RL wrote: > Franco <martelli...@gmail.com> writes: > >> " If you are comfortable with English language it's strongly >> recommended that you run the following command as soon as you start >> the 'script' session: >> >> # LC_ALL=C.UTF-8; LANGUAGE=; export LC_ALL LANGUAGE >> >> This will allow you to get command output messages in English into >> the script session. By doing so, it will help you for searching the >> web, during discussions or to submit a bug report." > > Are we *sure* this is a good idea - i have some doubts? > > -- "strongly recommended" by who? where did anyone previously make this > recommendation and why do they feel so strongly that it is needed in > 2024 when it has not been suggested before?
"Strongly" does seem a bit overconfident, but I've frequently seen it recommended for bug reporting. The fact that nobody's mentioned it in the specific context of the Release Notes may just mean that nobody's considered it - compare the way we've just been taking it for granted the user is running a sensible shell. > -- does anyone test the upgrade with this locale setting? What, C.UTF-8? Far more than test it under zh_TW.BIG-5 or whatever other locale the user might be using. If it fails on POSIX+unicode we've got real problems. > telling people to use something less tested seems like a bad > idea.eg, i have LANG=en_GB.UTF-8 would i still want to change > locales? This is one of the reasons I suggested starting with the problem, and saying "If you use a non-English locale..." (thus implying that it shouldn't matter for you). > -- why are we suggesting non-english message are somehow less clear? if > so, we should remove translations not hide them? Error messages in Korean or Norwegian or whatever are less likely to be intelligible to the average Debian developer reading a bug report, or to the average user reading the appeal for help on the debian-user mailing list. > -- why are we making the output harder to read for the user - that will > make it harder for them to fix their own issue. isnt this is the > opposite of what we want? Hence the "If you are comfortable using English..." caveat - I was hoping to avoid turning it into an explicit list of pros and cons for them to weigh up, but it's possible it does need some more. > -- isnt it better to say that if you get an error in non-english to > search the web for the english vesion? it's not like you cant look-up > the english version of the error message No, "better" would be if it was *less* bad! The idea is, if they're comfortable using English, and they can pick the relevant error out from its side-effects and the routine status messages (this is often the tricky bit), we want them to be able to quote what it says, not just give their best guess. > -- are we suggesting errors are likely? that isnt my experience with debian > upgrades Me too, but if nothing ever goes wrong, it hardly matters what the Release Notes say... > (maybe it makes more sense to do this if the upgrade fails and you can't > debug) Having a readable screen session makes things *much* easier to debug on the rare occasions when something does go wrong. -- JBR with qualifications in linguistics, experience as a Debian sysadmin, and probably no clue about this particular package