Search Results

Search found 25797 results on 1032 pages for 'source formatting'.

Page 17/1032 | < Previous Page | 13 14 15 16 17 18 19 20 21 22 23 24  | Next Page >

  • Why do some open source projects do not allow to report issues in a public issue tracker?

    - by linquize
    Why do some open source projects do not allow to report issues in a public issue tracker? Those projects requires the issues to be reported via email, and the issues may be forwarded to people in mailing list. Users may repeatedly report the same issue if there is no public issue tracker, as they have no easy way to know what have been reported before. The project team members need to spend extra time answering those repeated issues. Some projects do have a public issue tracker, but the issues are still reported through email and they are posted only by the project team only after filtering. It does not allow anyone to report directly in issue tracker. (example: SVN) Such arrangement is not transparent nor open, which I think it violates the philosophy of open source. And it is outdated.

    Read the article

  • How do you find partners for open source projects?

    - by static_rtti
    I've created a few open-source projects in the past, and had some success. The process was generally the following: I'd start alone, create something that works, promote it, and finally (maybe) get some contributions. I have less that to contribute to open-source now, but I still have ideas and can still code :) So I wonder if there is a place or a way to meet people (online), discuss ideas and design, and then start working together on the project? The reason I'm thinking about an online way to do this, with strangers, is that while I do have programmer friends, we only very rarely have the same needs and interests at the same time. It seems to me that it would be easier to find such in match in the online global community. Any insight?

    Read the article

  • Why do some open-source projects NOT have a public issue tracker?

    - by linquize
    Why do some open source projects not allow to report issues in a public issue tracker? Those projects require the issues to be reported via email, and the issues may be forwarded to people in a mailing list. Users may repeatedly report the same issue if there is no public issue tracker, as they have no easy way to know what has been reported before. The project team members need to spend extra time answering those repeated issues. Some projects do have a public issue tracker but the issues are still reported through email and they are posted by the project team only after filtering. It does not allow anyone to report directly in issue tracker (example: SVN). Such arrangement is not transparent nor open, which I think violates the philosophy of open source. And it is outdated.

    Read the article

  • How to go about "taking over" an open-source project?

    - by LuxuryMode
    There's an open-source project that I'm interested in and use regularly. It's licensed under the Apache License 2.0 and it has basically no activity any more. It's hosted on Google Code and I'm interested in continuing it's development. I'm new to the open-source process and I'm trying to figure out the appropriate way to go about this. Can I just check it out and push it to github so I can continue it's development in the open there? Should I contact the project "owner" first? Also, do I leave all the author information at the top of the classes, etc even though I'm going to be making changes..(I'm assuming the answer is yes)? Also, how do I practically adhere to the license requirement of "all modifications are clearly marked as being the work of the modifier"? Do I place a comment by every change I make?

    Read the article

  • Ready to Open source the software, but want to make some money out of it.

    - by byteofprash
    I'm an UG Student and an Entrepreneur. I have started a small company with which I'm trying to provide solutions, to various companies in my city, through my software. I put in 30-40 hours of coding and development. I've been a huge fan of Open Source projects and I've also built by software on python. But now when it comes to selling my software to my customers, I expect money. (Am I being a Hypocrite?) Is there any OSS license under which I can make money and also make my source open? Please suggest. Or any other way to make money with my softwares? (I guess I deserve some credit. If not the code, atleast the ideas? :D )

    Read the article

  • On what name should I claim copyright in open source software?

    - by ONOZ
    When I want to use the Apache 2.0 licence in my project, I should include this in the comments of my source code: Copyright [yyyy] [name of copyright owner] Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. What name should I fill in for [name of copyright owner]? I am currently working alone on this project, but I'm going to release the source code so there might be other contributors in the near future.

