Search Results

Search found 5634 results on 226 pages for 'conflicting libraries'.

Page 175/226 | < Previous Page | 171 172 173 174 175 176 177 178 179 180 181 182  | Next Page >

  • Is the science of Computer Science dead?

    - by Veaviticus
    Question : Is the science and art of CS dead? By that I mean, the real requirements to think, plan and efficiently solve problems seems to be falling away from CS these days. The field seems to be lowering the entry-barrier so more people can 'program' without having to learn how to truly program. Background : I'm a recent graduate with a BS in Computer Science. I'm working a starting position at a decent sized company in the IT department. I mostly do .NET and other Microsoft technologies at my job, but before this I've done Java stuff through internships and the like. I personally am a C++ programmer for my own for-fun projects. In Depth : Through the work I've been doing, it seems to me that the intense disciplines of a real science don't exist in CS anymore. In the past, programmers had to solve problems efficiently in order for systems to be robust and quick. But now, with the prevailing technologies like .NET, Java and scripting languages, it seems like efficiency and robustness have been traded for ease of development. Most of the colleagues that I work with don't even have degrees in Computer Science. Most graduated with Electrical Engineering degrees, a few with Software Engineering, even some who came from tech schools without a 4 year program. Yet they get by just fine without having the technical background of CS, without having studied theories and algorithms, without having any regard for making an elegant solution (they just go for the easiest, cheapest solution). The company pushes us to use Microsoft technologies, which take all the real thought out of the matter and replace it with libraries and tools that can auto-build your project for you half the time. I'm not trying to hate on the languages, I understand that they serve a purpose and do it well, but when your employees don't know how a hash-table works, and use the wrong sorting methods, or run SQL commands that are horribly inefficient (but get the job done in an acceptable time), it feels like more effort is being put into developing technologies that coddle new 'programmers' rather than actually teaching people how to do things right. I am interested in making efficient and, in my opinion, beautiful programs. If there is a better way to do it, I'd rather go back and refactor it than let it slide. But in the corporate world, they push me to complete tasks quickly rather than elegantly. And that really bugs me. Is this what I'm going to be looking forward to the rest of my life? Are there still positions out there for people who love the science and art of CS rather than just the paycheck? And on the same note, here's a good read if you haven't seen it before The Perils Of Java Schools

    Read the article

  • Open the SQL Server Error Log with PowerShell

    - by BuckWoody
    Using the Server Management Objects (SMO) library, you don’t even need to have the SQL Server 2008 PowerShell Provider to read the SQL Server Error Logs – in fact, you can use regular old everyday PowerShell. Keep in mind you will need the SMO libraries – which can be installed separately or by installing the Client Tools from the SQL Server install media. You could search for errors, store a result as a variable, or act on the returned values in some other way. Replace the Machine Name with your server and Instance Name with your instance, but leave the quotes, to make this work on your system: [reflection.assembly]::LoadWithPartialName("Microsoft.SqlServer.Smo") $machineName = "UNIVAC" $instanceName = "Production" $sqlServer = new-object ("Microsoft.SqlServer.Management.Smo.Server") "$machineName\$instanceName" $sqlServer.ReadErrorLog() Want to search for something specific, like the word “Error”? Replace the last line with this: $sqlServer.ReadErrorLog() | where {$_.Text -like "Error*"} 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 3 Hosting :: How to Upgrade ASP.NET MVC 2 Project to ASP.NET MVC 3

    - by mbridge
    ASP.NET MVC 3 can be installed side by side with ASP.NET MVC 2 on the same computer, which gives you flexibility in choosing when to upgrade an ASP.NET MVC 2 application to ASP.NET MVC 3. The simplest way to upgrade is to create a new ASP.NET MVC 3 project and copy all the views, controllers, code, and content files from the existing MVC 2 project to the new project and then to update the assembly references in the new project to match the old project. If you have made changes to the Web.config file in the MVC 2 project, you must also merge those changes with the Web.config file in the MVC 3 project. To manually upgrade an existing ASP.NET MVC 2 application to version 3, do the following: 1. In both Web.config files in the MVC 3 project, globally search and replace the MVC version. Find the following: System.Web.Mvc, Version=2.0.0.0 Replace it with the following System.Web.Mvc, Version=3.0.0.0 There are three changes in the root Web.config and four in the Views\Web.config file. 2. In Solution Explorer, delete the reference to System.Web.Mvc (which points to the version 2 DLL). Then add a reference to System.Web.Mvc (v3.0.0.0). 3. In Solution Explorer, right-click the project name and then select Unload Project. Then right-click again and select Edit ProjectName.csproj. 4. Locate the ProjectTypeGuids element and replace {F85E285D-A4E0-4152-9332-AB1D724D3325} with {E53F8FEA-EAE0-44A6-8774-FFD645390401}. 5. Save the changes and then right-click the project and select Reload Project. 6. If the project references any third-party libraries that are compiled using ASP.NET MVC 2, add the following highlighted bindingRedirect element to the Web.config file in the application root under the configuration section: <runtime>   <assemblyBinding >     <dependentAssembly>       <assemblyIdentity name="System.Web.Mvc"           publicKeyToken="31bf3856ad364e35"/>       <bindingRedirect oldVersion="2.0.0.0" newVersion="3.0.0.0"/>     </dependentAssembly>   </assemblyBinding> </runtime> Another ASP.NET MVC 3 article: - Rolling with Razor in MVC v3 Preview - Deploying ASP.NET MVC 3 web application to server where ASP.NET MVC 3 is not installed - RenderAction with ASP.NET MVC 3 Sessionless Controllers

    Read the article

  • Oracle releases new Java Embedded products

    - by Henrik Stahl
    With less than one week to go to JavaOne 2012, we've spiced things up a little by releasing not one but two net new embedded Java products. This is an important step towards realizing the vision of Java as the standard platform for the Internet of Things that I outlined in a recent blog post. The two new products are: Java ME Embedded 3.2. Based on same code as the widely deployed Oracle Java Wireless Client for feature phones, this new product provides a Java ME implementation optimized for very small microcontroller-based devices and adds - among other things - a new Device Access API that enables interaction with peripherals common in edge devices such as various types of sensors. In addition to the new Java ME Embedded platform, we have also released an update of the Java ME SDK which adds support for the development of small embedded devices. Java Embedded Suite 7.0. This is an integrated middleware stack for embedded devices, incorporating Java SE Embedded and versions of JavaDB, GlassFish and a Web Services stack optimized for remote operation and small footprint. A typical Internet of Things (or M2M) infrastructure contains three types of compute nodes: The edge device which is typically a sensor or control point of some kind. These devices can be connected directly to a backend through a mobile network if they are installed in - for example - a remote vending machine; or, they can be part of a local short-range network and be connected to the backend through a more powerful gateway device. A gateway is the second type of compute node and acts as an aggregator and control point for a local network. A good example of this could be a generalized home Internet access point, or home gateway. Gateways are mostly using normal wall power and are used for multiple applications, deployed by multiple service providers. Finally, the last type of compute node is the normal enterprise or cloud backend. Java ME Embedded and Java Embedded Suite are perfect base software stacks for the edge devices and the gateway respectively, providing the Java promise of a platform independent runtime and a complete set of libraries as well as allowing a programmer to focus on the business logic rather than plumbing. We are very thrilled with these new releases that open up exciting opportunities for Java developers to extend services and enterprise applications in ways that will make organizations more efficient and touch our daily lives. To find out more, come to the JavaOne conference (for technical content) and to the Java Embedded @ JavaOne subconference (for business content). There will be plenty of cool demos showing complete end-to-end applications, provided by Oracle and our partners, as well as keynotes and numerous sessions where you can learn more about the technology and business opportunities.

