В Mon, 06 Sep 2010 11:32:54 +0100, Roger Leigh написа: > I'd like to suggest that the best way to tackle the problem is to > simply stop generating the source file/line number comments by > default; I'm already doing this in some of my projects by adding > "--no-location" to XGETTEXT_OPTIONS in po/Makevars. It's a massive > improvement.
Improvement for you as a maintaner, probably. It's a massive disaster for translators. > A question I have is "what purpose does having the line number and > source file serve"? Automatically jumping to the source. That's invaluable. The only case where locations do not work (and do not make sense) is for XML gibberish like .glade/.ui files. The filename is still useful, though -- I'm using it to open the file with Glade, and click around to find the message I'm interested in. > And if the original source file(s) for a string need to be found, > grep(1) is pretty fast. Try this with a non-trivial (or even trivial) package, you'd be annoyed pretty quickly. > At least for me, the translators get mailed the po file, and never > look at the source, so it's not of *any practical benefit* to anyone Yes, unfortunately these days many translators take the "black box" approach -- not only they never look at the source (which leads to amusing translations with horrible quality), they don't even compile and run the program. That's not a valid reason to cripple the .po files, though, making the life significantly harder for those who take care to test their translations properly. Source code availability *is* a prerequisite for a decent translation, and the locations of the messages are a major convenience (with a capable PO editor, of course).
