Hi Michael, thanks for your feedback. Le dimanche, 25 novembre 2012 19.15:39, Michael Gilbert a écrit : > I personally prefer the one-line fix for wheezy due to simplicity. I > think the "nice" but large fix should wait till jessie. I very much disagree here. Let me explain why: first, if look closely, it's not a "big" fix: we drop the postinst completely in favour of a two-lines file that is installed in /usr/share/cups/ppd-updaters/. debian/control.in | 1 + debian/control.in.in | 1 + debian/ghostscript-cups.install | 3 ++ debian/ghostscript-cups.postinst | 61 ---------------------------------- debian/ghostscript-cups.ppd-updater | 2 + This mechanism is proven to work a) in general by 15 other packages that ship ppd-updater trigger files, b) by Ubuntu shipping that solution since "25 Jan 2012" (9.05~dfsg~20120125-0ubuntu1, released in Precise). Furthermore, as detailed in the #520753 bug (especially in #33), this is the technically correct solution: cups handles the updating of its printer queues specifications itself, using the smallest information needed, provided by the driver packages through the ppd-updater files in the trigger. Having this mechanism duplicated in 16 maintainer scripts was a bad solution and keeping that in ghostscript makes it an exception while all other driver packages migrated to using the cups trigger mechanism. Getting rid of that duplication for Wheezy is worth it IMHO. > Also, a > 1-day delay for an already fixed bug doesn't seem right. I have rescheduled it to 5 days (+4). Please test the new package in the mean time. > I would rather see a new bug opened describing exactly what is wrong > with the existing fix. Sorry, but the #520753 had all the information needed back when you uploaded the "one-line fix", that you uploaded without commenting on the bug (and apparently without reading the bug log). Cheers, OdyX
Attachment:
signature.asc
Description: This is a digitally signed message part.