    Read the article

  • Java Spotlight Episode 78: Jasper Potts on the JavaFX Scene Builder

    - by Roger Brinkley
    Tweet An interview with Jasper Potts about the new JavaFX Scene Builder. Joining us this week on the Java All Star Developer Panel are Dalibor Topic, Java Free and Open Source Software Ambassador and Arun Gupta, Java EE Guy. Right-click or Control-click to download this MP3 file. You can also subscribe to the Java Spotlight Podcast Feed to get the latest podcast automatically. If you use iTunes you can open iTunes and subscribe with this link:  Java Spotlight Podcast in iTunes. Show Notes News JavaFX Scene Builder Developer Preview available for testing. Java EE Unlock the Java EE 6 Platform using NetBeans 7.1 Tuning GlassFish for Production JSF 2.2 Update from Ed Burns John Rose at Microsoft's Lang.NEXT summit Recording of John's Java 8 presentation Jeroen Frijters' presentation on IKVM.NET Martin Odersky's keynote JVM Language Summit 2012 July 30 – August 1; Oracle Santa Clara (same as last year) CFP coming in a few days JVM Language Summit 2011 Presentations & Recordings Proposed development schedule for JDK 8 Say hello to Mathias Axelsson Events April 11, Cleveland JUG, Cleveland, OH April 12, GreenJUG, Greenville, SC April 17-18, JavaOne Russia, Moscow Russia April 18–20, Devoxx France, Paris, France April 17-20, GIDS, Bangalore April 21, Java Summit, Chennai April 26, Mix-IT, Lyon, France, May 3-4, JavaOne India, Hyderabad, India May 5, Bangalore, Pune, ?? - JUG outreach May 7, OTN Developer Day, Mumbai May 8, OTN Developer Day, Delhi Feature InterviewJasper Potts is the Developer Experience Architect for the Java Client Group at Oracle. Responsible for technical design for everything thats sis on the core platform including Controls, Tools, Samples and Blueprints. Formally a lead engineer on the JavaFX & Swing teams working on the new JavaFX UI Controls and Graphics frameworks. Also responsible for designing, developing and presenting demos during the keynotes at JavaOne and Devoxx. A JavaOne Rockstar presenter having presented many sessions on JavaFX and Swing at many conferences. Prior to Sun he founded Xerto a desktop applications company developing Imagery a Java professional photo management application. In this interview Jasper talks about the recently release JavaFX Scene Builder. Mail Bag What’s Cool Contribute to GlassFish in Five Different Ways Stephen Chin and James Weaver join Oracle Adam Bien - Building Java FX 2 Libraries From Source With Maven 3 Paul Sandoz - Java Boomerang Building Jigsaw on Mac OS X using VirtualBox Mandy Chung: Jigsaw for Mac OS X

    Read the article

  • GPL vs plugin interfaces not designed with a specific application in mind

    - by Kristóf Marussy
    I am not seeking or in need of legal advice, but an interesting though experiment came to my mind. Imagine the following situtation (I cannot really think about a concrete example and I am unsure if a real manifestation even exists): there is a free (libre) api A licensed under some permissive license or even LGPL. Non-free application B implements this api in order host plugins, but there are other free software doing the same thing. Moreover, there is plugin C acting as a plugin under api A. It links to library D, that is under GPL, so C is also under GPL. Plugins using A are loaded into hosts via a dlopen-like mechanism and use complex data structure for host-plugin communication. Neither B nor C distribute any files that may be required for A to function properly (like headers containing the structure definitions of A or dynamic libraries containing helper functions for A written by the authors of A), but such things may exist. Now some user installs application B and plugin C on his machine, along with anything that may be required for api A to function properly. Then he proceeds and loads C into B and creates some intellectual property with B which is not a piece of software. Did a GPL violation happend at some point, and if so, who violated GPL and why? The authors of C violate D's license by making C possible to be used in non-free host B? This is a possibility because they can't give and exception of GPL (like one described in http://www.gnu.org/licenses/gpl-faq.html#GPLPluginsInNF or http://www.gnu.org/licenses/gpl-faq.html#LinkingOverControlledInterface) due to D's license terms. The authors of B violate C's and D's license by making C possible to be loaded in B? This is a possibility because http://www.gnu.org/licenses/gpl-faq.html#NFUseGPLPlugins disallows the mechanisms A uses for communitation between the free and non-free modules. The authors of A, because the api may be used (and in this case, was used) for communication between GPL'd and non-free software. This would be extremely absurd. The user, because at the moment of loading B into C, he made a derived work of C. I think this is impossible, because he does not distribute it. But would the situation change is he decided to release a configuration file of B which makes B load C as a plugin? Nobody, because A counts as a 'system library', and both B and C directly interact only with A, not eachother. In a sane world, this would happen... A concrete example of A could be some kind of audio (think LADSPA) or image processing api. However, I could find no such interface (that is free software, generic and is also implemented by commercial tools). A real-world example could also be quite enlightening.

    Read the article

  • ArchBeat Link-o-Rama for 2012-06-22

    - by Bob Rhubart
    Guide to integration architecture | Stephanie Mann "The landscape of integration architecture is shifting as service-oriented and cloud-based architecture take the fore," says Stephanie Mann. "To ensure success, enterprise architects and developers are turning to lighter-weight infrastructure to support more complex integration projects." FY13 Oracle PartnerNetwork Kickoff - Tues June 26, 2012 Join us for a one-hour live online event hosted by the Oracle PartnerNetwork team as we kickoff FY13. Other dates/times for EMEA/LAD/JAPAN/APAC. Click the link for details. Why should you choose Oracle WebLogic 12c instead of JBoss EAP 6? | Ricardo Ferreira Okay, you would expect an Oracle guy to make this argument. But Ferreira takes a very deep, very detailed technical dive into the issue. So hear the man out, will ya? Hibernate4 and Coherence | Rene van Wijk According to Oracle ACE Rene van Wijk, "there are two ways to integrate Hibernate and Coherence." In this post he illustrates one of them. Simple Made Easy | Rich Hickey Rich Hickey discusses simplicity, why it is important, how to achieve it in design and how to recognize its absence in the tools, language constructs and libraries in this presentation from QCon London 2012. Starting a cluster | Mark Nelson Fusion Middleware A-Team blogger Mark Nelson looks at Oracle SOA Suite, Oracle BPM, and Oracle Coherence, three products that are " commonly clustered, and which have somewhat different requirements." Why building SaaS well means giving up your servers | GigaOM The biggest benefit to PaaS, reports GigaOM's Derrick Harris, "might be a better product because the company is able to focus on building the app rather than managing servers." Personas - what, why & how | Mascha van Oosterhout "To be able to create a successful, user-friendly website or application," says Mascha van Oosterhout, "every decision you take, whether you are part of the marketing team, the design team or the development team, should be based on what you know about the user." Thought for the Day "Machines take me by surprise with great frequency." — Alan Turing(June 23, 1912 - June 7, 1954) Source: Brainy Quote

    Read the article

  • SQL Server PowerShell Provider follows the Version of PowerShell on the Host and other errata

    - by BuckWoody
    There may be some misunderstanding on how the PowerShell Provider for SQL Server works. I’ve written an article or two explaining that you can use PowerShell with SQL Server, without having the SQL Server 2008 (or higher) provider around. After all, PowerShell just uses .NET, and SQL Server “Server Management Objects” or SMO listen to that interface as well. In SQL Server 2008 and higher we created a “MiniShell” for PowerShell that gives you the ability to treat a SQL Server Instance as a drive (called a “Provider” or path or drive) and a few commands (called command-lets). Using these two simple constructs you can move around SQL Server quickly and work with the objects it holds. I read the other day where someone stated that we had “re-compiled” PowerShell, so that you would have version 1.0 from SQL Server and 2.0 on your new server. Not so! Drop to a SQLPS prompt and a PowerShell prompt and type this in each: $PSVersionTable They should return the same value. You can think of a MiniShell as simply a compiled “profile” that gives you those providers and command-lets automatically – that’s all. In fact, you can load the SMO libraries yourself without the SQL Server 2008 Provider anywhere in sight. I do this all the time, since the MiniShell also has other restrictions. Also remember that if you run a PowerShell script as a SQL Agent Job step type (in 2008 and higher) that you’re running under the context of the account that starts Agent – I think most folks know this, but it’s good to keep in mind. There’s a re-written section of Books Online that goes over working with this very nicely – also covers the question “How to I connect to another server using the SQL Server PowerShell Provider” (hint: It’s just CD) and “How do I load all the SMO stuff if I don’t want to use the Provider” and more. Be sure and check out the note at the bottom that explains the firewall exceptions you’ll need to enable to CD to that remote server. Here’s that link: http://msdn.microsoft.com/en-us/library/cc281947.aspx Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • Nagging As A Strategy For Better Linking: -z guidance

