We are a Microsoft shop, concentrated on using C#. We have several projects, including websites, Windows services, and class libraries, that incorporate XML comments.
I'm looking to generate MSDN-style HTML documentation for each project and deploy it to a centralized location that all of the developers can easily access. I also want to automate these steps so they can be run at regular intervals so I, and the other developers, don't need to worry about remembering to generate and deploy new documentation whenever a change is made. I've looked at Sandcastle and Doxygen and both look to be good options for generating the documentation that I want, but I need advice on a good way to automate the generation of it, like in a nightly job or something.
Anyone out there doing something like this? I'm not sold on the end result being HTML; especially if there is a better idea.
EDIT:
I appreciate all of the good ideas. There are a couple of routes now that I can investigate, but I won't know which will work best until I get my hands dirty. The Sandcastle Help File Builder seems to offer me the best options for what I'm looking to do, so I'll give the nod to that suggestion. However, if I had more time to work on the XSLT and CSS solution to get the XML data looking just right, I'd pursue that suggestion first.
Thanks again everyone!
Take a look at Sandcastle Help File Builder. This uses standalone projects, so you can build them as often as you like (for example nightly, or as part of your continuous integration system every time a change is checked in).
You don't need any third-party tools to generate nice documentation: C# compiler can output the documentation from XML-comments in XML and all you need is to design a nice CSS to show it in browser. If you are not satisfied, you can also generate your own XSLT-transoformation before applying CSS, as described here (look for an example of CSS in comments!).
Alternatively you can take this XML documentation and improve it using something like NDoc. Here is a nice article about how to do it, unfortunately somewhat outdated.
You can manually build only the documentation of your project into a single XML file by running the compiler with /doc:documentationfilename.xml parameter.
You can also indicate in Visual Studio (Project Properties -> Documentation) to add this option to every build of your project, so that XML file will be generated every time you build your project. You can then hook post-build events to copy the XML file to your repository.
You can also set up your MS Build on your Team Foundation Server to build your documentation and copy it to the repository in a similar way (/p:DocumentationFile=fileName.xml);
I do that with our projects. Basically we use doxywizard to set up the configuration Doxyfile, which will set the specifications for the generated html. Then, in a build server step I invoke "doxygen doxyfile".
Our Doxyfile is configured to generate the doxygen files in an area visible from our webserver. So, each commit to trunk causes documentation to be rebuilt automatically.
Related
My team has a fairly large set of desktop applications with many shared libraries between them all in one common solution file in our repository. We'd like to use semantic versioning for a number of reasons, chief of which is to make it easier for our users to install updates. However, given the number of assemblies we're dealing with, we're finding it pretty tedious updating AssemblyInfo files for each one, especially if it's for a library that's a dependency for multiple applications.
I was wondering if there's an easy way to use git tags or some kind of external tool to tell the build server that, for example, XYZ has a bug fix and its patch number needs to be updated.
Use GitVersion : https://gitversion.readthedocs.io/en/latest/
It will do automatically the semantic versionning based on the last tag and git history.
You could use GitVersionTask if you use msbuild or (better) use it with build tools like fake or cake.net
Edit: you now have also alternatives easier to use : https://www.nuget.org/packages/Nerdbank.GitVersioning/, https://www.nuget.org/packages/GitInfo/,. ..
I am trying to automatically open an NDepend Project when the Solution builds in an automated build in TFS2010.
This stems from this previous question. The aforementioned post is where I tried (and failed) to integrate NDepend's code metrics software with an automated Team Build via messing with the XML of my solution.
I decided since I wasn't getting anywhere in messing with the XML, that I would try a different route. In another program I have developed, I used
System.Diagnostics.Process.Start("blah.txt");
to trigger Notepad to run and open the text file "blah.txt."
I figured I could use the same concept to possibly help me with this NDepend integration. So I researched MSDN to see if I can find out more about the Process.Start method. And using this example
Process.Start("IExplore.exe", "C:\\myPath\\myFile.htm");
I substituted in my own paths to what I believe should open the project file "myProj.ndproj" inside the VisualNDepend application like this
System.Diagnostics.Process.Start("C:\\tools\\NDepend\\VisualNDepend","C:\\myProj\\myProj.ndproj");
I may be taking that example and tweaking it out of context, I'm not sure, but it seemed to me that what I tried should work. The solution built fine without any errors, but VisualNDepend didn't run.
It finally hit me that I was trying to use this code that would only execute when the program ran when I really need it to execute when the program builds within TFS and Visual Studio.
I asked my coworkers if they knew of any built-in ways within TFS or VS that would recognize whether or not the solution was being built or not. And they didn't really know of anything in particular. I tried "Googling" this topic and couldn't find any information that was useful to me.
Does anyone know of how to accomplish this? Or am I chasing a lost cause by trying to execute some C# code behind the solution? In which case, is my best bet trying to tweak the XML like I had previously been attempting?
