT OUTER JOIN backupmediaset t5 ON ( t3.media_set_id = t5.media_set_id ) LEFT OUTER JOIN backupmediafamily t6 ON ( t6.media_set_id = t5.media_set_id ) ORDER BY backup_start_date DESC I'll munge this into my Excel PowerShell chart script tomorrow. Script Disclaimer, for people who need to be told this sort of thing: Never trust any script, including those that you find here, until you understand exactly what it does and how it will act on your systems. Always check the script on a test system or Virtual Machine, not a production system. Yes, there are always multiple ways to do things, and this script may not work in every situation, for everything. It’s just a script, people. All scripts on this site are performed by a professional stunt driver on a closed course. Your mileage may vary. Void where prohibited. Offer good for a limited time only. Keep out of reach of small children. Do not operate heavy machinery while using this script. If you experience blurry vision, indigestion or diarrhea during the operation of this script, see a physician immediately. Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • ASP.NET MVC 2 Model Binding for a Collection

    - by nmarun
    Yes, my yet another post on Model Binding (previous one is here), but this one uses features presented in MVC 2. How I got to writing this blog? Well, I’m on a project where we’re doing some MVC things for a shopping cart. Let me show you what I was working with. Below are my model classes: 1: public class Product 2: { 3: public int Id { get; set; } 4: public string Name { get; set; } 5: public int Quantity { get; set; } 6: public decimal UnitPrice { get; set; } 7: } 8:   9: public class Totals 10: { 11: public decimal SubTotal { get; set; } 12: public decimal Tax { get; set; } 13: public decimal Total { get; set; } 14: } 15:   16: public class Basket 17: { 18: public List<Product> Products { get; set; } 19: public Totals Totals { get; set;} 20: } The view looks as below:  1: <h2>Shopping Cart</h2> 2:   3: <% using(Html.BeginForm()) { %> 4: 5: <h3>Products</h3> 6: <% for (int i = 0; i < Model.Products.Count; i++) 7: { %> 8: <div style="width: 100px;float:left;">Id</div> 9: <div style="width: 100px;float:left;"> 10: <%= Html.TextBox("ID", Model.Products[i].Id) %> 11: </div> 12: <div style="clear:both;"></div> 13: <div style="width: 100px;float:left;">Name</div> 14: <div style="width: 100px;float:left;"> 15: <%= Html.TextBox("Name", Model.Products[i].Name) %> 16: </div> 17: <div style="clear:both;"></div> 18: <div style="width: 100px;float:left;">Quantity</div> 19: <div style="width: 100px;float:left;"> 20: <%= Html.TextBox("Quantity", Model.Products[i].Quantity)%> 21: </div> 22: <div style="clear:both;"></div> 23: <div style="width: 100px;float:left;">Unit Price</div> 24: <div style="width: 100px;float:left;"> 25: <%= Html.TextBox("UnitPrice", Model.Products[i].UnitPrice)%> 26: </div> 27: <div style="clear:both;"><hr /></div> 28: <% } %> 29: 30: <h3>Totals</h3> 31: <div style="width: 100px;float:left;">Sub Total</div> 32: <div style="width: 100px;float:left;"> 33: <%= Html.TextBox("SubTotal", Model.Totals.SubTotal)%> 34: </div> 35: <div style="clear:both;"></div> 36: <div style="width: 100px;float:left;">Tax</div> 37: <div style="width: 100px;float:left;"> 38: <%= Html.TextBox("Tax", Model.Totals.Tax)%> 39: </div> 40: <div style="clear:both;"></div> 41: <div style="width: 100px;float:left;">Total</div> 42: <div style="width: 100px;float:left;"> 43: <%= Html.TextBox("Total", Model.Totals.Total)%> 44: </div> 45: <div style="clear:both;"></div> 46: <p /> 47: <input type="submit" name="Submit" value="Submit" /> 48: <% } %> .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } Nothing fancy, just a bunch of div’s containing textboxes and a submit button. Just make note that the textboxes have the same name as the property they are going to display. Yea, yea, I know. I’m displaying unit price as a textbox instead of a label, but that’s beside the point (and trust me, this will not be how it’ll look on the production site!!). The way my controller works is that initially two dummy products are added to the basked object and the Totals are calculated based on what products were added in what quantities and their respective unit price. So when the page loads in edit mode, where the user can change the quantity and hit the submit button. In the ‘post’ version of the action method, the Totals get recalculated and the new total will be displayed on the screen. Here’s the code: 1: public ActionResult Index() 2: { 3: Product product1 = new Product 4: { 5: Id = 1, 6: Name = "Product 1", 7: Quantity = 2, 8: UnitPrice = 200m 9: }; 10:   11: Product product2 = new Product 12: { 13: Id = 2, 14: Name = "Product 2", 15: Quantity = 1, 16: UnitPrice = 150m 17: }; 18:   19: List<Product> products = new List<Product> { product1, product2 }; 20:   21: Basket basket = new Basket 22: { 23: Products = products, 24: Totals = ComputeTotals(products) 25: }; 26: return View(basket); 27: } 28:   29: [HttpPost] 30: public ActionResult Index(Basket basket) 31: { 32: basket.Totals = ComputeTotals(basket.Products); 33: return View(basket); 34: } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } That’s that. Now I run the app, I see two products with the totals section below them. I look at the view source and I see that the input controls have the right ID, the right name and the right value as well. 1: <input id="ID" name="ID" type="text" value="1" /> 2: <input id="Name" name="Name" type="text" value="Product 1" /> 3: ... 4: <input id="ID" name="ID" type="text" value="2" /> 5: <input id="Name" name="Name" type="text" value="Product 2" /> .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } So just as a regular user would do, I change the quantity value of one of the products and hit the submit button. The ‘post’ version of the Index method gets called and I had put a break-point on line 32 in the above snippet. When I hovered my mouse on the ‘basked’ object, happily assuming that the object would be all bound and ready for use, I was surprised to see both basket.Products and basket.Totals were null. Huh? A little research and I found out that the reason the DefaultModelBinder could not do its job is because of a naming mismatch on the input controls. What I mean is that when you have to bind to a custom .net type, you need more than just the property name. You need to pass a qualified name to the name property of the input control. I modified my view and the emitted code looked as below: 1: <input id="Product_Name" name="Product.Name" type="text" value="Product 1" /> 2: ... 3: <input id="Product_Name" name="Product.Name" type="text" value="Product 2" /> 4: ... 5: <input id="Totals_SubTotal" name="Totals.SubTotal" type="text" value="550" /> .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } Now, I update the quantity and hit the submit button and I see that the Totals object is populated, but the Products list is still null. Once again I went: ‘Hmm.. time for more research’. I found out that the way to do this is to provide the name as: 1: <%= Html.TextBox(string.Format("Products[{0}].ID", i), Model.Products[i].Id) %> 2: <!-- this will be rendered as --> 3: <input id="Products_0__ID" name="Products[0].ID" type="text" value="1" /> .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } It was only now that I was able to see both the products and the totals being properly bound in the ‘post’ action method. Somehow, I feel this is kinda ‘clunky’ way of doing things. Seems like people at MS felt in a similar way and offered us a much cleaner way to solve this issue. The simple solution is that instead of using a Textbox, we can either use a TextboxFor or an EditorFor helper method. This one directly spits out the name of the input property as ‘Products[0].ID and so on. Cool right? I totally fell for this and changed my UI to contain EditorFor helper method. At this point, I ran the application, changed the quantity field and pressed the submit button. Of course my basket object parameter in my action method was correctly bound after these changes. I let the app complete the rest of the lines in the action method. When the page finally rendered, I did see that the quantity was changed to what I entered before the post. But, wait a minute, the totals section did not reflect the changes and showed the old values. My status: COMPLETELY PUZZLED! Just to recap, this is what my ‘post’ Index method looked like: 1: [HttpPost] 2: public ActionResult Index(Basket basket) 3: { 4: basket.Totals = ComputeTotals(basket.Products); 5: return View(basket); 6: } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } A careful debug confirmed that the basked.Products[0].Quantity showed the updated value and the ComputeTotals() method also returns the correct totals. But still when I passed this basket object, it ended up showing the old totals values only. I began playing a bit with the code and my first guess was that the input controls got their values from the ModelState object. For those who don’t know, the ModelState is a temporary storage area that ASP.NET MVC uses to retain incoming attempted values plus binding and validation errors. Also, the fact that input controls populate the values using data taken from: Previously attempted values recorded in the ModelState["name"].Value.AttemptedValue Explicitly provided value (<%= Html.TextBox("name", "Some value") %>) ViewData, by calling ViewData.Eval("name") FYI: ViewData dictionary takes precedence over ViewData's Model properties – read more here. These two indicators led to my guess. It took me quite some time, but finally I hit this post where Brad brilliantly explains why this is the preferred behavior. My guess was right and I, accordingly modified my code to reflect the following way: 1: [HttpPost] 2: public ActionResult Index(Basket basket) 3: { 4: // read the following posts to see why the ModelState 5: // needs to be cleared before passing it the view 6: // http://forums.asp.net/t/1535846.aspx 7: // http://forums.asp.net/p/1527149/3687407.aspx 8: if (ModelState.IsValid) 9: { 10: ModelState.Clear(); 11: } 12:   13: basket.Totals = ComputeTotals(basket.Products); 14: return View(basket); 15: } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } What this does is that in the case where your ModelState IS valid, it clears the dictionary. This enables the values to be read from the model directly and not from the ModelState. So the verdict is this: If you need to pass other parameters (like html attributes and the like) to your input control, use 1: <%= Html.TextBox(string.Format("Products[{0}].ID", i), Model.Products[i].Id) %> .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } Since, in EditorFor, there is no direct and simple way of passing this information to the input control. If you don’t have to pass any such ‘extra’ piece of information to the control, then go the EditorFor way. The code used in the post can be found here.