    - by user9154181
    The link-editor (ld) in Solaris 11 has a new feature that we call guidance that is intended to help you build better objects. The basic idea behind guidance is that if (and only if) you request it, the link-editor will issue messages suggesting better options and other changes you might make to your ld command to get better results. You can choose to take the advice, or you can disable specific types of guidance while acting on others. In some ways, this works like an experienced friend leaning over your shoulder and giving you advice — you're free to take it or leave it as you see fit, but you get nudged to do a better job than you might have otherwise. We use guidance to build the core Solaris OS, and it has proven to be useful, both in improving our objects, and in making sure that regressions don't creep back in later. In this article, I'm going to describe the evolution in thinking and design that led to the implementation of the -z guidance option, as well as give a brief description of how it works. The guidance feature issues non-fatal warnings. However, experience shows that once developers get used to ignoring warnings, it is inevitable that real problems will be lost in the noise and ignored or missed. This is why we have a zero tolerance policy against build noise in the core Solaris OS. In order to get maximum benefit from -z guidance while maintaining this policy, I added the -z fatal-warnings option at the same time. Much of the material presented here is adapted from the arc case: PSARC 2010/312 Link-editor guidance The History Of Unfortunate Link-Editor Defaults The Solaris link-editor is one of the oldest Unix commands. It stands to reason that this would be true — in order to write an operating system, you need the ability to compile and link code. The original link-editor (ld) had defaults that made sense at the time. As new features were needed, command line option switches were added to let the user use them, while maintaining backward compatibility for those who didn't. Backward compatibility is always a concern in system design, but is particularly important in the case of the tool chain (compilers, linker, and related tools), since it is a basic building block for the entire system. Over the years, applications have grown in size and complexity. Important concepts like dynamic linking that didn't exist in the original Unix system were invented. Object file formats changed. In the case of System V Release 4 Unix derivatives like Solaris, the ELF (Extensible Linking Format) was adopted. Since then, the ELF system has evolved to provide tools needed to manage today's larger and more complex environments. Features such as lazy loading, and direct bindings have been added. In an ideal world, many of these options would be defaults, with rarely used options that allow the user to turn them off. However, the reality is exactly the reverse: For backward compatibility, these features are all options that must be explicitly turned on by the user. This has led to a situation in which most applications do not take advantage of the many improvements that have been made in linking over the last 20 years. If their code seems to link and run without issue, what motivation does a developer have to read a complex manpage, absorb the information provided, choose the features that matter for their application, and apply them? Experience shows that only the most motivated and diligent programmers will make that effort. We know that most programs would be improved if we could just get you to use the various whizzy features that we provide, but the defaults conspire against us. We have long wanted to do something to make it easier for our users to use the linkers more effectively. There have been many conversations over the years regarding this issue, and how to address it. They always break down along the following lines: Change ld Defaults Since the world would be a better place the newer ld features were the defaults, why not change things to make it so? This idea is simple, elegant, and impossible. Doing so would break a large number of existing applications, including those of ISVs, big customers, and a plethora of existing open source packages. In each case, the owner of that code may choose to follow our lead and fix their code, or they may view it as an invitation to reconsider their commitment to our platform. Backward compatibility, and our installed base of working software, is one of our greatest assets, and not something to be lightly put at risk. Breaking backward compatibility at this level of the system is likely to do more harm than good. But, it sure is tempting. New Link-Editor One might create a new linker command, not called 'ld', leaving the old command as it is. The new one could use the same code as ld, but would offer only modern options, with the proper defaults for features such as direct binding. The resulting link-editor would be a pleasure to use. However, the approach is doomed to niche status. There is a vast pile of exiting code in the world built around the existing ld command, that reaches back to the 1970's. ld use is embedded in large and unknown numbers of makefiles, and is used by name by compilers that execute it. A Unix link-editor that is not named ld will not find a majority audience no matter how good it might be. Finally, a new linker command will eventually cease to be new, and will accumulate its own burden of backward compatibility issues. An Option To Make ld Do The Right Things Automatically This line of reasoning is best summarized by a CR filed in 2005, entitled 6239804 make it easier for ld(1) to do what's best The idea is to have a '-z best' option that unchains ld from its backward compatibility commitment, and allows it to turn on the "best" set of features, as determined by the authors of ld. The specific set of features enabled by -z best would be subject to change over time, as requirements change. This idea is more realistic than the other two, but was never implemented because it has some important issues that we could never answer to our satisfaction: The -z best proposal assumes that the user can turn it on, and trust it to select good options without the user needing to be aware of the options being applied. This is a fallacy. Features such as direct bindings require the user to do some analysis to ensure that the resulting program will still operate properly. A user who is willing to do the work to verify that what -z best does will be OK for their application is capable of turning on those features directly, and therefore gains little added benefit from -z best. The intent is that when a user opts into -z best, that they understand that z best is subject to sometimes incompatible evolution. Experience teaches us that this won't work. People will use this feature, the meaning of -z best will change, code that used to build will fail, and then there will be complaints and demands to retract the change. When (not if) this occurs, we will of course defend our actions, and point at the disclaimer. We'll win some of those debates, and lose others. Ultimately, we'll end up with -z best2 (-z better), or other compromises, and our goal of simplifying the world will have failed. The -z best idea rolls up a set of features that may or may not be related to each other into a unit that must be taken wholesale, or not at all. It could be that only a subset of what it does is compatible with a given application, in which case the user is expected to abandon -z best and instead set the options that apply to their application directly. In doing so, they lose one of the benefits of -z best, that if you use it, future versions of ld may choose a different set of options, and automatically improve the object through the act of rebuilding it. I drew two conclusions from the above history: For a link-editor, backward compatibility is vital. If a given command line linked your application 10 years ago, you have every reason to expect that it will link today, assuming that the libraries you're linking against are still available and compatible with their previous interfaces. For an application of any size or complexity, there is no substitute for the work involved in examining the code and determining which linker options apply and which do not. These options are largely orthogonal to each other, and it can be reasonable not to use any or all of them, depending on the situation, even in modern applications. It is a mistake to tie them together. The idea for -z guidance came from consideration of these points. By decoupling the advice from the act of taking the advice, we can retain the good aspects of -z best while avoiding its pitfalls: -z guidance gives advice, but the decision to take that advice remains with the user who must evaluate its merit and make a decision to take it or not. As such, we are free to change the specific guidance given in future releases of ld, without breaking existing applications. The only fallout from this will be some new warnings in the build output, which can be ignored or dealt with at the user's convenience. It does not couple the various features given into a single "take it or leave it" option, meaning that there will never be a need to offer "-zguidance2", or other such variants as things change over time. Guidance has the potential to be our final word on this subject. The user is given the flexibility to disable specific categories of guidance without losing the benefit of others, including those that might be added to future versions of the system. Although -z fatal-warnings stands on its own as a useful feature, it is of particular interest in combination with -z guidance. Used together, the guidance turns from advice to hard requirement: The user must either make the suggested change, or explicitly reject the advice by specifying a guidance exception token, in order to get a build. This is valuable in environments with high coding standards. ld Command Line Options The guidance effort resulted in new link-editor options for guidance and for turning warnings into fatal errors. Before I reproduce that text here, I'd like to highlight the strategic decisions embedded in the guidance feature: In order to get guidance, you have to opt in. We hope you will opt in, and believe you'll get better objects if you do, but our default mode of operation will continue as it always has, with full backward compatibility, and without judgement. Guidance suggestions always offers specific advice, and not vague generalizations. You can disable some guidance without turning off the entire feature. When you get guidance warnings, you can choose to take the advice, or you can specify a keyword to disable guidance for just that category. This allows you to get guidance for things that are useful to you, without being bothered about things that you've already considered and dismissed. As the world changes, we will add new guidance to steer you in the right direction. All such new guidance will come with a keyword that let's you turn it off. In order to facilitate building your code on different versions of Solaris, we quietly ignore any guidance keywords we don't recognize, assuming that they are intended for newer versions of the link-editor. If you want to see what guidance tokens ld does and does not recognize on your system, you can use the ld debugging feature as follows: % ld -Dargs -z guidance=foo,nodefs debug: debug: Solaris Linkers: 5.11-1.2275 debug: debug: arg[1] option=-D: option-argument: args debug: arg[2] option=-z: option-argument: guidance=foo,nodefs debug: warning: unrecognized -z guidance item: foo The -z fatal-warning option is straightforward, and generally useful in environments with strict coding standards. Note that the GNU ld already had this feature, and we accept their option names as synonyms: -z fatal-warnings | nofatal-warnings --fatal-warnings | --no-fatal-warnings The -z fatal-warnings and the --fatal-warnings option cause the link-editor to treat warnings as fatal errors. The -z nofatal-warnings and the --no-fatal-warnings option cause the link-editor to treat warnings as non-fatal. This is the default behavior. The -z guidance option is defined as follows: -z guidance[=item1,item2,...] Provide guidance messages to suggest ld options that can improve the quality of the resulting object, or which are otherwise considered to be beneficial. The specific guidance offered is subject to change over time as the system evolves. Obsolete guidance offered by older versions of ld may be dropped in new versions. Similarly, new guidance may be added to new versions of ld. Guidance therefore always represents current best practices. It is possible to enable guidance, while preventing specific guidance messages, by providing a list of item tokens, representing the class of guidance to be suppressed. In this way, unwanted advice can be suppressed without losing the benefit of other guidance. Unrecognized item tokens are quietly ignored by ld, allowing a given ld command line to be executed on a variety of older or newer versions of Solaris. The guidance offered by the current version of ld, and the item tokens used to disable these messages, are as follows. Specify Required Dependencies Dynamic executables and shared objects should explicitly define all of the dependencies they require. Guidance recommends the use of the -z defs option, should any symbol references remain unsatisfied when building dynamic objects. This guidance can be disabled with -z guidance=nodefs. Do Not Specify Non-Required Dependencies Dynamic executables and shared objects should not define any dependencies that do not satisfy the symbol references made by the dynamic object. Guidance recommends that unused dependencies be removed. This guidance can be disabled with -z guidance=nounused. Lazy Loading Dependencies should be identified for lazy loading. Guidance recommends the use of the -z lazyload option should any dependency be processed before either a -z lazyload or -z nolazyload option is encountered. This guidance can be disabled with -z guidance=nolazyload. Direct Bindings Dependencies should be referenced with direct bindings. Guidance recommends the use of the -B direct, or -z direct options should any dependency be processed before either of these options, or the -z nodirect option is encountered. This guidance can be disabled with -z guidance=nodirect. Pure Text Segment Dynamic objects should not contain relocations to non-writable, allocable sections. Guidance recommends compiling objects with Position Independent Code (PIC) should any relocations against the text segment remain, and neither the -z textwarn or -z textoff options are encountered. This guidance can be disabled with -z guidance=notext. Mapfile Syntax All mapfiles should use the version 2 mapfile syntax. Guidance recommends the use of the version 2 syntax should any mapfiles be encountered that use the version 1 syntax. This guidance can be disabled with -z guidance=nomapfile. Library Search Path Inappropriate dependencies that are encountered by ld are quietly ignored. For example, a 32-bit dependency that is encountered when generating a 64-bit object is ignored. These dependencies can result from incorrect search path settings, such as supplying an incorrect -L option. Although benign, this dependency processing is wasteful, and might hide a build problem that should be solved. Guidance recommends the removal of any inappropriate dependencies. This guidance can be disabled with -z guidance=nolibpath. In addition, -z guidance=noall can be used to entirely disable the guidance feature. See Chapter 7, Link-Editor Quick Reference, in the Linker and Libraries Guide for more information on guidance and advice for building better objects. Example The following example demonstrates how the guidance feature is intended to work. We will build a shared object that has a variety of shortcomings: Does not specify all it's dependencies Specifies dependencies it does not use Does not use direct bindings Uses a version 1 mapfile Contains relocations to the readonly allocable text (not PIC) This scenario is sadly very common — many shared objects have one or more of these issues. % cat hello.c #include <stdio.h> #include <unistd.h> void hello(void) { printf("hello user %d\n", getpid()); } % cat mapfile.v1 # This version 1 mapfile will trigger a guidance message % cc hello.c -o hello.so -G -M mapfile.v1 -lelf As you can see, the operation completes without error, resulting in a usable object. However, turning on guidance reveals a number of things that could be better: % cc hello.c -o hello.so -G -M mapfile.v1 -lelf -zguidance ld: guidance: version 2 mapfile syntax recommended: mapfile.v1 ld: guidance: -z lazyload option recommended before first dependency ld: guidance: -B direct or -z direct option recommended before first dependency Undefined first referenced symbol in file getpid hello.o (symbol belongs to implicit dependency /lib/libc.so.1) printf hello.o (symbol belongs to implicit dependency /lib/libc.so.1) ld: warning: symbol referencing errors ld: guidance: -z defs option recommended for shared objects ld: guidance: removal of unused dependency recommended: libelf.so.1 warning: Text relocation remains referenced against symbol offset in file .rodata1 (section) 0xa hello.o getpid 0x4 hello.o printf 0xf hello.o ld: guidance: position independent (PIC) code recommended for shared objects ld: guidance: see ld(1) -z guidance for more information Given the explicit advice in the above guidance messages, it is relatively easy to modify the example to do the right things: % cat mapfile.v2 # This version 2 mapfile will not trigger a guidance message $mapfile_version 2 % cc hello.c -o hello.so -Kpic -G -Bdirect -M mapfile.v2 -lc -zguidance There are situations in which the guidance does not fit the object being built. For instance, you want to build an object without direct bindings: % cc -Kpic hello.c -o hello.so -G -M mapfile.v2 -lc -zguidance ld: guidance: -B direct or -z direct option recommended before first dependency ld: guidance: see ld(1) -z guidance for more information It is easy to disable that specific guidance warning without losing the overall benefit from allowing the remainder of the guidance feature to operate: % cc -Kpic hello.c -o hello.so -G -M mapfile.v2 -lc -zguidance=nodirect Conclusions The linking guidelines enforced by the ld guidance feature correspond rather directly to our standards for building the core Solaris OS. I'm sure that comes as no surprise. It only makes sense that we would want to build our own product as well as we know how. Solaris is usually the first significant test for any new linker feature. We now enable guidance by default for all builds, and the effect has been very positive. Guidance helps us find suboptimal objects more quickly. Programmers get concrete advice for what to change instead of vague generalities. Even in the cases where we override the guidance, the makefile rules to do so serve as documentation of the fact. Deciding to use guidance is likely to cause some up front work for most code, as it forces you to consider using new features such as direct bindings. Such investigation is worthwhile, but does not come for free. However, the guidance suggestions offer a structured and straightforward way to tackle modernizing your objects, and once that work is done, for keeping them that way. The investment is often worth it, and will replay you in terms of better performance and fewer problems. I hope that you find guidance to be as useful as we have.

