Re: Bug in cygport? debuginfo package skipped even though a debug file is present

On 23/04/2026 14:14, Jon Turney wrote:
> On 20/04/2026 16:27, Hamish McIntyre-Bhatty via Cygwin-apps wrote:
>> Hi there,
>>
>> I think I might have come across a bug in cygport around the handling 
>> of the debuginfo packages. For pycairo, I've found that it says it's 
>> skipping the debuginfo package even though it later reports there's a 
>> file not in any package, which is a debug file. The cygport file can 
>> be found at: https://gitlab.com/hamishmb/cygwin-python-pycairo
>>
>> The full output is attached.
>>
>> Any help would be greatly appreciated, assuming this isn't a bug, as 
>> I am a bit lost right now.
> Interesting,
>
> So "No debug files, skipping debuginfo subpackage" really means "While 
> enumerating the source files referenced by the debuginfo, I didn't 
> find any starting with the expected /usr/src/debug/ and copy them 
> there from $S (or $B, in the case of generated files)"
>
> (It's arranged for these pathnames to start with /usr/src/debug/ by 
> passing -ffile-prefix-map to the compiler in CFLAGS etc. This avoids 
> capturing the specific working directory for the package build into 
> the built debuginfo package)
>
> So this is perhaps a meson-python build integration problem (assuming 
> it's taking note of these flags, it still looks like it's creating its 
> own temporary subdirectory to do the build in, which seems to have 
> been removed by the time we'd want to do this).
>
> I'm guessing there's a way to make this work with meson-python, as 
> it's otherwise useless with packaging systems which generally want to 
> do something like this, but it'll probably take some digging to find 
> out what that is...
>
>
> As a workaround for the moment, you could list the .dbg file in 
> PKG_IGNORE [1], which should allow you to generate a package (albeit 
> without debuginfo).
>
> [1] https://cygwin.github.io/cygport/pkg_pkg_cygpart.html#PKG_IGNORE

That's fair enough, mostly I just wanted to make you aware of the problem.

I'll go ahead and do that and then submit an ITP.

Best,

Hamish