    Read the article

  • BAM design pointers

    - by Kavitha Srinivasan
    In working recently with a large Oracle customer on SOA and BAM, I discovered that some BAM best practices are not quite well known as I had always assumed ! There is a doc bug out to formally incorporate those learnings but here are a few notes..  EMS-DO parity When using EMS (Enterprise Message Source) as a BAM feed, the best practice is to use one EMS to write to one Data Object. There is a possibility of collisions and duplicates when multiple EMS write to the same row of a DO at the same time. This customer had 17 EMS writing to one DO at the same time. Every sensor in their BPEL process writes to one topic but the Topic was read by 1 EMS corresponding to one sensor. They then used XSL within BAM to transform the payload into the BAM DO format. And hence for a given BPEL instance, 17 sensors fired, populated 1 JMS topic, was consumed by 17 EMS which in turn wrote to 1 DataObject.(You can image what would happen for later versions of the application that needs to send more information to BAM !).  We modified their design to use one Master XSL based on sensorname for all sensors relating to a DO- say Data Object 'Orders' and were able to thus reduce the 17 EMS to 1 with a master XSL. For those of you wondering about how squeaky clean this design is, you are right ! This is indeed not squeaky clean and that brings us to yet another 'inferred' best practice. (I try very hard not to state the obvious in my blogs with the hope that everytime I blog, it is very useful but this one is an exception.) Transformations and Calculations It is optimal to do transformations within an engine like BPEL. Not only does this provide modelling ease with a nice GUI XSL mapper in JDeveloper, the XSL engine in BPEL is quite efficient at runtime as well. And so, doing XSL transformations in BAM is not quite prudent.  The same is true for any non-trivial calculations as well. It is best to do all transformations,calcuations and sanitize the data in a BPEL or like layer and then send this to BAM (via JMS, WS etc.) This then delegates simply the function of report rendering and mechanics of real-time reporting to the Oracle BAM reporting tool which it is most suited to do. All nulls are not created equal Here is yet another possibly known fact but reiterated here. For an EMS with an Upsert operation: a) If Empty tags or tags with no value are sent like <Tag1/> or <Tag1></Tag1>, the DO will be overwritten with --null-- b) If Empty tags are suppressed ie not generated at all, the corresponding DO field will NOT be overwritten. The field will have whatever value existed previously.  For an EMS with an Insert operation, both tags with an empty value and no tags result in –null-- being written to the DO. Hope this helps .. Happy 4th!