    Read the article

  • Netbeans source code formatting

    - by lucky
    I am Netbeans user for PHP. I have installed Netbeans IDE 6.8. Everything is working well. There is an option in SOURCE -- FORMAT. This action formats the code properly. It is quite handy. Unfortunately, this functionality is not working for me. Whenever some php code comes in between, it additionally adding some empty rows in between and making it more untidy. Can you please help to set up the same functionality. Restore the settings to default.

    Read the article

  • Data Source Security Part 1

    - by Steve Felts
    I’ve written a couple of articles on how to store data source security credentials using the Oracle wallet.  I plan to write a few articles on the various types of security available to WebLogic Server (WLS) data sources.  There are more options than you might think! There have been several enhancements in this area in WLS 10.3.6.  There are a couple of more enhancements planned for release WLS 12.1.2 that I will include here for completeness.  This isn’t intended as a teaser.  If you call your Oracle support person, you can get them now as minor patches to WLS 10.3.6.   The current security documentation is scattered in a few places, has a few incorrect statements, and is missing a few topics.  It also seems that the knowledge of how to apply some of these features isn’t written down.  The goal of these articles is to talk about WLS data source security in a unified way and to introduce some approaches to using the available features.  Introduction to WebLogic Data Source Security Options By default, you define a single database user and password for a data source.  You can store it in the data source descriptor or make use of the Oracle wallet.  This is a very simple and efficient approach to security.  All of the connections in the connection pool are owned by this user and there is no special processing when a connection is given out.  That is, it’s a homogeneous connection pool and any request can get any connection from a security perspective (there are other aspects like affinity).  Regardless of the end user of the application, all connections in the pool use the same security credentials to access the DBMS.   No additional information is needed when you get a connection because it’s all available from the data source descriptor (or wallet). java.sql.Connection conn =  mydatasource.getConnection(); Note: You can enter the password as a name-value pair in the Properties field (this not permitted for production environments) or you can enter it in the Password field of the data source descriptor. The value in the Password field overrides any password value defined in the Properties passed to the JDBC Driver when creating physical database connections. It is recommended that you use the Password attribute in place of the password property in the properties string because the Password value is encrypted in the configuration file (stored as the password-encrypted attribute in the jdbc-driver-params tag in the module file) and is hidden in the administration console.  The Properties and Password fields are located on the administration console Data Source creation wizard or Data Source Configuration tab. The JDBC API can also be used to programmatically specify a database user name and password as in the following.  java.sql.Connection conn = mydatasource.getConnection(“user”, “password”); According to the JDBC specification, it’s supposed to take a database user and associated password but different vendors implement this differently.  WLS, by default, treats this as an application server user and password.  The pair is authenticated to see if it’s a valid user and that user is used for WLS security permission checks.  By default, the user is then mapped to a database user and password using the data source credential mapper, so this API sort of follows the specification but database credentials are one-step removed from the application code.  More details and the rationale are described later. While the default approach is simple, it does mean that only one database user is doing all of the work.  You can’t figure out who actually did the update and you can’t restrict SQL operations by who is running the operation, at least at the database level.   Any type of per-user logic will need to be in the application code instead of having the database do it.  There are various WLS data source features that can be configured to provide some per-user information about the operations to the database. WebLogic Data Source Security Options This table describes the features available for WebLogic data sources to configure database security credentials and a brief description.  It also captures information about the compatibility of these features with one another. Feature Description Can be used with Can’t be used with User authentication (default) Default getConnection(user, password) behavior – validate the input and use the user/password in the descriptor. Set client identifier Proxy Session, Identity pooling, Use database credentials Use database credentials Instead of using the credential mapper, use the supplied user and password directly. Set client identifier, Proxy session, Identity pooling User authentication, Multi Data Source Set Client Identifier Set a client identifier property associated with the connection (Oracle and DB2 only). Everything Proxy Session Set a light-weight proxy user associated with the connection (Oracle-only). Set client identifier, Use database credentials Identity pooling, User authentication Identity pooling Heterogeneous pool of connections owned by specified users. Set client identifier, Use database credentials Proxy session, User authentication, Labeling, Multi-datasource, Active GridLink Note that all of these features are available with both XA and non-XA drivers. Currently, the Proxy Session and Use Database Credentials options are on the Oracle tab of the Data Source Configuration tab of the administration console (even though the Use Database Credentials feature is not just for Oracle databases – oops).  The rest of the features are on the Identity tab of the Data Source Configuration tab in the administration console (plan on seeing them all in one place in the future). The subsequent articles will describe these features in more detail.  Keep referring back to this table to see the big picture.

