I have duplicate classes into a used library. How can I specify on the ecf file that I want to exclude a path?
Error code: VD71
Configuration error
Duplicate class: BASIC
/home/pg/data/solarity/sit-dev/eiffel/lib/thirdparty/json/examples/performance/basic.e
/home/pg/data/solarity/sit-dev/eiffel/lib/thirdparty/json/examples/basic/basic.e
Configuration: /home/pg/data/solarity/sit-dev/eiffel/sit-platform/sit_platform_lib.ecf
one way I found can be
<library name="json" location=".\..\lib\thirdparty\json\library\json.ecf" readonly="false"/>
<file_rule>
<exclude>
/examples$
</exclude>
</file_rule>
But can't find the definition of excluding paths and its syntax
There is no mechanism to control source code outside a library. The file rules to include and exclude source files are supported only for clusters.
You need to fix the library configuration to make sure is has no duplicate classes.
Note. Project settings dialog is a recommended way to change project settings, because it reflects only supported options, and ensures the configuration file does not get broken. In particular, it shows that clusters have exclusion rules in Source code section, whereas libraries do not.
Related
I am using Visual Studio (2017) and I need the following.
I have a folder where a code generator puts the .h and .c files obtained from a formal model. This folder is not controlled by me, e.g. I cannot write in it, but it is updated by another team member.
By using the /I compiler options (or Additional include directories in the project properties) I managed to import all the generated header files in my VS project. What I am supposed to do is to integrate this generated code into a specific platform, this means that I have to compile both the generated code and the integration code on the target platform. The problem is, the compiler is not able to resolve the generated function definitions of the generated code as it only sees the .h files. What I got is a linking error (external symbol not resolved)
To solve the problem, I added the existing .c files manually, one by one. The obvious problems that comes with this solution are
manual boring work
when new files are generated, I need to manually import the new files
Question is: is there an option that can be set in order to specify the path of the source files without passing them one by one?
note: just copying and pasting the generated code in the VS project folder is not an acceptable solution.
Thanks
If you look at https://learn.microsoft.com/en-us/cpp/ide/working-with-project-properties?view=vs-2017 then you see there is a Source Directories property that has $(VC_SourcePath) as a default but (I think) to which you can add additional paths. The documentation is unclear whether that means all source files in such a path will be included for compilation.
At the bottom of the documentation it explains how to override certain project properties by providing an external properties file. It seems you can override the targets/sources using such a file. You can generate the file using a small tool that reads the filenames in those directories and adds them to the file.
You could also analyze the .vcproj file and build a small tool that wil re-write the part with your generated source directories, reading the filenames in those directories and adding them to the section in the .vcproj file.
I've been attempting to modify my project's .csproj file, that will include one of two different resource(resx) files that will depend on what the build configuration is at the current time.
I've tried to use conditionals to differentiate the inclusion, something like:
Then proceeded to add the files that were located in different directories. However, when building I've either gotten errors or the build just ignored the conditionals and added the default resx files.
I was hoping someone might know how to conditionally add resx files depending on the current build configuration?
Thank you!
I need to use tcl.h in my C project (VS 2012) for registering function in SQLite. I have downloaded Tcl for windows and installed it. But my project still doesn't recognize when I write #include . What should I do to get access to this header?
Thanks in advance.
Open the project properties and do :
Right click on your project, go to properties, go to C/C++ session. On the field Additional Include Directories, add the path to the header you want, and use the header with "".
After that, you should go to the Linker session. On General, go to the Additional Library directories and add the path to the libs you need to add. Then go to Input and add the lib.
Sometimes you will need to copy a dll to the same folder of your executable. You can do that on Custom Build Step - Post Build Events.
You need to do it for each configuration and platform you want to build - one tip to make it easier is to use those variables $(Configuration) and $(Platform) when you are specifying the paths, and select on the dropdown list of the top of the properties window All configurations and All platforms, so you don't need to do it several times per project.
I am working with a MATLAB project, and would like to be able to configure variables and paths without re-creating the executable (currently done by using MATLAB's built-in deploytool). I realize I can create a simple text file, listing key=value pairs, and read that in during run-time. There doesn't seem to be any built-in support in MATLAB for creating configuration files that follow a standard format.
Are there any standard existing ways to use configuration files for MATLAB-generated executables?
1) Add your resource to the package in DeployTool in the shared resources part
2) Read by using:
fullfile(ctfroot(),'config.txt');
Edit : That is suitable if your configuration is "private", that is the user should not edit it. In this case it is deployed together with your program as a resource. If you want a "public" config for the users, I know of no better way than to do what you offered.
I can't find a way to use the source server tools from the Debugging Tools for Windows on a static library project,
which is built separately from the solutions actually using that library:
The output of "ssindex.cmd" always displays "zero source files found" for the PDB file generated for the library
(using compiler options /ZI and /Fd).
Running "srctool.exe -r" on this PDB displays nothing, which probably means that the PDB file does not contain any source file information.
Running the same command on the PDB file of a test application which is also build as part of the the same solution
yields a list of all expected source files.
Is there a way to use source indexing for a static library project when it should be built seperately from the solutions using it?
Thanks for any suggestions!
You can use the "/Save" and "/Load" options to store and load source information for a static library, respectively. Using these options allows you to store information for your library and then later import it when indexing a project that links against your library.
When indexing your library solution, you specify the "/Save" flag with a directory in which to store index information about the library's source files. For example (assuming you are using Subversion for source control),
ssindex.cmd /System=SVN /Save=c:\source\libproj\srcinfo /Source=c:\source\libproj /Symbols=c:\source\libproj\Release\*.pdb
When later indexing your project that includes your library, you specify the "/Load" flag with the directory containing the library's source file information. For example,
ssindex.cmd /System=SVN /Load=c:\source\libproj\srcinfo /Source=c:\source\binproj /Symbols=c:\source\binproj\Release\*.pdb
There are two potential issues that may affect your ability to use this technique. First, it appears that some source control providers may not support saving and loading source control information. I know that the Subversion provider does and it looks like the SourceSafe provider does, but I haven't checked any others.
Second, this technique appears to only work for one external static library out-of-the-box. There does not seem to be a way to load information from multiple directories and the scripts currently overwrite the contents of the directory each time you use the "/Save" option. You could probably edit the source control provider module to append to the files in the save directory rather than overwrite them, but I have not tried it.
Also, note as you mentioned above that you only need to do this if your library is being built as part of a separate solution. If the static library is part of the solution you are indexing, its source files will be included if they are in the path specified by the "/Source" option.
It probably means you haven't inputed the correct directories when running "ssindex" so for ssindex you need to have: /source=C:/SourceCode/ /symbols=C:/SourceCode/bin/Debug I'm not sure if the "source" has an upper case S or not but that should be it!
when run svnindex.cmd, it always tell you "zero source files found"
after a painful diggin into svn.pm (the perl module to deal with svn), i found that:
first, svn.pm invokes "svn info -R $SourceRoot" to get all version info of files in $SourceRoot (passed by /source option),
then svn.pm stores all files in a dictionary which using the local file path as key
svnindex.cmd call srctool -r to get all source files info in *.pdb, and use the source file name as a key to query info saved in step2
the problem is:
svn.pm uses relative path, but *.pdb uses absolute path, so you will never find a svn log info for any file, then "zero source files found"
fixup:
change svn.pm line 162:
$LocalFile = lc $1;
to
$LocalFile = $SourceRoot . "\" . lc $1; #make path absolute