    Read the article

  • SQL SERVER – Inviting Ideas for SQL in Sixty Seconds – 12/12/12

    - by pinaldave
    Today is 12/12/12 – I am not sure when will I write this kind of date again – maybe never. This opportunity comes once in a lifetime when we have the same date, month and year all have same digit. December 12th is one of the most fantastic day in my personal life. Four years ago, this day I got married to my wife – Nupur Dave.  Here are photos of our wedding (Dec 12, 2008). Here is a very interesting photo of myself earlier this year. It is not photoshoped or modified photo. The only modification I have done here is to add arrow and speech bubble. Every Wednesday I tried to put one SQL in Sixty Seconds Video. The journey has been fantastic and so far I have put a total of 35 SQL in Sixty Seconds Video. The goal of the video is to learn something in 1 minute. In our daily life we are all very busy and hardly have time for anything. No matter how much we are busy – we all have one minute of time. Sometime we wait for a minute in elevators, at the escalator, at a coffee shop, or just waiting for our phone reboot. Today is a fantastic day – 12/12/12. Let me invite all of you submits SQL in Sixty Seconds idea. If I like your idea and create a sixty second video over it – you will win surprise learning material from me. There are two very simple rules of the contest: - I should have not have already recorded the tip. The tip should be descriptive. Do not just suggest to cover “Performance Tuning” or “How to Create Index” or “More of reporting services”. The tip should have around 100 words of description explaining SQL Tip. The contest is open forever. The winner will be announced whenever I use the tip to convert to video. If I use your tip, I will for sure mention in the blog post that it is inspired from your suggestion. Meanwhile, do not forget to subscribe YouTube Channel. Here are my latest three videos from SQL in Sixty Seconds. Reference: Pinal Dave (http://blog.sqlauthority.com) Filed under: About Me, PostADay, SQL, SQL Authority, SQL in Sixty Seconds, SQL Query, SQL Server, SQL Tips and Tricks, T SQL, Technology, Video

    Read the article

  • SQL SERVER – Sharing your ETL Resources Across Applications with Ease

    - by pinaldave
    Frequently an organization will find that the same resources are used in multiple ETL applications, for example, the same database, general purpose processing logic, or file system locations.  Creating an easy way to reuse these resources across multiple applications would increase efficiency and reduce errors.  Moreover, not every ETL developer has the same skill set, and it is likely that one developer will be more adept at writing code while another is more comfortable configuring database connections.  Real productivity gains will come when these developers are able to work independently while still making their work available to others assigned to the same project.  These are the benefits of a centralized version control system. Of course, most version control systems could be used to store and serve files, but the real need is to store and serve entire ETL applications so that each developer’s ongoing work can immediately benefit from another developer’s completed work.  In other words, the version control system needs to be tightly integrated with the tools used to develop the ETL application. The following screen shot shows such a tool. Desktop ETL tool that tightly integrates with a central version control system Developers can checkout or commit entire projects or just a single artifact.  Each artifact may be managed independently so if you need to go back to an earlier version of one artifact, changes you may have made to other artifacts are not lost.  By being tightly integrated into the graphical environment used to create and edit the project artifacts, it is extremely easy and straight-forward to move your files to and from the version control system and there is no dependency on another vendor’s version control system.  The built in version control system is optimized for managing the artifacts of ETL applications. It is equally important that the version control system supports all of the actions one typically performs such as rollbacks, locking and unlocking of files, and the ability to resolve conflicts.  Note that this particular ETL tool also has the capability to switch back and forth between multiple version control systems. It also needs to be easy to determine the status of an artifact.  Not just that it has been committed or modified, but when and by whom.  Generally you must query the version control system for this information, but having it displayed within the development environment is more desirable. Who’s ETL tool works in this fashion?  Last month I mentioned the data integration solution offered by expressor software.  The version control features I described in this post are all available in their just released expressor 3.1 Standard Edition through the integration of their expressor Studio development environment with a centralized metadata repository and version control system. You can download their Studio application, which is free, or evaluate the full Standard Edition on your own hardware.  It may be worth your time. Reference: Pinal Dave (http://blog.SQLAuthority.com) Filed under: Pinal Dave, PostADay, SQL, SQL Authority, SQL Query, SQL Server, SQL Tips and Tricks, T SQL, Technology