    Read the article

  • As a web designer, which language should I learn first for my feature career? (PHP or JavaScript) [closed]

    - by kdevs3
    Possible Duplicates: Best Programming Language for Web Development How can I choose a web development language? What language will you choose if you are going to build something big? What is the right option of programming languages and tools for building our website? What is the easiest web programing language at....? Well, I'm more of a basic web designer. I know the easy stuff pretty well. (Ya know, html, css) But I've been trying to take it to the next step and I'm contemplating about what I should learn that will help me out the most in my future web design/programming career, should it be JavaScript or maybe I should try to learn a back end programming language such as PHP. Lately, I have been hearing about a lot how JavaScript is so great and useful now, because of libraries such as jQuery and what possibility's it can bring by using Node.js and other frameworks. I've only learned the most basic of JavaScript and used some jQuery (mostly plugins) so i wouldn't know at all of what it can actually do. Would JS being so popular as it is now and useful, be a reason to stick with JavaScript and only learn it that for now? Or as a web designer, how important would it be to learn how to make a web application/website operate and functional, and know how to work with servers, etc? (Such as getting forms to work and sending data to the server and back) I've took a look at frameworks such as Code Igniter before, and looks really simple to get started with if I try to learn PHP, But I'm not sure how important it is for my career and what I would gain out of it. I'm asking because I can't decide what I should learn first. When I select it, I really want to take my time and learn the language. I don't want to spend time on learning multiple languages at the same time, so I need to pick wisely. I'm trying to turn the right direction so my career can hopefully be successful in the feature. (If money/gaining a job asked if its important, then its a yeah, it is a bit) I'm hoping I can get opinions and suggestions on this question, thanks for giving me your thoughts also.