    Read the article

  • How to implement multi-source XSLT mapping in 11g BPEL

    - by [email protected]
    In SOA 11g, you can create a XSLT mapper that uses multiple sources as the input. To implement a multi-source mapper, just follow the instructions below, Drag and drop a Transform Activity to a BPEL process Double-click on the Transform Activity, the Transform dialog window appears. Add source variables by clicking the Add icon and selecting the variable and part of the variable as needed. You can select multiple input variables. The first variable represents the main XML input to the XSL mapping, while additional variables that are added here are defined in the XSL mapping as input parameters. Select the target variable and its part if available. Specify the mapper file name, the default file name is xsl/Transformation_%SEQ%.xsl, where %SEQ% represents the sequence number of the mapper. Click OK, the xls file will be opened in the graphical mode. You can map the sources to the target as usual. Open the mapper source code, you will notice the variable representing the additional source payload, is defined as the input parameter in the map source spec and body<mapSources>    <source type="XSD">      <schema location="../xsd/po.xsd"/>      <rootElement name="PurchaseOrder" namespace="http://www.oracle.com/pcbpel/po"/>    </source>    <source type="XSD">      <schema location="../xsd/customer.xsd"/>      <rootElement name="Customer" namespace="http://www.oracle.com/pcbpel/Customer"/>      <param name="v_customer" />    </source>  </mapSources>...<xsl:param name="v_customer"/> Let's take a look at the BPEL source code used to execute xslt mapper. <assign name="Transform_1">            <bpelx:annotation>                <bpelx:pattern>transformation</bpelx:pattern>            </bpelx:annotation>            <copy>                <from expression="ora:doXSLTransformForDoc('xsl/Transformation_1.xsl',bpws:getVariableData('v_po'),'v_customer',bpws:getVariableData('v_customer'))"/>                <to variable="v_invoice"/>            </copy>        </assign> You will see BPEL uses ora:doXSLTransformForDoc XPath function to execute the XSLT mapper.This function returns the result of  XSLT transformation when the xslt template matching the document. The signature of this function is  ora:doXSLTransformForDoc(template,input, [paramQName, paramValue]*).Wheretemplate is the XSLT mapper nameinput is the string representation of xml input, paramQName is the parameter defined in the xslt mapper as the additional sourceparameterValue is the additional source payload. You can add more sources to the mapper at the later stage, but you have to modify the ora:doXSLTransformForDoc in the BPEL source code and make sure it passes correct parameter and its value pair that reflects the changes in the XSLT mapper.So the best practices are : create the variables before creating the mapping file, therefore you can add multiple sources when you define the transformation in the first place, which is more straightforward than adding them later on. Review ora:doXSLTransformForDoc code in the BPEL source and make sure it passes the correct parameters to the mapper.

    Read the article

  • Pure C# open source PCM to Ogg convertor?

    - by Ole Jak
    Microsoft Silverlight 4 is in beta. It supports PCM audio output. It would be madness to stream PCM over internet (for ex in P2P chart webApp) so we need Pure C# open source PCM to Ogg convertor. No unmanaged code, nothing going out of .net sandbox. So does any one know such Pure C# open source PCM to Ogg convertor? What do I need: Open Source Libs for encoding. Tutorials and blog articles on How to do it, about etc. BTW: why Pure C#? - because Silverlight 4 does not support unmanaged or just not C# DLL's. BTW2: this question is similar to this one but it is different because Ogg is Open Source, free while mp3 will not be free until 2010

    Read the article

  • Difference between putting variables in header vs putting variables in source

    - by Mohit Deshpande
    Say I declare a header file with a variable: int count; Then in the source file, I want to use count. Do I have to declare it as: extern int count Or can I just use it in my source file? All assuming that I have #include "someheader.h". Or should I just declare it in the source file? What is the difference between putting count in the header file vs the source file? Or does it not matter?