    Read the article

  • Much Ado About Nothing: Stub Objects

    - by user9154181
    The Solaris 11 link-editor (ld) contains support for a new type of object that we call a stub object. A stub object is a shared object, built entirely from mapfiles, that supplies the same linking interface as the real object, while containing no code or data. Stub objects cannot be executed — the runtime linker will kill any process that attempts to load one. However, you can link to a stub object as a dependency, allowing the stub to act as a proxy for the real version of the object. You may well wonder if there is a point to producing an object that contains nothing but linking interface. As it turns out, stub objects are very useful for building large bodies of code such as Solaris. In the last year, we've had considerable success in applying them to one of our oldest and thorniest build problems. In this discussion, I will describe how we came to invent these objects, and how we apply them to building Solaris. This posting explains where the idea for stub objects came from, and details our long and twisty journey from hallway idea to standard link-editor feature. I expect that these details are mainly of interest to those who work on Solaris and its makefiles, those who have done so in the past, and those who work with other similar bodies of code. A subsequent posting will omit the history and background details, and instead discuss how to build and use stub objects. If you are mainly interested in what stub objects are, and don't care about the underlying software war stories, I encourage you to skip ahead. The Long Road To Stubs This all started for me with an email discussion in May of 2008, regarding a change request that was filed in 2002, entitled: 4631488 lib/Makefile is too patient: .WAITs should be reduced This CR encapsulates a number of cronic issues with Solaris builds: We build Solaris with a parallel make (dmake) that tries to build as much of the code base in parallel as possible. There is a lot of code to build, and we've long made use of parallelized builds to get the job done quicker. This is even more important in today's world of massively multicore hardware. Solaris contains a large number of executables and shared objects. Executables depend on shared objects, and shared objects can depend on each other. Before you can build an object, you need to ensure that the objects it needs have been built. This implies a need for serialization, which is in direct opposition to the desire to build everying in parallel. To accurately build objects in the right order requires an accurate set of make rules defining the things that depend on each other. This sounds simple, but the reality is quite complex. In practice, having programmers explicitly specify these dependencies is a losing strategy: It's really hard to get right. It's really easy to get it wrong and never know it because things build anyway. Even if you get it right, it won't stay that way, because dependencies between objects can change over time, and make cannot help you detect such drifing. You won't know that you got it wrong until the builds break. That can be a long time after the change that triggered the breakage happened, making it hard to connect the cause and the effect. Usually this happens just before a release, when the pressure is on, its hard to think calmly, and there is no time for deep fixes. As a poor compromise, the libraries in core Solaris were built using a set of grossly incomplete hand written rules, supplemented with a number of dmake .WAIT directives used to group the libraries into sets of non-interacting groups that can be built in parallel because we think they don't depend on each other. From time to time, someone will suggest that we could analyze the built objects themselves to determine their dependencies and then generate make rules based on those relationships. This is possible, but but there are complications that limit the usefulness of that approach: To analyze an object, you have to build it first. This is a classic chicken and egg scenario. You could analyze the results of a previous build, but then you're not necessarily going to get accurate rules for the current code. It should be possible to build the code without having a built workspace available. The analysis will take time, and remember that we're constantly trying to make builds faster, not slower. By definition, such an approach will always be approximate, and therefore only incremantally more accurate than the hand written rules described above. The hand written rules are fast and cheap, while this idea is slow and complex, so we stayed with the hand written approach. Solaris was built that way, essentially forever, because these are genuinely difficult problems that had no easy answer. The makefiles were full of build races in which the right outcomes happened reliably for years until a new machine or a change in build server workload upset the accidental balance of things. After figuring out what had happened, you'd mutter "How did that ever work?", add another incomplete and soon to be inaccurate make dependency rule to the system, and move on. This was not a satisfying solution, as we tend to be perfectionists in the Solaris group, but we didn't have a better answer. It worked well enough, approximately. And so it went for years. We needed a different approach — a new idea to cut the Gordian Knot. In that discussion from May 2008, my fellow linker-alien Rod Evans had the initial spark that lead us to a game changing series of realizations: The link-editor is used to link objects together, but it only uses the ELF metadata in the object, consisting of symbol tables, ELF versioning sections, and similar data. Notably, it does not look at, or understand, the machine code that makes an object useful at runtime. If you had an object that only contained the ELF metadata for a dependency, but not the code or data, the link-editor would find it equally useful for linking, and would never know the difference. Call it a stub object. In the core Solaris OS, we require all objects to be built with a link-editor mapfile that describes all of its publically available functions and data. Could we build a stub object using the mapfile for the real object? It ought to be very fast to build stub objects, as there are no input objects to process. Unlike the real object, stub objects would not actually require any dependencies, and so, all of the stubs for the entire system could be built in parallel. When building the real objects, one could link against the stub objects instead of the real dependencies. This means that all the real objects can be built built in parallel too, without any serialization. We could replace a system that requires perfect makefile rules with a system that requires no ordering rules whatsoever. The results would be considerably more robust. We immediately realized that this idea had potential, but also that there were many details to sort out, lots of work to do, and that perhaps it wouldn't really pan out. As is often the case, it would be necessary to do the work and see how it turned out. Following that conversation, I set about trying to build a stub object. We determined that a faithful stub has to do the following: Present the same set of global symbols, with the same ELF versioning, as the real object. Functions are simple — it suffices to have a symbol of the right type, possibly, but not necessarily, referencing a null function in its text segment. Copy relocations make data more complicated to stub. The possibility of a copy relocation means that when you create a stub, the data symbols must have the actual size of the real data. Any error in this will go uncaught at link time, and will cause tragic failures at runtime that are very hard to diagnose. For reasons too obscure to go into here, involving tentative symbols, it is also important that the data reside in bss, or not, matching its placement in the real object. If the real object has more than one symbol pointing at the same data item, we call these aliased symbols. All data symbols in the stub object must exhibit the same aliasing as the real object. We imagined the stub library feature working as follows: A command line option to ld tells it to produce a stub rather than a real object. In this mode, only mapfiles are examined, and any object or shared libraries on the command line are are ignored. The extra information needed (function or data, size, and bss details) would be added to the mapfile. When building the real object instead of the stub, the extra information for building stubs would be validated against the resulting object to ensure that they match. In exploring these ideas, I immediately run headfirst into the reality of the original mapfile syntax, a subject that I would later write about as The Problem(s) With Solaris SVR4 Link-Editor Mapfiles. The idea of extending that poor language was a non-starter. Until a better mapfile syntax became available, which seemed unlikely in 2008, the solution could not involve extentions to the mapfile syntax. Instead, we cooked up the idea (hack) of augmenting mapfiles with stylized comments that would carry the necessary information. A typical definition might look like: # DATA(i386) __iob 0x3c0 # DATA(amd64,sparcv9) __iob 0xa00 # DATA(sparc) __iob 0x140 iob; A further problem then became clear: If we can't extend the mapfile syntax, then there's no good way to extend ld with an option to produce stub objects, and to validate them against the real objects. The idea of having ld read comments in a mapfile and parse them for content is an unacceptable hack. The entire point of comments is that they are strictly for the human reader, and explicitly ignored by the tool. Taking all of these speed bumps into account, I made a new plan: A perl script reads the mapfiles, generates some small C glue code to produce empty functions and data definitions, compiles and links the stub object from the generated glue code, and then deletes the generated glue code. Another perl script used after both objects have been built, to compare the real and stub objects, using data from elfdump, and validate that they present the same linking interface. By June 2008, I had written the above, and generated a stub object for libc. It was a useful prototype process to go through, and it allowed me to explore the ideas at a deep level. Ultimately though, the result was unsatisfactory as a basis for real product. There were so many issues: The use of stylized comments were fine for a prototype, but not close to professional enough for shipping product. The idea of having to document and support it was a large concern. The ideal solution for stub objects really does involve having the link-editor accept the same arguments used to build the real object, augmented with a single extra command line option. Any other solution, such as our prototype script, will require makefiles to be modified in deeper ways to support building stubs, and so, will raise barriers to converting existing code. A validation script that rederives what the linker knew when it built an object will always be at a disadvantage relative to the actual linker that did the work. A stub object should be identifyable as such. In the prototype, there was no tag or other metadata that would let you know that they weren't real objects. Being able to identify a stub object in this way means that the file command can tell you what it is, and that the runtime linker can refuse to try and run a program that loads one. At that point, we needed to apply this prototype to building Solaris. As you might imagine, the task of modifying all the makefiles in the core Solaris code base in order to do this is a massive task, and not something you'd enter into lightly. The quality of the prototype just wasn't good enough to justify that sort of time commitment, so I tabled the project, putting it on my list of long term things to think about, and moved on to other work. It would sit there for a couple of