On 8/15/26 12:44 PM, Rafael Laboissière wrote:
* Sebastiaan Couwenberg <[email protected]> [2026-08-14 07:46]:
On 8/14/26 6:53 AM, Sebastiaan Couwenberg wrote:
On 8/13/26 8:21 PM, Bas Couwenberg wrote:
The pycontainer patch seems to be the culprit:
"
I can't reproduce the issue with SWIG 4.5.0 built from official SWIG
tarballs, but looking at
https://salsa.debian.org/debian/swig/-/blob/debian/master/debian/patches/include-pycontainer.patch
, I believe it is the reason of the failure as it unconditionnally includes a
SWIG .swg file only meant for C++. This patch to the SWIG package looks wrong
to me
"
Dropping include-pycontainer.patch (e.g. as per the attached patch) is
confirmed the fix the issue.
Has the issue been discussed with upstream or the patch forwarded? They likely
can suggest a better fix.
I don't know if it's possible to make the include conditional for C++, that
might be a good alternative too.
The attached patch does that and also resolves the GDAL build failure.
Thanks for you patch. I committed it at your name [1] and released version
4.5.0-2 of the package. I hope there will not be other collateral effects.
I hope so too, that's why it's good to forward the patch upstream to get their
feedback, there might be better ways to fix the issue reported in #1024555.
Either of these patches should be forwarded upstream to get their input on the
best solution.
As you are already in contact with the upstream authors regarding this issue,
could you please forward your patch to them?
To clarify, I was in contact with GDAL upstream, not SWIG upstream.
I'd prefer if you forward the patch to SWIG upstream as you're in a better
position to coordinate with them, I'd rather not be the man in the middle.
If you are unable or unwilling to upstream the patch I can do that despite the
drawbacks.
Kind Regards,
Bas
--
PGP Key ID: 4096R/6750F10AE88D4AF1
Fingerprint: 8182 DE41 7056 408D 6146 50D1 6750 F10A E88D 4AF1
