[gdal-dev] Alpine, gcc 32-bit in Linux Build Actions

[Abel Pau]

After some investigations, I've concluded that I was doing two things that I've improved lately:
	1) Addressing a memory leak (in a non-typical case in my previous tests). CORRECTED
	2) I was requesting an amount of memory suitable for ogr2ogr of large layers but not for numerous translations on a machine with limited memory. I've reduced the amount of memory required by /10. And now Alpine 32 bits is in green colour!

There are still some tests to review, but I'm optimistic about submitting a pull request soon!


-----Mensaje original-----
De: gdal-dev <gdal-dev-bounces at lists.osgeo.org> En nombre de Abel Pau via gdal-dev
Enviado el: divendres, 1 de març de 2024 16:18
Para: Even Rouault <even.rouault at spatialys.com>; gdal-dev at lists.osgeo.org
Asunto: Re: [gdal-dev] Alpine, gcc 32-bit in Linux Build Actions

Yes, I hope it's only  ONE single mistake because I have my code very protected against that kind of things...
Before you said that I am using LOG_STR(); from the test and I've filled it in all possible place it could be something.

Remember I have windows and in Visual Studio it works fine...
I don't know how to debug in my Docker linux (only command line) with Cmake compiler...

I think I'll catch the problem. Sooner or later!

I've been there before (years ago with a big problem (not my fault)). Finally I win

-----Mensaje original-----
De: Even Rouault <even.rouault at spatialys.com> Enviado el: divendres, 1 de març de 2024 16:12
Para: Abel Pau <a.pau at creaf.uab.cat>; gdal-dev at lists.osgeo.org
Asunto: Re: [gdal-dev] Alpine, gcc 32-bit in Linux Build Actions

Your code must be something terrible so that even gdb doesn't catch it
:-) (SIGKILL cannot be caught...)

 From the error messages, things might go wrong starting with
https://github.com/OSGeo/gdal/blob/master/apps/test_ogrsf.cpp#L782

So maybe set a breakpoint at that line

b test_ogrsf.cpp:782

and then use "step", and "next" to single step and locate where this crashes. This is going to be a bit tedious...

(you might also want to modify slightly test_ogrsf so that this
TestCreateLayer() method exists early when !EQUAL(poDriver->GetDescription(), "miramon") to avoid debugging other
drivers)

You might install "ddd", as a GUI front-end for gdb, so that this is slightly more user friendly.

-- 

http://www.spatialys.com
My software is free, but my time generally not.

_______________________________________________
gdal-dev mailing list
gdal-dev at lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/gdal-dev