years. Semi-coincidentally, one of the projects I tacked after that was to create a new mapfile syntax for the Solaris link-editor. We had wanted to do something about the old mapfile syntax for many years. Others before me had done some paper designs, and a great deal of thought had already gone into the features it should, and should not have, but for various reasons things had never moved beyond the idea stage. When I joined Sun in late 2005, I got involved in reviewing those things and thinking about the problem. Now in 2008, fresh from relearning for the Nth time why the old mapfile syntax was a huge impediment to linker progress, it seemed like the right time to tackle the mapfile issue. Paving the way for proper stub object support was not the driving force behind that effort, but I certainly had them in mind as I moved forward. The new mapfile syntax, which we call version 2, integrated into Nevada build snv_135 in in February 2010: 6916788 ld version 2 mapfile syntax PSARC/2009/688 Human readable and extensible ld mapfile syntax In order to prove that the new mapfile syntax was adequate for general purpose use, I had also done an overhaul of the ON consolidation to convert all mapfiles to use the new syntax, and put checks in place that would ensure that no use of the old syntax would creep back in. That work went back into snv_144 in June 2010: 6916796 OSnet mapfiles should use version 2 link-editor syntax That was a big putback, modifying 517 files, adding 18 new files, and removing 110 old ones. I would have done this putback anyway, as the work was already done, and the benefits of human readable syntax are obvious. However, among the justifications listed in CR 6916796 was this We anticipate adding additional features to the new mapfile language that will be applicable to ON, and which will require all sharable object mapfiles to use the new syntax. I never explained what those additional features were, and no one asked. It was premature to say so, but this was a reference to stub objects. By that point, I had already put together a working prototype link-editor with the necessary support for stub objects. I was pleased to find that building stubs was indeed very fast. On my desktop system (Ultra 24), an amd64 stub for libc can can be built in a fraction of a second: % ptime ld -64 -z stub -o stubs/libc.so.1 -G -hlibc.so.1 \ -ztext -zdefs -Bdirect ... real 0.019708910 user 0.010101680 sys 0.008528431 In order to go from prototype to integrated link-editor feature, I knew that I would need to prove that stub objects were valuable. And to do that, I knew that I'd have to switch the Solaris ON consolidation to use stub objects and evaluate the outcome. And in order to do that experiment, ON would first need to be converted to version 2 mapfiles. Sub-mission accomplished. Normally when you design a new feature, you can devise reasonably small tests to show it works, and then deploy it incrementally, letting it prove its value as it goes. The entire point of stub objects however was to demonstrate that they could be successfully applied to an extremely large and complex code base, and specifically to solve the Solaris build issues detailed above. There was no way to finesse the matter — in order to move ahead, I would have to successfully use stub objects to build the entire ON consolidation and demonstrate their value. In software, the need to boil the ocean can often be a warning sign that things are trending in the wrong direction. Conversely, sometimes progress demands that you build something large and new all at once. A big win, or a big loss — sometimes all you can do is try it and see what happens. And so, I spent some time staring at ON makefiles trying to get a handle on how things work, and how they'd have to change. It's a big and messy world, full of complex interactions, unspecified dependencies, special cases, and knowledge of arcane makefile features... ...and so, I backed away, put it down for a few months and did other work... ...until the fall, when I felt like it was time to stop thinking and pondering (some would say stalling) and get on with it. Without stubs, the following gives a simplified high level view of how Solaris is built: An initially empty directory known as the proto, and referenced via the ROOT makefile macro is established to receive the files that make up the Solaris distribution. A top level setup rule creates the proto area, and performs operations needed to initialize the workspace so that the main build operations can be launched, such as copying needed header files into the proto area. Parallel builds are launched to build the kernel (usr/src/uts), libraries (usr/src/lib), and commands. The install makefile target builds each item and delivers a copy to the proto area. All libraries and executables link against the objects previously installed in the proto, implying the need to synchronize the order in which things are built. Subsequent passes run lint, and do packaging. Given this structure, the additions to use stub objects are: A new second proto area is established, known as the stub proto and referenced via the STUBROOT makefile macro. The stub proto has the same structure as the real proto, but is used to hold stub objects. All files in the real proto are delivered as part of the Solaris product. In contrast, the stub proto is used to build the product, and then thrown away. A new target is added to library Makefiles called stub. This rule builds the stub objects. The ld command is designed so that you can build a stub object using the same ld command line you'd use to build the real object, with the addition of a single -z stub option. This means that the makefile rules for building the stub objects are very similar to those used to build the real objects, and many existing makefile definitions can be shared between them. A new target is added to the Makefiles called stubinstall which delivers the stub objects built by the stub rule into the stub proto. These rules reuse much of existing plumbing used by the existing install rule. The setup rule runs stubinstall over the entire lib subtree as part of its initialization. All libraries and executables link against the objects in the stub proto rather than the main proto, and can therefore be built in parallel without any synchronization. There was no small way to try this that would yield meaningful results. I would have to take a leap of faith and edit approximately 1850 makefiles and 300 mapfiles first, trusting that it would all work out. Once the editing was done, I'd type make and see what happened. This took about 6 weeks to do, and there were many dark days when I'd question the entire project, or struggle to understand some of the many twisted and complex situations I'd uncover in the makefiles. I even found a couple of new issues that required changes to the new stub object related code I'd added to ld. With a substantial amount of encouragement and help from some key people in the Solaris group, I eventually got the editing done and stub objects for the entire workspace built. I found that my desktop system could build all the stub objects in the workspace in roughly a minute. This was great news, as it meant that use of the feature is effectively free — no one was likely to notice or care about the cost of building them. After another week of typing make, fixing whatever failed, and doing it again, I succeeded in getting a complete build! The next step was to remove all of the make rules and .WAIT statements dedicated to controlling the order in which libraries under usr/src/lib are built. This came together pretty quickly, and after a few more speed bumps, I had a workspace that built cleanly and looked like something you might actually be able to integrate someday. This was a significant milestone, but there was still much left to do. I turned to doing full nightly builds. Every type of build (open, closed, OpenSolaris, export, domestic) had to be tried. Each type failed in a new and unique way, requiring some thinking and rework. As things came together, I became aware of things that could have been done better, simpler, or cleaner, and those things also required some rethinking, the seeking of wisdom from others, and some rework. After another couple of weeks, it was in close to final form. My focus turned towards the end game and integration. This was a huge workspace, and needed to go back soon, before changes in the gate would made merging increasingly difficult. At this point, I knew that the stub objects had greatly simplified the makefile logic and uncovered a number of race conditions, some of which had been there for years. I assumed that the builds were faster too, so I did some builds intended to quantify the speedup in build time that resulted from this approach. It had never occurred to me that there might not be one. And so, I was very surprised to find that the wall clock build times for a stock ON workspace were essentially identical to the times for my stub library enabled version! This is why it is important to always measure, and not just to assume. One can tell from first principles, based on all those removed dependency rules in the library makefile, that the stub object version of ON gives dmake considerably more opportunities to overlap library construction. Some hypothesis were proposed, and shot down: Could we have disabled dmakes parallel feature? No, a quick check showed things being build in parallel. It was suggested that we might be I/O bound, and so, the threads would be mostly idle. That's a plausible explanation, but system stats didn't really support it. Plus, the timing between the stub and non-stub cases were just too suspiciously identical. Are our machines already handling as much parallelism as they are capable of, and unable to exploit these additional opportunities? Once again, we didn't see the evidence to back this up. Eventually, a more plausible and obvious reason emerged: We build the libraries and commands (usr/src/lib, usr/src/cmd) in parallel with the kernel (usr/src/uts). The kernel is the long leg in that race, and so, wall clock measurements of build time are essentially showing how long it takes to build uts. Although it would have been nice to post a huge speedup immediately, we can take solace in knowing that stub objects simplify the makefiles and reduce the possibility of race conditions. The next step in reducing build time should be to find ways to reduce or overlap the uts part of the builds. When that leg of the build becomes shorter, then the increased parallelism in the libs and commands will pay additional dividends. Until then, we'll just have to settle for simpler and more robust. And so, I integrated the link-editor support for creating stub objects into snv_153 (November 2010) with 6993877 ld should produce stub objects PSARC/2010/397 ELF Stub Objects followed by the work to convert the ON consolidation in snv_161 (February 2011) with 7009826 OSnet should use stub objects 4631488 lib/Makefile is too patient: .WAITs should be reduced This was a huge putback, with 2108 modified files, 8 new files, and 2 removed files. Due to the size, I was allowed a window after snv_160 closed in which to do the putback. It went pretty smoothly for something this big, a few more preexisting race conditions would be discovered and addressed over the next few weeks, and things have been quiet since then. Conclusions and Looking Forward Solaris has been built with stub objects since February. The fact that developers no longer specify the order in which libraries are built has been a big success, and we've eliminated an entire class of build error. That's not to say that there are no build races left in the ON makefiles, but we've taken a substantial bite out of the problem while generally simplifying and improving things. The introduction of a stub proto area has also opened some interesting new possibilities for other build improvements. As this article has become quite long, and as those uses do not involve stub objects, I will defer that discussion to a future article.