    Read the article

  • eclipse debugger: attaching source-code of maven managed libraries

    - by meriton
    I'd like to use the source code of maven-managed dependencies when debugging our webapp in eclipse. I have managed to attach the sources to the libraries in the "Maven Managed Depedencies" classpath container, i.e. when I open a class file from a depedency (e.g. using Ctrl-Shift-T), I see the source code. However, when I launch the tomcat within eclipse, and execution halts on a breakpoint in that same class, the editor pane only displays the text "source not found", and a button to edit the "source lookup path". I have attempted to manually add the "Maven Managed Dependencies" classpath container, only to be told "Use maven project settings to configure depedency resolution". However, I see no useful setting in that property pane ... How can I attach those sources?

    Read the article

  • What is your contribution to open source projects?

    - by Yuval A
    I was always wondering about this seemingly utopic world of open source. Assuming the vast majority of users here are professional software engineers which need some sort of income source, I assume most of us hold stable, money-making jobs. So who are the key players in the open source community? Who are the people which devote their precious time to these projects? What is their benefit? Are the majority just people who see a bug, fix it, submit, and forget about the project? Or are they people constantly involved in the process of building the product? How do you find yourself contributing to open source projects?

    Read the article

  • .Net Library for parsing source code files?

    - by Jörg Battermann
    Does anyone know of a good .NET library that allows me to parse source code files, but not only .NET source code files (like java, perl, ruby, etc)? I need programmatic access to the contents of various source code files (e.g. class/method /parameter names, types, etc.). Has anyone come across something like this? I know within .NET it is reasonably possible and there are some libraries out there, but I need that to be abstracted to more types of programming languages.

    Read the article

  • Visual Studio - Attach Source Code to Reference

    - by Joe
    My C# project references a third-party DLL for which I have the source code. Can I somehow tell Visual Studio the location of that source code, so that, for example, when I press F12 to open the definition of a method in the DLL, it will open up the source code, instead of opening up the "Class [from metadata]" stub code?

    Read the article

  • How to build Lucene / Solr from source code in windows environment in order to add patches

    - by Simon
    I have successfully implemented Apache’s Solr for free text searching a database driven web site build for windows platforms using Visual Studio in c#. I am trying to get a version Solr working with field collapsing (which is not in the release version). There are patches available from apache and discussions on the web of people successfully doing this for the version I am using but my problem is cannot get the build to work. I am a c# coder on windows platforms so java development is new to me. I understand I need to get the correct source code (and revision) from SVN, add the appropriate patches, then build the war file to deploy to my system. I cannot seem to get the source to build and produce the deployment code including jar (and subsequent war) files. My system is: Windows 7 Ultimate for development Visual Studio 2010 for c# / javascript development MyEclipse 8.6 / Eclipse 3.5 for the java build from source Subecplise 1.6x SVN plugin to get the source from apache’s SVN Apache Solr 1.4.1 So far I have: Found the right patches for the function I need: https://issues.apache.org/jira/browse/SOLR-236 Specifically I need to patch: field_collapsing_1.1.0.patch HTTPS //issues.apache.org/jira/secure/attachment/12357681/field_collapsing_1.1.0.patch and SOLR-236-1_4_1.patch HTTPS //issues.apache.org/jira/secure/attachment/12448216/SOLR-236-1_4_1.patch I downloaded the Lucene trunk version from the day before the patch was released (revision 958303 from 28/6/10) via subeclipse into a java package in myeclipse from: HTTPS //svn.apache.org/repos/asf/lucene/dev/trunk (Solr is the web implementation of Lucene and is in the subfolder solr/) I can apply patches to the solr directory once it has downloaded but the parent Lucene project doesn’t build the war files, copy the jar or other files into the bin folder (it stays empty). The build process starts, but doesn’t do anything apart from creating the folders bin and src. I am building the whole Lucene project, which contains Solr. I have tried building the source without patching and the same happens. If I copy out the Solr directory into a new project, it runs the build and copies all the related files, tests, etc but fails with 4,500 errors and does not produce the jar files or war file, which I assume is because it can’t find the Lucene trunk files which it depends on. I have two interrelated problems 1) I can't get the Lucene downloaded trunk to build 2) The jar, war and associated files are not created Can anyone help with what I am missing to build the war file? I have spent 2 days to get this far as the help online is extremely patchy and I can’t find a walk though tutorial on building a java war file from source in a windows environment. Any help will be much appreciated. Simon