    Read the article

  • Upgrading Code from 2007 to 2010

    - by MOSSLover
    So I’ve been doing some upgrades just to see if things will work from 2007 to 2010.  So far most of the stuff I want works, but obviously there are some things that break.  Did you guys know that in 2007 you could add a webpart to the view pages for lists and libraries without losing the toolbar?  In 2010 the ribbon disappears every time you add a webpart.  So if you are using Scot Hillier’s Codeplex project to hide buttons it will not work the same way, because the ribbon is going to disappear altogether. I have also learned another reason why standalone installations are the bane of my existence.  Nine times out of ten the installation is done using Network Service as the application pool account.  You are wondering why is this bad?  Well, let’s just say the site collection administrator with local admin rights wants to attach the IIS Worker process and debug say a webpart.  Visual Studio 2010 will throw a nasty error that tells you that you are not an administrator.  You will say, but I am an administrator?  I have all the correct group permissions on the server and on SQL and in SharePoint.  Then you will go in and decide let’s add my own admin account just to see if I can attach the debugger and you will notice that works properly.  So the morale of the story is create a separate account on your development environment to run all the SharePoint Services and such.  You don’t need to go all out and create the best practices amount of accounts if it’s just your dev environment.  I would at least create one single account to run all your SharePoint process (Services, SQL, and App Pool).  Also, don’t run a standalone install unless you want to kill kittens (this is a quote from Todd Klindt).  We love kittens they are cute and awesome.  Besides you learn more if you click Complete and just skip standalone.  You will learn how to setup SQL Server 2008 and you will learn how to configure your environment.  It will help you in the long run.  So I have ranted enough for today I figure these are enough tidbits for you this time around.  The two of you who read my blog and I know some of you are friends who don’t understand SharePoint.  I might as well have just done “wahwahwahwah” in Charlie Brown adult speak.  Thanks for reading as usual.  I’ll catch you all when I complain more about the upgrade process and share more tidbits, which will inevitably become a presentation at a conference or two. Technorati Tags: Upgrade Code SharePoint 2007 to 2010,Visuaul Studio 2010,SharePoint 2010

    Read the article

  • Review: Backbone.js Testing

    - by george_v_reilly
    Title: Backbone.js Testing Author: Ryan Roemer Rating: $stars(4.5) Publisher: Packt Copyright: 2013 ISBN: 178216524X Pages: 168 Keywords: programming, testing, javascript, backbone, mocha, chai, sinon Reading period: October 2013 Backbone.js Testing is a short, dense introduction to testing JavaScript applications with three testing libraries, Mocha, Chai, and Sinon.JS. Although the author uses a sample application of a personal note manager written with Backbone.js throughout the book, much of the material would apply to any JavaScript client or server framework. Mocha is a test framework that can be executed in the browser or by Node.js, which runs your tests. Chai is a framework-agnostic TDD/BDD assertion library. Sinon.JS provides standalone test spies, stubs and mocks for JavaScript. They complement each other and the author does a good job of explaining when and how to use each. I've written a lot of tests in Python (unittest and mock, primarily) and C# (NUnit), but my experience with JavaScript unit testing was both limited and years out of date. The JavaScript ecosystem continues to evolve rapidly, with new browser frameworks and Node packages springing up everywhere. JavaScript has some particular challenges in testing—notably, asynchrony and callbacks. Mocha, Chai, and Sinon meet those challenges, though they can't take away all the pain. The author describes how to test Backbone models, views, and collections; dealing with asynchrony; provides useful testing heuristics, including isolating components to reduce dependencies; when to use stubs and mocks and fake servers; and test automation with PhantomJS. He does not, however, teach you Backbone.js itself; for that, you'll need another book. There are a few areas which I thought were dealt with too lightly. There's no real discussion of Test-driven_development or Behavior-driven_development, which provide the intellectual foundations of much of the book. Nor does he have much to say about testability and how to make legacy code more testable. The sample Notes app has plenty of testing seams (much of this falls naturally out of the architecture of Backbone); other apps are not so lucky. The chapter on automation is extremely terse—it could be expanded into a very large book!—but it does provide useful indicators to many areas for exploration. I learned a lot from this book and I have no hesitation in recommending it. Disclosure: Thanks to Ryan Roemer and Packt for a review copy of this book.

    Read the article

  • WebLogic Partner Community Newsletter June 2013

    - by JuergenKress
    Dear WebLogic Partner Community member, This month newsletter contains tons of Java material with the availability of Java EE7! Including many articles in the May/June Java Magazine and the Duke Award nomination! Thanks for all your efforts to become certified and Specialized. For all the experts who achieved the WebLogic Server 12c Certified Implementation Specialist  or ADF 11g Certified Implementation Specialist you can download a logo for your blog or business card at the Competence Center. For all the companies who achieved a WebLogic and ADF Specialization you can request a nice plaque for your office. Summer time is training time – make sure you attend our Fusion Middleware Summer Camps or one of our upcoming Exalogic Implementation Workshop by PTS in Spain, Turkey and Middle East. For those who can not make it we offers plenty of online courses like the New ADF Academy. Or the webcast The value of Engineered Systems for SAP. At our WebLogic Community Workspace (WebLogic Community membership required) you can find Engineered Systems Strategy presentation. Thanks to the community for all the WebLogic best practice papers and tools like Automated provision of Oracle Weblogic Server Platform & Frank’s Coherence videos on YouTube & Create and deploy applications on the Oracle Java Cloud & WebLogic Application redeployment using shared libraries - without downtime & Oracle Traffic Director. The first tests deployments for WebLogic on Oracle Database Appliance are on the way, thanks to all who are their experience: Sizing & Configuration and Traffic Director. Virtualised Oracle Database Appliance Proof of Concept - #1 Planning from Simon Haslam. Would be great if you can also share your experience via twitter @wlscommunity! In the ADF section of the newsletter you find Tuning Application Module Pools and Connection Pools & List View - Cool Looking ADF PS6 Component for Collections & Insider Essential & User Interface & Skinning & Oracle Forms to ADF Modernization reference. Great summer time! Jürgen Kress Oracle WebLogic Partner Adoption EMEA To read the newsletter please visit http://tinyurl.com/WebLogicnewsJune2013 (OPN Account required) To become a member of the WebLogic Partner Community please register at http://www.oracle.com/partners/goto/wls-emea ( OPN account required). If you need support with your account please contact the Oracle Partner Business Center. Blog Twitter LinkedIn Mix Forum Wiki Technorati Tags: WebLogic Community newsletter,newsletter,WebLogic,WebLogic Community,Oracle,OPN,Jürgen Kress

    Read the article

  • Should CSS be listed on your resume under Languages?

    - by Sandeepan Nath
    I have some doubts like Whether CSS should be put under Languages or not? Although Wikipedia says Cascading Style Sheets (CSS) is a style sheet language ... But do they write CSS under the languages section of the resume, along with PHP, etc? Similarly what about HTML? I have some doubt and I don't want to sound like someone who is not aware of the trends. Just to give an example, currently I have the following languages,frameworks, technologies, etc. listed under the "Technical Expertise" section of my resume - Technical Expertise * Languages - Proficient - PHP 5, Javascript, HTML ?*,CSS ?*,Sass ?*. Beginner - Linux Bash. * Databases - MySQL 5. * Technologies - AJAX. * Frameworks/Libraries - Symfony, jQuery. * CMSes - Wordpress. Although my domain is Web-development/design, I welcome domain-agnostic answers which can provide some generic ideas/reasoning. I have seen, a lot of people messing up these sections (even more serious than my doubts :) ), putting things under wrong sub-headings and thus putting a big question mark on their understanding of those things. I don't know much about XML, Comet Technology etc. Considering those are included too, What things should be put under Languages? E.g. Should CSS be put under Languages? Please give some reasoning to support your views. Where should the others (XML, Comet, cURL etc. ) be put? I welcome some examples of how you put it. Or do you have an additional Keywords section where you write all the unsortables ? Considering a set of standards like W3C standards, etc. do you have a standards sub-heading? I guess I have put the contents of other sections Okay. But do let me know about your ideas and reasoning. After all, I understand there may not be a single answer to this, but let's see what is the trend. Thanks Updates Further, do you mention design patters you have used? Web Services etc.? Where do you mention SOAP, XML etc...

