Hi Jamenson, At 2025-03-24T18:29:35-0300, Jamenson Espindula wrote: > Yes, I am aware of the 'po4a' (PO for anything) tool. Unfortunately, > 'po4a' does not gettext manpages very well. I tried to gettext the GNU > Bash manual page and 'po4a' did fail with no results at all. To be > clear, I transcript below a short text from the > 'Locale::Po4a::Man(3pm)' manual page: [snip]
Acknowledged. I think your preferences are sensible ones. It's also sensible for the po4a developers to _not_ want to interpret a full nroff interpreter. I pitched my vision of The Right Thing To Do to Martin Quinson, po4a author/maintainer, in a GitHub issue. https://github.com/mquinson/po4a/issues/527 ...and he seemed to agree. It means some work on the _groff_ side, but, I suspect, not much more than a weekend's worth,[1] once we have a flexible alternative output format command-line option. https://savannah.gnu.org/bugs/?66242 (Implementing that would also open the door to an "-A tree" option, addressing a question I asked this list's subscribers over the weekend.) https://lists.gnu.org/archive/html/groff/2025-03/msg00053.html >> The reason I haven't already committed this [dumping the node list of >> every output line] is because it requires an interface decision. Use >> an environment variable? If so, named what? Use a command-line >> option? If so, which letter do we want to permanently eat for it? As you can see, I already had an answer in mind, but wanted to solicit the untainted feedback of readers. Unfortunately, I didn't get any yet. > I intend to translate several man pages into Brazilian Portuguese and > I think many people could translate them into several languages if the > man page format would not be difficult to handle with. I agree. > I submit it to higher appreciation. I think your goal is worthy and your diagnosis of existing barriers to their achievement, sound. Regards, Branden [1] In the GitHub po4a issue, I characterized it as a "Google Summer of Code"-scale project. Having just familiarized myself with every node type in the formatter, and having an idea of the effort it would take to implement member functions to dump POD directly from GNU troff, I no longer asses the task's magnitude as that large. There are other details of GNU troff internals that I didn't characterize correctly in the GitHub issue. That was 6 months ago. I've improved my understanding since then.
signature.asc
Description: PGP signature