    Read the article

  • WebCenter Spaces 11g PS2 Template Customization

    - by javier.ductor(at)oracle.com
    Recently, we have been involved in a WebCenter Spaces customization project. Customer sent us a prototype website in HTML, and we had to transform Spaces to set the same look&feel as in the prototype.Protoype:First of all, we downloaded a Virtual Machine with WebCenter Spaces 11g PS2, same version as customer's. The next step was to download ExtendWebCenterSpaces application. This is a webcenter application that customizes several elements of WebCenter Spaces: templates, skins, landing page, etc. jDeveloper configuration is needed, we followed steps described in Extended Guide, this blog was helpful too. . After that, we deployed the application using WebLogic console, we created a new group space and assigned the ExtendedWebCenterSpaces template (portalCentricSiteTemplate) to it. The result was this:As you may see there is a big difference with the prototype, which meant a lot of work to do from our side.So we needed to create a new Spaces template with its skin.Using jDeveloper, we created a new template based on the default template. We removed all menus we did not need and inserted 'include'  tags for header, breadcrumb and footers. Each of these elements was defined in an isolated jspx file.In the beginning, we faced a problem: we had to add code from prototype (in plain HTML) to jspx templates (JSF language). We managed to workaround this issue using 'verbatim' tags with 'CDATA' surrounding HTML code in header, breadcrumb and footers.Once the template was modified, we added css styles to the default skin. So we had some styles from portalCentricSiteTemplate plus styles from customer's prototype.Then, we re-deployed the application, assigned this template to a new group space and checked the result. After testing, we usually found issues, so we had to do some modifications in the application, then it was necessary to re-deploy the application and restart Spaces server. Due to this fact, the testing process takes quite a bit of time.Added to the template and skin customization, we also customized the Landing Page using this application and Members task flow using another application: SpacesTaskflowCustomizationApplication. I will talk about this task flow customization in a future entry.After some issues and workarounds, eventually we managed to get this look&feel in Spaces:P.S. In this customization I was working with Francisco Vega and Jose Antonio Labrada, consultants from Oracle Malaga International Consulting Centre.