    Read the article

  • Is there a better term than "smoothness" or "granularity" to describe this language feature?

    - by Chris
    One of the best things about programming is the abundance of different languages. There are general purpose languages like C++ and Java, as well as little languages like XSLT and AWK. When comparing languages, people often use things like speed, power, expressiveness, and portability as the important distinguishing features. There is one characteristic of languages I consider to be important that, so far, I haven't heard [or been able to come up with] a good term for: how well a language scales from writing tiny programs to writing huge programs. Some languages make it easy and painless to write programs that only require a few lines of code, e.g. task automation. But those languages often don't have enough power to solve large problems, e.g. GUI programming. Conversely, languages that are powerful enough for big problems often require far too much overhead for small problems. This characteristic is important because problems that look small at first frequently grow in scope in unexpected ways. If a programmer chooses a language appropriate only for small tasks, scope changes can require rewriting code from scratch in a new language. And if the programmer chooses a language with lots of overhead and friction to solve a problem that stays small, it will be harder for other people to use and understand than necessary. Rewriting code that works fine is the single most wasteful thing a programmer can do with their time, but using a bazooka to kill a mosquito instead of a flyswatter isn't good either. Here are some of the ways this characteristic presents itself. Can be used interactively - there is some environment where programmers can enter commands one by one Requires no more than one file - neither project files nor makefiles are required for running in batch mode Can easily split code across multiple files - files can refeence each other, or there is some support for modules Has good support for data structures - supports structures like arrays, lists, and especially classes Supports a wide variety of features - features like networking, serialization, XML, and database connectivity are supported by standard libraries Here's my take on how C#, Python, and shell scripting measure up. Python scores highest. Feature C# Python shell scripting --------------- --------- --------- --------------- Interactive poor strong strong One file poor strong strong Multiple files strong strong moderate Data structures strong strong poor Features strong strong strong Is there a term that captures this idea? If not, what term should I use? Here are some candidates. Scalability - already used to decribe language performance, so it's not a good idea to overload it in the context of language syntax Granularity - expresses the idea of being good just for big tasks versus being good for big and small tasks, but doesn't express anything about data structures Smoothness - expresses the idea of low friction, but doesn't express anything about strength of data structures or features Note: Some of these properties are more correctly described as belonging to a compiler or IDE than the language itself. Please consider these tools collectively as the language environment. My question is about how easy or difficult languages are to use, which depends on the environment as well as the language.

    Read the article

  • Are there deprecated practices for multithread and multiprocessor programming that I should no longer use?

    - by DeveloperDon
    In the early days of FORTRAN and BASIC, essentially all programs were written with GOTO statements. The result was spaghetti code and the solution was structured programming. Similarly, pointers can have difficult to control characteristics in our programs. C++ started with plenty of pointers, but use of references are recommended. Libraries like STL can reduce some of our dependency. There are also idioms to create smart pointers that have better characteristics, and some version of C++ permit references and managed code. Programming practices like inheritance and polymorphism use a lot of pointers behind the scenes (just as for, while, do structured programming generates code filled with branch instructions). Languages like Java eliminate pointers and use garbage collection to manage dynamically allocated data instead of depending on programmers to match all their new and delete statements. In my reading, I have seen examples of multi-process and multi-thread programming that don't seem to use semaphores. Do they use the same thing with different names or do they have new ways of structuring protection of resources from concurrent use? For example, a specific example of a system for multithread programming with multicore processors is OpenMP. It represents a critical region as follows, without the use of semaphores, which seem not to be included in the environment. th_id = omp_get_thread_num(); #pragma omp critical { cout << "Hello World from thread " << th_id << '\n'; } This example is an excerpt from: http://en.wikipedia.org/wiki/OpenMP Alternatively, similar protection of threads from each other using semaphores with functions wait() and signal() might look like this: wait(sem); th_id = get_thread_num(); cout << "Hello World from thread " << th_id << '\n'; signal(sem); In this example, things are pretty simple, and just a simple review is enough to show the wait() and signal() calls are matched and even with a lot of concurrency, thread safety is provided. But other algorithms are more complicated and use multiple semaphores (both binary and counting) spread across multiple functions with complex conditions that can be called by many threads. The consequences of creating deadlock or failing to make things thread safe can be hard to manage. Do these systems like OpenMP eliminate the problems with semaphores? Do they move the problem somewhere else? How do I transform my favorite semaphore using algorithm to not use semaphores anymore?

    Read the article

  • Is throwing an error in unpredictable subclass-specific circumstances a violation of LSP?

    - by Motti Strom
    Say, I wanted to create a Java List<String> (see spec) implementation that uses a complex subsystem, such as a database or file system, for its store so that it becomes a simple persistent collection rather than an basic in-memory one. (We're limiting it specifically to a List of Strings for the purposes of discussion, but it could extended to automatically de-/serialise any object, with some help. We can also provide persistent Sets, Maps and so on in this way too.) So here's a skeleton implementation: class DbBackedList implements List<String> { private DbBackedList() {} /** Returns a list, possibly non-empty */ public static getList() { return new DbBackedList(); } public String get(int index) { return Db.getTable().getRow(i).asString(); // may throw DbExceptions! } // add(String), add(int, String), etc. ... } My problem lies with the fact that the underlying DB API may encounter connection errors that are not specified in the List interface that it should throw. My problem is whether this violates Liskov's Substitution Principle (LSP). Bob Martin actually gives an example of a PersistentSet in his paper on LSP that violates LSP. The difference is that his newly-specified Exception there is determined by the inserted value and so is strengthening the precondition. In my case the connection/read error is unpredictable and due to external factors and so is not technically a new precondition, merely an error of circumstance, perhaps like OutOfMemoryError which can occur even when unspecified. In normal circumstances, the new Error/Exception might never be thrown. (The caller could catch if it is aware of the possibility, just as a memory-restricted Java program might specifically catch OOME.) Is this therefore a valid argument for throwing an extra error and can I still claim to be a valid java.util.List (or pick your SDK/language/collection in general) and not in violation of LSP? If this does indeed violate LSP and thus not practically usable, I have provided two less-palatable alternative solutions as answers that you can comment on, see below. Footnote: Use Cases In the simplest case, the goal is to provide a familiar interface for cases when (say) a database is just being used as a persistent list, and allow regular List operations such as search, subList and iteration. Another, more adventurous, use-case is as a slot-in replacement for libraries that work with basic Lists, e.g if we have a third-party task queue that usually works with a plain List: new TaskWorkQueue(new ArrayList<String>()).start() which is susceptible to losing all it's queue in event of a crash, if we just replace this with: new TaskWorkQueue(new DbBackedList()).start() we get a instant persistence and the ability to share the tasks amongst more than one machine. In either case, we could either handle connection/read exceptions that are thrown, perhaps retrying the connection/read first, or allow them to throw and crash the program (e.g. if we can't change the TaskWorkQueue code).

