New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
I cannot build GDLMM (hardcoded D:/a/_temp path to m4.exe) #9615
Comments
You can see the procedure about how to build that project here https://github.com/msys2/MINGW-packages/tree/master/mingw-w64-gdlmm2. See the PKGBUILD file. |
I apologize for any confusion but I had installed that package and it seemed it was not the same version. Thus my attempt to build from source. That package seems to build gdlmm-1.0 but my project is expecting gdlmm-3.0. Are they the same? The .pc file fails my cmake pkgconfig_check for gdlmm. Also reading the PKGBUILD file I am not sure where the variables are coming from as I do not have experience with building packages on mingw. I think I'll review the patch it tried to apply and see if that helps me doing it manually. I will report back. |
To acidtonic did you update your MSys2 packages?
Tim S. |
No, the gdlmm2 package is version 2 not version 3. But, the package might help you. To get rid of the error you posted you need to first update your msys2 packages because that issue has likely been fixed. Tim S. |
I'm having a similar problem (different program, but same hardcoded path, and same idea of trying to compile something locally that is normally installed via pacman) -- a search of /mingw64 for this path yields over 5000 hits. If memory serves, that looks like the path that GitHub Actions uses for its builds, so maybe this is a problem with the build artifacts? |
See also: #9411 |
I ran pacman -Syuu and only had 3 packages updated (pacman itselt, a tty library, and something else non-related). Still have the same issue if I try again with a fresh checkout. Is there a way to specify the path to m4 or edit a makefile or something? I find it odd the path gets embedded inside an EXE file that generates code. |
The other issue is similar but the suggested fixes are for missing symbols which I do not have a problem with. I do see this path if I run gcc --print-search-dirs and I'd presume it's in other places as well. |
Alright well I decided I couldn't wait so I started going deep into this... Managed to hack it and got it to build.
Changed to
Now she builds. So perl inside glibmm is trying to do magical things to get the m4 path and it's flat out WRONG. Have a nice day :) |
Agree. This seems to be an issue of glibmm package. Some files of that package contain full path reference of build environment. e.g.
|
What command did you use to build the gdlmm project? In my case,
|
./autogen.sh |
Also in case it matters, at one point I ran an autoreconf in there per another suggestion via a different resource. |
Can anyone test if the pull request (#9731) fixes this issue? |
I have mingw64 on windows64 trying to build gdlmm and I cannot get the autotools code generation to work. It keeps building a path to a non-existent D drive for the m4 command which then fails and the build cannot complete.
$ clone git repo from -> https://github.com/Fabo/gdlmm.git
$ ./autogen.sh
$ ./configure
$ make
(various entering directory lines)
sh: line 1: D:/a/_temp/msys64/usr/bin/m4.EXE: No such file or directory
m4 failed with exit code 127. Aborting...
I believe this is related to the issue #7481 with hardcoded paths but in that case it with Python, but for me it's M4. Any help or workarounds would be appreciated. I have a binary release build to drop shortly and this is a blocker.
Cheers
The text was updated successfully, but these errors were encountered: