Open mariuszskon opened 3 years ago
can you provide a rationale for this change? what effect are we trying to avoid here?
In short I'm trying to avoid return code 1 when doing ./get launchpad
. Launchpad is the only one suffering from this error (debian might but all downloads succeed).
@mariuszskon sure but why does sleeping help? I'd understand if you would introduce some kind of retry mechanism with a back-off, but 100ms seems like some magic value
I'm not 100% sure but I believe the sleep return code is used which leads to 0.
I agree the 100ms is a magic value, but I saw it in all the other get_all_*
functions.
To clarify, I don't think this is a great solution either, but I saw the sleep .1
in the other functions and it has the desired side effect of having a zero return code.
I want #49 solved, this is perhaps not the way to do it.
I'm not 100% sure but I believe the sleep return code is used which leads to 0.
I don't think that's accurate. So it's not an obvious fix for me. will have to look at the actual error case.
Do we even have a consistent exit code set for the downloads? I mean, there has been an error, so a non-null exit code would be fine - but we've ignored that error and continued downloading other packages.
Maybe just adding an exit 0
at the end of the get
script does the trick? Otherwise reducing the set of ubuntu versions tracked by the launchpad downloader would work too, to exclude some really old versions.
This is ensures we consistently perform a sleep regardless what categories we download.
Naively fixes https://github.com/niklasb/libc-database/issues/49.