Re: make alters subprocess environment

This seems a make-w32 problem, which is better addressed to the make-w32 
mailing list http://lists.gnu.org/mailman/listinfo/make-w32

Kees Zeelenberg

----- Original Message ----- 
From: "Павел Сутырин" <[email protected]>
To: <[email protected]>
Sent: Wednesday, September 19, 2007 2:36 PM
Subject: [GnuWin32-Users] make alters subprocess environment


>
>
> Hi all!
>
> Here is what I'm interested in -- make *seems to alter subprocess 
> environment somehow (current dir and variables supposedly). Here is the 
> story.
>
> For testing of a Delphi5 application under WindowsXP I've established 
> following call chain (on my machine make is renamed to gmake):
>
> gmake.exe -> testApp.exe -> ..\..\db_comparator\db_comparator.py <some 
> params>
>
> The latter is a separate utility intended to test the equality of some 
> database tables during tests, a thing written in python just for the case.
>
> I've discovered, that CreateProcess() (used in testApp.exe for calls to 
> python utility) cannot parse relative paths to executables. Then I've 
> modified chain:
>
> gmake.exe -> testApp.exe -> db_comparator.bat -> 
> ..\..\db_comparator\db_comparator.py <some params>
>
> where db_comparator.bat is in the same dir with testApp.exe and obviously 
> contains:
>
> ..\..\db_comparator.py %*
>
> this configuration seemed to work, but ONLY if i run testApp.exe manually. 
> When gmake runs it during all the tests -- it fails. Manual rerun (from 
> cmd) of db_comparator.bat or ..\..\db_comparator.py with the same 
> parameters is successful.
>
> Summarizing:
>
> gmake.exe -> testApp.exe -> db_comparator.bat -> 
> ..\..\db_comparator\db_comparator.py <some params>
>  FAILS
>
> testApp.exe -> db_comparator.bat -> ..\..\db_comparator\db_comparator.py 
> <some params>
>  WORKS
>
> Another issue connected. If I want interrupt a make run by pressing Ctrl-C 
> at the cmd terminal, windows reports: "gmake.exe -- error encountered. The 
> application will be closed", and when i close this window, at the termnial 
> make says:
>
>  gmake: *** [test] Error -1073741510
>
> Moreover, this sequence (an error window and a line at terminal) happens 
> once for every nested gmake, that was active at the moment of 
> interruption, so it may result in this:
>
> gmake[3]: *** [testApp.exe] Error -1073741510
> gmake[2]: *** [build] Error -1073741819
> gmake[1]: *** [test] Error -1073741819
> gmake: *** [test] Error -1073741819
>
> These two issues seem to be connected with the way make controls its 
> subprocesses. What is known?..
>
> APPENDIX:
>
>  C:\>gmake --version
>  GNU Make version 3.75, by Richard Stallman and Roland McGrath.
>
> Thank you for reading all this length,
>
> --pavel.
>
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2005.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> GnuWin32-Users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/gnuwin32-users
>
>
> 


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
GnuWin32-Users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gnuwin32-users