Debian Patches
Status for tango/10.3.3+dfsg1-4
| Patch | Description | Author | Forwarded | Bugs | Origin | Last update |
|---|---|---|---|---|---|---|
| 0001-assets-CMakeLists.txt-Use-new-package-root-variable-.patch | [PATCH 1/2] assets/CMakeLists.txt: Use new package root variable to override location From 5192f5ff7a95cae1ae22f2452eb39fd4ba924794 Mon Sep 17 00:00:00 2001 Since the move to using find_package in cppTango, ca0754ba (Merge branch 'use-find_package-for-deps' into 'release/9.5.0', 2023-09-14), our hack to override the tangoidl include directory does not work anymore. This was hidden as the tsd/debian12:v1 docker image used for testing also contained tangoidl. Fix that by using the correct tangoidl_ROOT variable for overriding. |
Thomas Braun <thomas.braun@byte-physics.de> | no | https://gitlab.com/tango-controls/TangoSourceDistribution/-/merge_requests/157 | 2023-12-08 | |
| remove-hoverxref-extension.patch | Remove the sphinx-hoverxref extension since it has been deprecated upstream, and it is currently causing build issues: . /usr/lib/python3/dist-packages/hoverxref/extension.py:67: RemovedInSphinx90Warning: The '_Opt' object tuple interface is deprecated, use attribute access instead for 'default', 'rebuild', and 'valid_types'. context[attr] = app.config.values[attr][0] =================================================================== |
Santiago Ruano Rincón <santiago@freexian.com> | no | 2025-08-20 | ||
| use-a-local-copy-of-remote-intersphinx-inventories.patch | Use a local copy of intersphinx inventory objects This is required to avoid requiring access to Internet while building the package, and to avoid warnings avoid missing references that produce broken links. =================================================================== |
Santiago Ruano Rincón <santiago@freexian.com> | not-needed | 2025-09-03 | ||
| sphinx-output.patch | Don't write warnings and errors to a file stderr should be enough. And this creates a file that is not cleaned by make clean. =================================================================== |
Santiago Ruano Rincón <santiago@freexian.com> | not-needed | 2025-09-05 | ||
| sphinx-allow-warnings.patch | Don't fail on Sphinx warnings | Colin Watson <cjwatson@debian.org> | not-needed | 2026-03-08 | ||
| sort-glossary.patch | Sort input to glossary generator This makes the documentation build process reproducible. |
Colin Watson <cjwatson@debian.org> | yes | 2026-03-11 | ||
| detach-from-the-event-system-on-destruction.patch | CallBack: Detach from the event system on destruction The event system keeps a raw CallBack pointer in every subscription and checks it only against nullptr before calling push_event(). When the owner releases the callback while a subscription is still alive, the keep alive thread or the ZMQ consumer thread jumps through a dangling vptr. Language bindings hit this regularly, because they can free the callback before the DeviceProxy destructor gets as far as unsubscribing. Clear those pointers from the destructor instead. Every dispatch path holds map_modification_lock, respectively async_event_lock, across the call to push_event(), so taking both as a writer also waits for an invocation which is already running. The subscriptions are kept so that their event ids stay valid for a later unsubscribe_event(); an event for a detached subscription is dropped. Waiting for an event thread is only correct when that thread can still make progress. A binding whose callback needs an interpreter lock must therefore release it around the destruction of a subscribed callback, just as it has to around unsubscribe_event(). To keep that requirement confined to the case which needs it, track the callbacks a subscription refers to in a set with a mutex of its own, never held while a callback runs, and take no event system lock at all for a callback which was never subscribed. Destroying such a callback is what a binding does on every asynchronous reply. Destroying a callback from within one of its own callback methods cannot be made safe and would dead lock on the lock its dispatcher already holds, so that case is detected and reported instead. Closes #1813 |
Yury Matveev <yury.matveev@desy.de> | no | backport, https://gitlab.com/tango-controls/cppTango/-/merge_requests/1787 | 2026-10-23 | |
| 0001-debian-fix_database_scripts.patch | debian/fix_database_scripts update the database creation script to deal with the debian path |
Picca Frédéric-Emmanuel | no | 2011-03-22 | ||
| 0002-debian-my.cnf.in.patch | debian/my.cnf.in set the right parameters for the mysql configuration file |
Picca Frédéric-Emmanuel <picca@debian.org> | no | 2012-09-15 | ||
| 0003-remove-upstream-built-doc-from-cmake.patch | doc/html is being excluded in d/copyright. So we need to remove this subdir from doc/CMakeLists.txt |
Santiago Ruano Rincón <santiago@freexian.com> | not-needed | 2024-09-25 |
All known versions for source package 'tango'
- 10.3.3+dfsg1-4 (sid, forky)
- 10.0.2+dfsg1-2+deb13u1 (trixie)
- 10.0.2+dfsg1-2+deb13u1~bpo12+1 (bookworm-backports)
- 9.3.4+dfsg1-2+deb12u1 (bookworm)