    Read the article

  • How To: Spell Check InfoPath web form in SharePoint 2010

    - by Jeremy Ramos
    Originally posted on: http://geekswithblogs.net/JeremyRamos/archive/2013/11/07/how-to-spell-check-infopath-web-form-in-sharepoint-2010.aspxThis is a sequel to my 2011 post about How To: Spell Check InfoPath Web Form in SharePoint. This time I will share how I managed to achieve Spell Checking in SharePoint 2010. This time round, we have changed our Online Forms strategy to use Custom lists instead of Form Libraries. I thought everything will be smooth sailing as we are using all OOTB features. So, we customised a Custom list form using InfoPath and added a few Rich Text Boxes (Spell Check is a requirement for this specific project). All is good in the InfoPath client including the Spell Checker so, happy days, I published straight away.Here comes the surprises now. I browsed to my Custom List and clicked Add New Item. This launched my Form in a modal dialog format. I went to my Rich Text Boxes to check the spell checker, and voila, it's disabled!I tried hacking the FormServer.aspx and the CustomSpellCheckEntirePage.js again but the new FormServer.aspx behaves differently than of MOSS 2007's. I searched for answers in many blogs to no avail. Often ending up being linked to my old blog post. I also tried placing the spell check javascript into a Content Editor Webpart of the Item's New Form and Edit form. It is launching the Spell Check dialog but it's not spellchecking the page correctly.At this point, I decided I needed to get my project across ASAP so enough with experimentations and logged a ticket with Microsoft Premier Support.On a call with the Support Engineer, I browsed through the Custom List and to the item to demonstrate my problem. Suddenly, the Spell Check tab in the toolbar is now Enabled! Surprised? Not much, it's Microsoft!Anyway, to cut my story short, here is a summary of my solution:Navigate to your Custom ListIn the Ribbon Toolbar, navigate to List > Customize List > Form Web Parts > Content Type Forms > (Item) New Form. This will display the newifs.aspx which is the page displayed when Add New Item is clicked. This page, just like any other SharePoint page, contains webparts. In this case, we have the InfoPath Form Web Part.Add a Content Editor Web Part (CEWP) on top of the InfoPath Form Web Part. (A blank CEWP would do for this example)Navigate to Page and click Stop EditingClick Add New Item again and navigate to a Rich Text box. Tadah! The Spell Check tab is now enabled!Do the same steps for the (Item) Edit Form to enable Spell Checks when editing an item.This "no code" solution discovered purely by accident!

    Read the article

  • Your interesting code tricks/ conventions? [closed]

    - by Paul
    What interesting conventions, rules, tricks do you use in your code? Preferably some that are not so popular so that the rest of us would find them as novelties. :) Here's some of mine... Input and output parameters This applies to C++ and other languages that have both references and pointers. This is the convention: input parameters are always passed by value or const reference; output parameters are always passed by pointer. This way I'm able to see at a glance, directly from the function call, what parameters might get modified by the function: Inspiration: Old C code int a = 6, b = 7, sum = 0; calculateSum(a, b, &sum); Ordering of headers My typical source file begins like this (see code below). The reason I put the matching header first is because, in case that header is not self-sufficient (I forgot to include some necessary library, or forgot to forward declare some type or function), a compiler error will occur. // Matching header #include "example.h" // Standard libraries #include <string> ... Setter functions Sometimes I find that I need to set multiple properties of an object all at once (like when I just constructed it and I need to initialize it). To reduce the amount of typing and, in some cases, improve readability, I decided to make my setters chainable: Inspiration: Builder pattern class Employee { public: Employee& name(const std::string& name); Employee& salary(double salary); private: std::string name_; double salary_; }; Employee bob; bob.name("William Smith").salary(500.00); Maybe in this particular case it could have been just as well done in the constructor. But for Real WorldTM applications, classes would have lots more fields that should be set to appropriate values and it becomes unmaintainable to do it in the constructor. So what about you? What personal tips and tricks would you like to share?

    Read the article

  • Library to fake intermittent failures according to tester-defined policy?

    - by crosstalk
    I'm looking for a library that I can use to help mock a program component that works only intermittently - usually, it works fine, but sometimes it fails. For example, suppose I need to read data from a file, and my program has to avoid crashing or hanging when a read fails due to a disk head crash. I'd like to model that by having a mock data reader function that returns mock data 90% of the time, but hangs or returns garbage otherwise. Or, if I'm stress-testing my full program, I could turn on debugging code in my real data reader module to make it return real data 90% of the time and hang otherwise. Now, obviously, in this particular example I could just code up my mock manually to test against a random() routine. However, I was looking for a system that allows implementing any failure policy I want, including: Fail randomly 10% of the time Succeed 10 times, fail 4 times, repeat Fail semi-randomly, such that one failure tends to be followed by a burst of more failures Any policy the tester wants to define Furthermore, I'd like to be able to change the failure policy at runtime, using either code internal to the program under test, or external knobs or switches (though the latter can be implemented with the former). In pig-Java, I'd envision a FailureFaker interface like so: interface FailureFaker { /** Return true if and only if the mocked operation succeeded. Implementors should override this method with versions consistent with their failure policy. */ public boolean attempt(); } And each failure policy would be a class implementing FailureFaker; for example there would be a PatternFailureFaker that would succeed N times, then fail M times, then repeat, and a AlwaysFailFailureFaker that I'd use temporarily when I need to simulate, say, someone removing the external hard drive my data was on. The policy could then be used (and changed) in my mock object code like so: class MyMockComponent { FailureFaker faker; public void doSomething() { if (faker.attempt()) { // ... } else { throw new RuntimeException(); } } void setFailurePolicy (FailureFaker policy) { this.faker = policy; } } Now, this seems like something that would be part of a mocking library, so I wouldn't be surprised if it's been done before. (In fact, I got the idea from Steve Maguire's Writing Solid Code, where he discusses this exact idea on pages 228-231, saying that such facilities were common in Microsoft code of that early-90's era.) However, I'm only familiar with EasyMock and jMockit for Java, and neither AFAIK have this function, or something similar with different syntax. Hence, the question: Do such libraries as I've described above exist? If they do, where have you found them useful? If you haven't found them useful, why not?

    Read the article

  • What can a Service do on Windows?

    - by Akemi Iwaya
    If you open up Task Manager or Process Explorer on your system, you will see many services running. But how much of an impact can a service have on your system, especially if it is ‘corrupted’ by malware? Today’s SuperUser Q&A post has the answers to a curious reader’s questions. Today’s Question & Answer session comes to us courtesy of SuperUser—a subdivision of Stack Exchange, a community-driven grouping of Q&A web sites. The Question SuperUser reader Forivin wants to know how much impact a service can have on a Windows system, especially if it is ‘corrupted’ by malware: What kind malware/spyware could someone put into a service that does not have its own process on Windows? I mean services that use svchost.exe for example, like this: Could a service spy on my keyboard input? Take screenshots? Send and/or receive data over the internet? Infect other processes or files? Delete files? Kill processes? How much impact could a service have on a Windows installation? Are there any limits to what a malware ‘corrupted’ service could do? The Answer SuperUser contributor Keltari has the answer for us: What is a service? A service is an application, no more, no less. The advantage is that a service can run without a user session. This allows things like databases, backups, the ability to login, etc. to run when needed and without a user logged in. What is svchost? According to Microsoft: “svchost.exe is a generic host process name for services that run from dynamic-link libraries”. Could we have that in English please? Some time ago, Microsoft started moving all of the functionality from internal Windows services into .dll files instead of .exe files. From a programming perspective, this makes more sense for reusability…but the problem is that you can not launch a .dll file directly from Windows, it has to be loaded up from a running executable (exe). Thus the svchost.exe process was born. So, essentially a service which uses svchost is just calling a .dll and can do pretty much anything with the right credentials and/or permissions. If I remember correctly, there are viruses and other malware that do hide behind the svchost process, or name the executable svchost.exe to avoid detection. Have something to add to the explanation? Sound off in the comments. Want to read more answers from other tech-savvy Stack Exchange users? Check out the full discussion thread here.

