Bug#1144963: hugin FTBFS with swig 4.5.0

[Andreas Metzler]

Control: tags -1 pending

On 2026-08-20 Adrian Bunk <bunk at debian.org> wrote:
> Source: hugin
> Version: 2025.0.1+dfsg-2
> Severity: serious
> Tags: ftbfs

> https://buildd.debian.org/status/fetch.php?pkg=hugin&arch=arm64&ver=2025.0.1%2Bdfsg-2%2Bb2&stamp=1787084882&raw=0

> ...
> /build/reproducible-path/hugin-2025.0.1+dfsg/src/hugin_script_interface/hpi_classes.cpp: In member function ‘int hpi::python_interface::call_hpi(const char*, PyObject*)’:
> /build/reproducible-path/hugin-2025.0.1+dfsg/src/hugin_script_interface/hpi_classes.cpp:115:22: error: ‘PyInt_AsLong’ was not declared in this scope; did you mean ‘PyLong_AsLong’?
>   115 |             result = PyInt_AsLong ( pValue ) ;
>       |                      ^~~~~~~~~~~~
>       |                      PyLong_AsLong
> ...

Hello Adrian,

there is an patch in upstream HG that fixes this, I habe vcherrypicked
it in GIT.

However hugin currently FTBFS due to the openexr/imath breakage
(#1144763 - Thanks for your NMU). I have tried building hugin against
the pending NMU openxer version but it fails with:

[ 22%] Linking CXX executable celeste_standalone
cd /tmp/HUGIN/hugin/obj-x86_64-linux-gnu/src/celeste && /usr/bin/cmake -E cmake_link_script CMakeFiles/celeste_standalone.dir/link.txt --verbose=1
/usr/bin/x86_64-linux-gnu-ld.bfd: /usr/lib/x86_64-linux-gnu/libvigraimpex.so: undefined reference to `Imf_3_4::Header::Header(Imath_3_1::Box<Imath_3_1::Vec2<int> > const&, Imath_3_1::Box<Imath_3_1::Vec2<int> > const&, float, Imath_3_1::Vec2<float> const&, float, Imf_3_4::LineOrder, Imf_3_4::Compression)'
collect2: error: ld returned 1 exit status

Rebuilding libvigraimpex against the new openexr let the hugin build
succeed.

cu Andreas
-- 
"You people are noisy," Nia said.
I made the gesture of agreement.