    Read the article

  • Put Unused Tabs on Hold in Firefox

    - by Asian Angel
    If you have a lot of content heavy webpages open in Firefox, it soon adds up on memory usage. The BarTab extension puts unused tabs on hold and keeps them unloaded until you are ready to access them. Before When you have a few tabs open things are not so bad. But if you have something like 30+ tabs open then Firefox is definitely going to have a much larger memory footprint. For the moment all six tabs are active in our example… After Once the extension is installed you can see quite a difference in the tab bar. Five of the six websites are literally on hold and display as about:blank. In the screenshot above all six tabs were restored from a previous session, but to put new tabs on hold the preferences will have to be modified. Notice that you can have active tabs placed on hold if you do not access them within a set time (“x” seconds, minutes, hours, or days). You may also add your favorite websites as exclusions. Newly opened tabs can now be placed on hold until you are ready for them. Need to place a currently accessed tab on hold? Use the tab context menu to unload the page. Once you use the tab context menu to place a tab on hold your browser will shift over to the closest active tab. You will notice the difference in memory usage when you put a lot of intensive pages on hold. Conclusion If you keep a high number of content heavy webpages open in Firefox then the BarTab extension will help reduce memory usage while browsing. Links Download the BarTab extension (Mozilla Add-ons) Similar Articles Productive Geek Tips Quick Tip: Use Tab Characters in Textarea Boxes in FirefoxYou Really Want to Completely Disable Tabs in Firefox?Quick Hits: 11 Firefox Tab How-TosQuick Tip: Save Windows and Tabs When Restarting FirefoxMake Firefox Use Multiple Rows of Tabs TouchFreeze Alternative in AutoHotkey The Icy Undertow Desktop Windows Home Server – Backup to LAN The Clear & Clean Desktop Use This Bookmarklet to Easily Get Albums Use AutoHotkey to Assign a Hotkey to a Specific Window Latest Software Reviews Tinyhacker Random Tips CloudBerry Online Backup 1.5 for Windows Home Server Snagit 10 VMware Workstation 7 Acronis Online Backup Gmail Button Addon (Firefox) Hyperwords addon (Firefox) Backup Outlook 2010 Daily Motivator (Firefox) FetchMp3 Can Download Videos & Convert Them to Mp3 Use Flixtime To Create Video Slideshows

    Read the article

  • SQL SERVER – Creating All New Database with Full Recovery Model

    - by pinaldave
    Sometimes, complex problems have very simple solutions. Let us see the following email which I received recently. “Hi Pinal, In our system when we create new database, by default, they are all created with the Simple Recovery Model. We have to manually change the recovery model after we create the database. We used the following simple T-SQL code: CREATE DATABASE dbname. We are very frustrated with this situation. We want all our databases to have the Full Recovery Model option by default. We are considering the following methods; please suggest the most efficient one among them. 1) Creating a Policy; when it is violated, the database model can be fixed 2) Triggers at Server Level 3) Automated Job which goes through all the databases and checks their recovery model; if the DBA has not changed the model, then the job will list the Databases and change their recovery model Also, we have a situation where we need a database in the Simple Recovery Model as well – how to white list them? Please suggest the best method.” Indeed, an interesting email! The answer to their question, i.e., which is the best method to fit their needs (white list, default, etc)? It will be NONE of the above. Here is the solution in one line and also the easiest way: Just go to your Model database: Path in SSMS >> Databases > System Databases >> model >> Right Click Properties >> Options >> Recovery Model - Select Full from dropdown. Every newly created database takes its base template from the Model Database. If you create a custom SP in the Model Database, when you create a new database, it will automatically exist in that database. Any database that was already created before making changes in the Model Database will not be affected at all. Creating Policy is also a good method, and I will blog about this in a separate blog post, but looking at current specifications of the reader, I think the Model Database should be modified to have a Full Recovery Option. While writing this blog post, I remembered my another blog post where the model database log file was growing drastically even though there were no transactions SQL SERVER – Log File Growing for Model Database – model Database Log File Grew Too Big. NOTE: Please do not touch the Model Database unnecessary. It is a strict “No.” If you want to create an object that you need in all the databases, then instead of creating it in model database, I suggest that you create a new database called maintenance and create the object there. Reference: Pinal Dave (http://blog.SQLAuthority.com) Filed under: PostADay, Readers Question, SQL, SQL Authority, SQL Query, SQL Server, SQL Tips and Tricks, T SQL, Technology

    Read the article

  • const vs. readonly for a singleton

    - by GlenH7
    First off, I understand there are folk who oppose the use of singletons. I think it's an appropriate use in this case as it's constant state information, but I'm open to differing opinions / solutions. (See The singleton pattern and When should the singleton pattern not be used?) Second, for a broader audience: C++/CLI has a similar keyword to readonly with initonly, so this isn't strictly a C# type question. (Literal field versus constant variable in C++/CLI) Sidenote: A discussion of some of the nuances on using const or readonly. My Question: I have a singleton that anchors together some different data structures. Part of what I expose through that singleton are some lists and other objects, which represent the necessary keys or columns in order to connect the linked data structures. I doubt that anyone would try to change these objects through a different module, but I want to explicitly protect them from that risk. So I'm currently using a "readonly" modifier on those objects*. I'm using readonly instead of const with the lists as I read that using const will embed those items in the referencing assemblies and will therefore trigger a rebuild of those referencing assemblies if / when the list(s) is/are modified. This seems like a tighter coupling than I would want between the modules, but I wonder if I'm obsessing over a moot point. (This is question #2 below) The alternative I see to using "readonly" is to make the variables private and then wrap them with a public get. I'm struggling to see the advantage of this approach as it seems like wrapper code that doesn't provide much additional benefit. (This is question #1 below) It's highly unlikely that we'll change the contents or format of the lists - they're a compilation of things to avoid using magic strings all over the place. Unfortunately, not all the code has converted over to using this singleton's presentation of those strings. Likewise, I don't know that we'd change the containers / classes for the lists. So while I normally argue for the encapsulations advantages a get wrapper provides, I'm just not feeling it in this case. A representative sample of my singleton public sealed class mySingl { private static volatile mySingl sngl; private static object lockObject = new Object(); public readonly Dictionary<string, string> myDict = new Dictionary<string, string>() { {"I", "index"}, {"D", "display"}, }; public enum parms { ABC = 10, DEF = 20, FGH = 30 }; public readonly List<parms> specParms = new List<parms>() { parms.ABC, parms.FGH }; public static mySingl Instance { get { if(sngl == null) { lock(lockObject) { if(sngl == null) sngl = new mySingl(); } } return sngl; } } private mySingl() { doSomething(); } } Questions: Am I taking the most reasonable approach in this case? Should I be worrying about const vs. readonly? is there a better way of providing this information?