    Read the article

  • Data Structures usage and motivational aspects

    - by Aubergine
    For long student life I was always wondering why there are so many of them yet there seems to be lack of usage at all in many of them. The opinion didn't really change when I got a job. We have brilliant books on what they are and their complexities, but I never encounter resources which would actually give a good hint of practical usage. I perfectly understand that I have to look at problem , analyse required operations, look for data structure that does them efficiently. However in practice I never do that, not because of human laziness syndrome, but because when it comes to work I acknowledge time priority over self-development. Over time I thought that when I would be better developer I will automatically use more of them - that didn't happen at all or maybe I just didn't. Then I found that the colleagues usually in the same plate as me - knowing more or less some three of data structures and being totally happy about it and refusing to discuss this matter further with me, coming back to conversations about 'cool new languages' 'libraries that do jobs for you' and the joy to work under scrumban etc. I am stuck with ArrayLists, Arrays and SortedMap , which no matter what I do always suffice or either I tweak them to be capable of fulfilling my task. Yes, it might be inefficient but do we really have to care if Intel increases performance over years no matter if we improve our skills? Does new Xeon or IBM machines really care what we use? What if I like build things, but I am not particularly excited whether it is n log(n) or just n? Over twenty years the processing power increased enormously, which gives us freedom of not being critical about which one to use? On top of that new more optimized languages appear which support multiple cores more efficiently. To be more specific: I would like to find motivational material on complex real areas/cases of possible effective usages of data structures. I would be really grateful if you would provide relevant resources. There is similar question ,but in the end the links again mostly describe or do dumb example(vehicles, students or holy grail quest - yes, very relevant) them and people keep referring to the "scenario decides the data structure to use". I want to know these complex scenarios to be able to identify similarities to my scenario and then use them. The complex scenarios where it really matters and not necessarily of quantitive nature. It seems that data structures only concern is efficiency and nothing else? There seems to be no particular convenience for developer in use one over another. (only when I found scientific resources on why exactly simple carbohydrates are evil I stopped eating sugar and candies completely replacing it with less harmful fruits - I hope you can see the analogy)

    Read the article

  • Referencing a picture in another DLL in Silverlight and Windows Phone 7

    - by Laurent Bugnion
    This one has burned me a few times, so here is how it works for future reference: Usually, when I add an Image control into a Silverlight application, and the picture it shows is local (as opposed to loaded from the web), I set the picture’s Build Action to Content, and the Copy to Output Directory to Copy if Newer. What the compiler does then is to copy the picture to the bin\Debug folder, and then to pack it into the XAP file. In XAML, the syntax to refer to this local picture is: <Image Source="/Images/mypicture.jpg" Width="100" Height="100" /> And in C#: return new BitmapImage(new Uri( "/Images/mypicture.jpg", UriKind.Relative)); One of the features of Silverlight is to allow referencing content (pictures, resource dictionaries, sound files, movies etc…) located in a DLL directly. This is very handy because just by using the right syntax in the URI, you can do this in XAML directly, for example with: <Image Source="/MyApplication;component/Images/mypicture.jpg" Width="100" Height="100" /> In C#, this becomes: return new BitmapImage(new Uri( "/MyApplication;component/Images/mypicture.jpg", UriKind.Relative)); Side note: This kind of URI is called a pack URI and they have been around since the early days of WPF. There is a good tutorial about pack URIs on MSDN. Even though it refers to WPF, it also applies to Silverlight Side note 2: With the Build Action set to Content, you can rename the XAP file to ZIP, extract all the files, change the picture (but keep the same name), rezip the whole thing and rename again to XAP. This is not possible if the picture is embedded in an assembly! So what’s the catch? Well the catch is that this does not work if you set the Build Action to Content. It’s actually pretty simple to explain: The pack URI above tells the Silverlight runtime to look within an assembly named MyOtherAssembly for a file named MyPicture.jpg in the Images folder. If the file is included as Content, however, it is not in the assembly. Silverlight does not find it, and silently returns nothing. The image is not displayed. And the fix? The fix, for class libraries, is to set the Build Action to Resource. With this, the picture will gets packed into the DLL itself. Of course, this will increase the size of the DLL, and any change to the picture will require recompiling the class library, which is not ideal. But in the cases where you want to distribute pictures (icons etc) together with a plug-in assembly, well, this is a good way to have everything in the same place Happy coding, Laurent   Laurent Bugnion (GalaSoft) Subscribe | Twitter | Facebook | Flickr | LinkedIn

    Read the article

  • Developer – Cross-Platform: Fact or Fiction?

    - by Pinal Dave
    This is a guest blog post by Jeff McVeigh. Jeff McVeigh is the general manager of Performance Client and Visual Computing within Intel’s Developer Products Division. His team is responsible for the development and delivery of leading software products for performance-centric application developers spanning Android*, Windows*, and OS* X operating systems. During his 17-year career at Intel, Jeff has held various technical and management positions in the fields of media, graphics, and validation. He also served as the technical assistant to Intel’s CTO. He holds 20 patents and a Ph.D. in electrical and computer engineering from Carnegie Mellon University. It’s not a homogenous world. We all know it. I have a Windows* desktop, a MacBook Air*, an Android phone, and my kids are 100% Apple. We used to have 2.5 kids, now we have 2.5 devices. And we all agree that diversity is great, unless you’re a developer trying to prioritize the limited hours in the day. Then it’s a series of trade-offs. Do we become brand loyalists for Google or Apple or Microsoft? Do we specialize on phones and tablets or still consider the 300M+ PC shipments a year when we make our decisions on where to spend our time and resources? We weigh the platform options, monetization opportunities, APIs, and distribution models. Too often, I see developers choose one platform, or write to the lowest common denominator, which limits their reach and market success. But who wants to be ?me too”? Cross-platform coding is possible in some environments, for some applications, for some level of innovation—but it’s not all-inclusive, yet. There are some tricks of the trade to develop cross-platform, including using languages and environments that ?run everywhere.” HTML5 is today’s answer for web-enabled platforms. However, it’s not a panacea, especially if your app requires the ultimate performance or native UI look and feel. There are other cross-platform frameworks that address the presentation layer of your application. But for those apps that have a preponderance of native code (e.g., highly-tuned C/C++ loops), there aren’t tons of solutions today to help with code reuse across these platforms using consistent tools and libraries. As we move forward with interim solutions, they’ll improve and become more robust, based, in no small part, on our input. What’s your answer to the cross-platform challenge? Are you fully invested in HTML5 now? What are your barriers? What’s your vision to navigate the cross-platform landscape?  Here is the link where you can head next and learn more about how to answer the questions I have asked: https://software.intel.com/en-us Republished with permission from here. Reference: Pinal Dave (http://blog.sqlauthority.com)Filed under: PostADay, SQL, SQL Authority, SQL Query, SQL Server, SQL Tips and Tricks, T SQL Tagged: Intel

    Read the article

  • NuGet JustMock

    - by mehfuzh
    As most of us already know JustMock got  a free edition. The free edition is not a stripped down of the features of the full edition but I would rather say its a strip down of the type you can mock. Technically, free version runs on  proxy as full version runs on proxy + profiler. In full version, It switches to profiler when you are mocking final methods or sealed class or anything else that can not be done using inheritance. Like in full version you can mock non public methods , in free version you can still do it but it has to be virtual for protected or must be done through InternalsVisibleTo attribute for internal virtual methods (If you have access to the source and can apply the attribute). Now, you can get a copy of free edition from the product page. Install it and off you go. But it is also exposed to NuGet. Those of you are not familiar with NuGet (that will be odd). But still NuGet is the centralized package manager from Microsoft that cuts the workflow of manual inclusion of  libraries in your project. I think NuGet in future will limit the scope of  “.vsi” packages and installers because of its ease (except in some cases). Its similar to ruby gems. In ruby, virtually you can install any library in this way “gems  install <target_library>” and you are off to go. It will check the dependencies, install them or less prompt with the steps you need to do.   Now sticking to the post, to get started you first need to install NuGet package manager. Once you have completed the step pressing “Ctrl + W, Ctrl + Z” it will bring up an console like one below:   Once you are here, you just have to type “install-package justmock” Next, it will should print the confirmation when the installation is complete: Moving to visual studio solution explorer, you will now see:   Finally, NuGet is still in its early ages and steps that are shown here may not remain the same in coming releases, but feel free to enjoy what is out there right now. Regarding JustMock free edition, there is a nice post by Phil Japikse at Introducing JustMock Free Edition. I think its worth checking if not already.   Have fun and happy holidays!

    Read the article

< Previous Page | 171 172 173 174 175 176 177 178 179 180 181 182  | Next Page >