Repository navigation
Allow motion's %$ conversion specifier - #3273
Conversation
|
@copilot |
|
Great #3274 is ready to be merged into whichever PR adds this change. That way we avoid any additional translation needs. A blueprint for whenever we change source strings in a sufficiently trivial way, or in a way which does not render the translations entirely wrong. If some native speaker shall better have a look, we can let Copilot add the |
Signed-off-by: MichaIng <micha@dietpi.com>
16291f9 to
98f7513
Compare
…pecifier (#3274) Signed-off-by: MichaIng <micha@dietpi.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
|
I'll merge this now. I asked study orga to forward the info to the participants, but people are busy, and so far no one reacted. The changes are rather small, and we can craft a branch with the state from before this commit, to allow the true author to reapply it formally for the commit history, if needed. But I want to get these cleared now for 0.44.0 release. |
…3377) The image/movie file-name validator rejected the %I (12-hour hour), %p (AM/PM) and %l (space-padded 12-hour hour) conversion specifiers, so valid Motion file-name patterns such as "%Y-%m-%d/%I-%M-%S-%p" were refused in the UI with "Make sure all the configuration options are valid!", even though Motion saves the files correctly when the pattern is set directly in the config. Add I, p and l to filenameValidRegExp in both config.py and static/js/main.js (kept in sync per the existing requirement) and list the new specifiers in the validation failure message. Fixes #2070 * Update translation catalogs for the new file-name specifiers Insert %I %p %l into the filenameFailMessage specifier list across all motioneye.po catalogs (both the msgid and the translated msgstr where one exists) and motioneye.pot, so translators do not need to re-translate the whole string when only literal specifiers are added. Mirrors #3273. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit was done with an anonymous account in context of a study. The actual author might want to open a PR with its true account for proper credits, after the study ended. However, I am opening this here to show that this issue has been addressed, so no one starts doubled work 😉.
My only concern was that it re-introduced this RCE vulnerability: GHSA-j945-qm58-4gjx
However, the
%$seems to be replaced bymotionbefore it is piped through the shell, so%$VARIALEand%$(command)are expanded toCamera1VARIABLEandCamera1(command)respectively, no variable or command substitution hence happens.