I would recommend writing a custom build task (or tasks). You can essentially make the task do anything you'd like -- run a process, spit out results, etc., and it can be invoked directly from your MSBuild script.
I'm not sure if I'm answering your question (or if I even have a grasp on what you're trying to do), but that's probably the area I'd be looking to find my solution.
I have a project , there is several classes. I want to search used and not used properties in project.
is visual studio have any way to search ueed properties.
Programing language is C#.net
ReSharper
I suppose you have to use some third party tools available in the market.Just Google it and you will find many.I recommend resharper which gives you many added functionality
Find all references.
Just use the "Run Code Analysis" option. It will show the unused calls (if they are not public and available to everyone.
Even if code analysis tools might help in searching this properties, you can't rely on code analysis tools completely, because a client code might access a property with reflection.
This includes data contracts your client proxies might have in case of WCF, javasript json serialization dependencies and the like.
You can try comment out properties suspected to have no usages, but you have not only to make sure it compiles, but also have all the the tests to pass.
Hope you have a decent test coverage and intergated test suites in place.
I have a number of legacy web controls (ascx) that contains huge amounts of inline C#. The forms contain a number of repeated and duplicate code. Our first plan is to move the code into code behinds per file, then refactor etc... were doing this to upgrade the client to the latest version of their cms
At the moment we are going to have to manually copy and paste from hundreds of files, create a code behind, copy the code, add the namespaces based on the client-side imports and then do any tidying up
does anybody PLEASE know of a tool that can do the majority of this work for us ?
Thanks
The DMS Software Reengineering Toolkit is designed to support the the construction of automated mass change tools.
It can parse HTML, and C#, (and build corresponding ASTs) and could be configured parse .ascx files. With the parse trees, one could likely generate your desired results.
Is this easy? No. is is practical? Yes, if the amount of code you are reengineering is significant, as it seems in your case.
I think your best hope here is to create a custom conversion utility. I don't know of any tools that would do this out of the box.
Use a language like Perl or Rebol which can do parsing. Rebol is the easiest.
Resharper has many features for moving and locating code for different refactorings. I don't have it installed on the machine I'm at right now, so can't verify, but it may be able to at least speed up the process and make it more full proof, even if it can't be fully automated.
I think your best bet would be to look at the Codesmith tools, or just T4 templates in general..you shouldn't need to write a custom app...though it looks like someone on CodeProject did just that : http://www.codeproject.com/KB/aspnet/InlineCodeVSCodeBehind.aspx?msg=997793
I'm a newbie to SSIS / C# (I'm generally a Java developer) so apologies if this is a really stupid question.
Essentially the problem is this: I have two Data Flow tasks which load data up and export them to a legacy flat file format. The formatting is done by a Script Task (C#).
What I'd like to do is share some common code between the two. e.g. I could create a common base class and then extend it for my two different script tasks.
However it seems that SSIS doesn't really make provision for this.
Does anyone know if there is a way of accomplishing what I want to do?
You're correct that there is not a straightforward way to do this directly from SSIS.
In a recent project, we took two different approaches, which both worked fairly well depending on what you need to do:
Create a utility class (as a simple class library) and reference it from your script tasks. This is done pretty much the same as any other sort of reference. If you use .NET 3.5, remember that you'll have to update the version manually in the script tasks since SSIS defaults to 2.0. We also found that if we wanted some manner of reusability in the utility assembly (not relying on hardcoded variable names, etc.) then the package still had to have a fairly large amount of "setup" boilerplate to use the utility scripts.
Create a custom data flow component. This is a much more involved process, but ultimately will do the best in terms of avoiding code duplication. Generally, coding the actual data flow is fairly simple and not that much different than a script component, but the various setup code you'll need can tend to make things complicated. There's also not a lot of support in SSIS for when something goes wrong. Led to a lot of detective work on our project.
If you plan on using something a whole lot, and are committed to getting rid of boilerplate code as much as possible, 2 is the preferred option. If it's being used a few places here and there, consider the simple approach of 1.
I am pretty sure it's possible to access .NET assemblies in SSIS scripts. So you could do it this way. See the article "Accessing .NET assemblies with SSIS" on SQL Server Central.
I believe you will have to create an assembly or webservice for this to work.
This does not completely solve your issue but it does help in not having to recreate all the classes every time you need them (I also do not want to deploy referenced assemblies for my current project ). Firstly you need a master copy of your classes, you can copy them from an existing Script Task using the same process below but in reverse.
Open the Editor for the Script Task and on the Property Explorer click on the Project File (the st_[Guid] ), in the Properties window you’ll see the Project Folder location. (This location gets recreated every time you edit the script task)
In explorer, copy your classes to this folder
On the Project Explorer, click on the “Show All Files” icon
Right click on your files and add to Project
Probably way too late to answer this, but you can click on the solution and add a class there. Then when you go into your scripts you can say add existing object and search for that class you created earlier. For me it was located by the solution for the project. Haven't gone through the deployment or anything for this, but at least you can access the class through the individual scripts.