    Read the article

  • Java is open-source, so what?

    - by typoknig
    Hi all, I always here that Java being open-source is a big benefit, but I fail to see how Java being open-source should draw me to use it as opposed to .NET which is closed-source. This website has some Q&A sections (What is the significance of these developments to the industry? in particular) that give a little info, but is being free the only (or the biggest) advantage to Java being open-source? Since I am a beginner, have any of you pros noticed any major difference since the change was made?

    Read the article

  • How do I edit error messages in GCC source

    - by user299274
    Hi, I am trying to create a C Compiler in my native language which I intend to put up as open source. I want to do this by downloading the GCC source code and then manually translating the error messages and warnings into my target language. I am a beginner to GCC. Any idea where the error messages are located in the source code and how can I edit them?

    Read the article

  • Open Source Project & Language Selection

    - by James
    I'm getting ready to start an open-source project that will target .NET/Mono. For those who have started their own open source venture... Do you let the fact that a project is going to be open-source weigh on the decision of what language to use? For example. Most .NET open-source projects are written in C#. However, if you were more comfortable with VB.NET, Boo, Nemerle, etc... would you use it? What other considerations are there? This particular project will be a core library and application for geocaching. Similar to GSAK.

    Read the article

  • Source operator doesn't work in if construction in bash

    - by Igor
    Hello, I'd like to include a config file in my bash script with 2 conditions: 1) the config file name is constructed on-the-fly and stored in variable, and 2) in case if config file doesn't exist, the script should fail: config.cfg: CONFIGURED=yes test.sh: #!/bin/sh $CFG=config.cfg echo Source command doesn't work here: [ -f $CFG ] && ( source $CFG ) || (echo $CFG doesnt exist; exit 127) echo $CONFIGURED echo ... but works here: source $CFG echo $CONFIGURED What's wrong in [...] statement?

    Read the article

  • Is there an Open Source licence matrix?

    - by tobsen
    I am trying to find a simple matrix overview of the various licenses around open source software but can't find anything useful. I would like to know whether software under a certain licence can be used in commercial closed source software, if there has to be the original source code included, if the licence has to be redistirbuted, etc. I know this is a difficult topic because I guess only the minority of programmers are also moonlighting lawyers - neither am I.

    Read the article

  • Why would LaTeX ignore the font size in the documentclass

    - by Rory
    I have a LaTeX file. I'm experimenting with trying to reduce the font size (this is related to my other question here http://stackoverflow.com/questions/2636647/latex-changing-the-font-size-for-a-document-but-in-the-preamble-not-the-docum ). The LaTeX file is generated from another programme. I have edited it to start with \documentclass[4pt,a4paper,english]{report} i.e. I am trying to make the text really small. However it doesn't work. I change that 4pt to anything and the font size is the same. When running pdflatex on it, I get this message printed out. LaTeX Warning: Unused global option(s): [4pt]. That might explain why the error message is What could be going on here? How do I make it use the font size in the documentclass definition?

    Read the article

< Previous Page | 13 14 15 16 17 18 19 20 21 22 23 24  | Next Page >