test/functional/test_image.py::TestCompareXLSXFiles::test_image58 failure on 32bit architectures about libxlsxwriter

[jmcnamara]

Thanks for the report. I'll look into it.

[hosiet]

I see changes in 42700ea , which has failed CI tests.

If the variable to be printed has fixed length (e.g., in64_t), the printf format string should use functionalities in <inttype.h> with the PRId64 macro. See https://stackoverflow.com/questions/9225567/how-to-portably-print-a-int64-t-type-in-c for more explanation. Using %ld or %lld are not portable across different platforms, and will cause build failure on one platform or another. Of course, we need to pay attention to whether lxw_snprintf() is compatible with such format strings, though.

In the meanwhile, the dirty hack of Ignore %lld warning with gcc in xmlwriter.c. in src/Makefile shall be removed.

[jmcnamara]

I see changes in 42700ea , which has failed CI tests.

That was just a WIP attempt before I went to bed. :-)

See https://stackoverflow.com/questions/9225567/how-to-portably-print-a-int64-t-type-in-c for more explanation.

I don't think any of those suggestions work with ANSI-C (whether I should still try to support that is a separate question).

Anyway, I went with a fix based on a double conversion/format. That will work in the integer range required for those large uint64_t offsets. The test case that was failing is more or less the maximum value for that attribute.

I tested it on a 32bit ubuntu system. Could you try it when you get a chance.

[hosiet]

Thanks, I will try it and the rebuild result for all architectures will be available in 24 hours.

[jmcnamara]

Thanks, I will try it and the rebuild result for all architectures will be available in 24 hours.

Ok. Let me know if you need it in a new release.

[jmcnamara]

I have added a 32bit github action to hopefully catch issues like this in future.

[hosiet]

Thanks, the new patch seems to have fixed the bug.

I don't mind having a new release; if you feel like postponing the release, I can always backport the patch only.

[jmcnamara]

I don't mind having a new release; if you feel like postponing the release, I can always backport the patch only.

I'm still accumulating fixes for a new release so it is probably best to work with the patch/fix for now. I'll let you know when there is an update.