Hi Beatrice, Am 9. August 2025 13:03:41 MESZ schrieb Beatrice Torracca <[email protected]>: >I can use > make update-po LANGS=it >to update the .po file for Italian. That also changes all the .pot files in >source/locales/ > >I always wonder if I should commit those too, and in doing so, make all other >languages incomplete, but at the same time sort of activate an update for all >languages.
The other languages are not getting incomplete or outdated, because you are doing the 'make update-po' call. Indeed they were imcomplete/outdated before already - because the English original has changed (and this the only way, they are getting implete/outdated). >Or if I should just update and commit the Italian .po files, and wonder if >that will cause problem having them not corresponding to the .pot in the >repository. Yes, you can just commit your Italiano files, and leave the rest aside. No breakage will happen because of this. The only purpose behind the pot files is: they serve as empty templates for people starting a translation for a new language. >I don't want to cause any prob, if anyone working on the original content can >give me an advice on what to do without disrupting the workflow for English >and all the other translations, I would be very grateful. The ideal workflow would be, to call 'make update-po' for all languages (and also commit the pot files in that step as well), when English has changed (or once a day at least, or similar), and I did that in the past indeed several times. Holger PS: I'm a quite unhappy with the 'make update-po' functionality in Sphinx/reST. Comparing that with the old docbook-based variant it is lacking an important feature for translators: it does not add the old English text to the po files as comment (this is what docbook/po4a does via the '--previous' commandline option). That feature allows translators easily to see what has changed in the English original text in their po editor (I use lokalize for that, for example). I opened an issue against sphinx on github to ask for such feature, but the upsteam authors seem to think different about this: <https://github.com/sphinx-doc/sphinx-intl/issues/121> -- Sent from /e/ OS on Fairphone3