    Read the article

  • What are the alternatives to "overriding a method" when using composition instead of inheritance?

    - by Sebastien Diot
    If we should favor composition over inheritance, the data part of it is clear, at least for me. What I don't have a clear solution to is how overwriting methods, or simply implementing them if they are defined in a pure virtual form, should be implemented. An obvious way is to wrap the instance representing the base-class into the instance representing the sub-class. But the major downsides of this are that if you have say 10 methods, and you want to override a single one, you still have to delegate every other methods anyway. And if there were several layers of inheritance, you have now several layers of wrapping, which becomes less and less efficient. Also, this only solve the problem of the object "client"; when another object calls the top wrapper, things happen like in inheritance. But when a method of the deepest instance, the base class, calls it's own methods that have been wrapped and modified, the wrapping has no effect: the call is performed by it's own method, instead of by the highest wrapper. One extreme alternative that would solve those problems would be to have one instance per method. You only wrap methods that you want to overwrite, so there is no pointless delegation. But now you end up with an incredible amount of classes and object instance, which will have a negative effect on memory usage, and this will require a lot more coding too. So, are there alternatives (preferably alternatives that can be used in Java), that: Do not result in many levels of pointless delegation without any changes. Make sure that not only the client of an object, but also all the code of the object itself, is aware of which implementation of method should be called. Does not result in an explosion of classes and instances. Ideally puts the extra memory overhead that is required at the "class"/"particular composition" level (static if you will), rather than having every object pay the memory overhead of composition. My feeling tells me that the instance representing the base class should be at the "top" of the stack/layers so it receives calls directly, and can process them directly too if they are not overwritten. But I don't know how to do it that way.

    Read the article

  • ASP.NET Hosting :: ASP.NET File Upload Control

    - by mbridge
    The asp.net FileUpload control allows a user to browse and upload files to the web server. From developers perspective, it is as simple as dragging and dropping the FileUpload control to the aspx page. An extra control, like a Button control, or some other control is needed, to actually save the file. <asp:FileUploadID="FileUpload1"runat="server"/> <asp:ButtonID="B1"runat="server"Text="Save"OnClick="B1_Click"/> By default, the FileUpload control allows a maximum of 4MB file to be uploaded and the execution timeout is 110 seconds. These properties can be changed from within the web.config file’s httpRuntime section. The maxRequestLength property determines the maximum file size that can be uploaded. The executionTimeout property determines the maximum time for execution. <httpRuntimemaxRequestLength="8192"executionTimeout="220"/> From code behind, the mime type, size of the file, file name and the extension of the file can be obtained. The maximum file size that can be uploaded can be obtained and modified using the System.Web.Configuration.HttpRuntimeSection class. Files can be alternatively saved using the System.IO.HttpFileCollection class. This collection class can be populated using the Request.Files property. The collection contains HttpPostedFile class which contains a reference to the class. using System; using System.Collections.Generic; using System.Linq; using System.Web; using System.Web.UI; using System.Web.UI.WebControls; using System.IO; using System.Configuration; using System.Web.Configuration;   namespace WebApplication1 {     public partial class WebControls : System.Web.UI.Page     {         protected void Page_Load(object sender, EventArgs e)         {         }           //Using FileUpload control to upload and save files         protected void B1_Click(object sender, EventArgs e)         {             if (FileUpload1.HasFile && FileUpload1.PostedFile.ContentLength > 0)             {                 //mime type of the uploaded file                 string mimeType = FileUpload1.PostedFile.ContentType;                   //size of the uploaded file                 int size = FileUpload1.PostedFile.ContentLength; // bytes                   //extension of the uploaded file                 string extension = System.IO.Path.GetExtension(FileUpload1.FileName);                                  //save file                 string path = Server.MapPath("path");                                 FileUpload1.SaveAs(path + FileUpload1.FileName);                              }             //maximum file size allowed             HttpRuntimeSection rt = new HttpRuntimeSection();             rt.MaxRequestLength = rt.MaxRequestLength * 2;             int length = rt.MaxRequestLength;                     //execution timeout             TimeSpan ts = rt.ExecutionTimeout;             double secomds = ts.TotalSeconds;           }           //Using Request.Files to save files         private void AltSaveFile()         {             HttpFileCollection coll = Request.Files;             for (int i = 0; i < coll.Count; i++)             {                 HttpPostedFile file = coll[i];                   if (file.ContentLength > 0)                     ;//do something             }         }     } }

    Read the article

  • How to include an apache library with my opensource code?

    - by OscarRyz
    I have this opensource code with MIT license that uses an Apache 2.0 licensed library. I want to include this in my project, so it can be built right away. In the point 4 of that license explains how to redistribute it: excerpt: 4 . Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions: You must give any other recipients of the Work or Derivative Works a copy of this License; and You must cause any modified files to carry prominent notices stating that You changed the files; and You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding those notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License. You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License. I'm not creating a derivative work ( I plan to provide it as it is ). I don't have a NOTICE file, just my my own LICENSE.txt file. Question: Where should I put something along the lines: "This project uses Xyz library distributed under Apache2.0 ..."? What's recommented? Should I provide the apache license file too? Or would be enough if I just say "Find the license online here...http://www.apache.org/licenses/LICENSE-2.0.html" I hope someone who has done this in the past may shed some light on the matter.

    Read the article

< Previous Page | 75 76 77 78 79 80 81 82 83 84 85 86  | Next Page >