Debian Patches

Status for wxmaxima/26.07.1+ds-10

Patch Description Author Forwarded Bugs Origin Last update
upstream-fix-commit-e8fd25c.patch Fix two 32-bit-only bugs found by GH #2208, and run tests on 32-bit CI

CellPtr's cmpObjects() subtracted two unrelated Observed* pointers
directly to get an ordering; that is undefined behavior and could
silently produce the wrong sign once the address space is only 32 bits
wide, which the much sparser 64-bit address space essentially never
triggers. Compare the addresses as unsigned integers instead.

EditorCell::FindNext() and its Tab-key indent handler stored a
wxString::find()/rfind() result (size_t) into a signed 64-bit variable
and compared it against wxNOT_FOUND. On 64-bit builds size_t is also
64 bits, so wxString::npos's bit pattern happens to reinterpret as -1;
on 32-bit builds it zero-extends into a large positive number instead,
making every failed search look like a match at a huge offset. Keep
both in size_t/wxString::npos terms.

Both were only caught by actually running the unit test suite on a
real i386 cross-build -- CI never did that before, it only compiled
the tests. compile_ubuntu_32bit.yml now runs `ctest -L unittest` under
a virtual X server (most of these tests instantiate real GTK widgets),
matching the pattern already used for compile_windows.yml.		 Claude <noreply@anthropic.com> no 2026-08-06
upstream-fix-XDG_CONFIG_HOME-inappropriate_permissions.patch upstream: fix: XDG_CONFIG_HOME: inappropriate permissions It appears that `wxmaxima` makes the XDG_CONFIG_HOME directory with
inappropriate permissions when this directory does not exist. This
patch resolves this issue. Actually the upstream source somehow
made this directory with the good permissions but they were typed
in hexadecimal. Thanks to Daniel Lange <DLange@debian.org> for
spotting the inappropriate code
( <https://lists.debian.org/debian-devel/2026/08/msg00006.html> ).
A closer look reaveals the issue repeated itself for TEMPDIR.		 Jerome Benoit <calculus@rezozer.net> invalid debian 2026-08-07
upstream-test-fix-testbench_simple.patch upstream: test: fix: testbench_simple The testbench_simple test bench currently emits some `Error' messages.
This patch solves the Error messages related to the absence of the
`{a,b,c,d}.png` images by introducing `TESTBENCH_FILES` list in the
concerned `test/CMakeLists.txt` along some cleaning/rationalization
(wrt to the style of cmake script used by upstream).		 Jerome Benoit <calculus@rezozer.net> yes debian 2026-07-25
upstream-test-harden-multithreadtest.patch upstream: test: harden: multithreadtest The multithreadtest test currently fails due to a concurrency issue
associated with loading and compiling of the `draw` maxima package.
This patch solves this issue by preloading the `draw` maxima package.		 Jerome Benoit <calculus@rezozer.net> yes debian 2026-07-25
debianization.patch debianization Meant to maintain a minimal debian/rules, to fix warnings,
to address Debian specific stuff in general.		 Jerome Benoit <calculus@rezozer.net> not-needed debian 2026-08-02
debianization-tests-WXM_TEST_TIMEOUT.patch debianization; tests: WXM_TEST_TIMEOUT This patch adjusts the WXM_TEST_TIMEOUT duration
as some exotic architectures supported by Debian
appear to be slower than upstream computers.
This patch is an adhocery: it would be better to pass
adquate WXM_TEST_TIMEOUT durations via an option
(or something).		 Jerome Benoit <calculus@rezozer.net> not-needed debian 2026-08-04
debianization-tests-relax.patch debianization; tests: relax This patch relaxes a very minor check. Jerome Benoit <calculus@rezozer.net> not-needed debian 2026-07-24
debianization-tests-ImageFormatTest-adhoc.patch debianization: tests: imageFormat test: ad hoc This patch corrects a misnamed issue in the image-test.
The path was not submitted to upstream, but the issue was
reported.		 Jerome Benoit <calculus@rezozer.net> yes debian 2026-07-25

All known versions for source package 'wxmaxima'

Links