retitle 661937 [pgstatus] P-A-S matching could be smarter thanks On 03/03/2012 01:03 PM, Adam D. Barratt wrote: > > On 02.03.2012 18:48, Samuel Bronson wrote: >> At the moment, the bottom of >> <http://buildd.debian-ports.org/status/package.php?p=tack> shows >> this enlightening tidbit: >> >> ,---- | Occurrences found in Packages-arch-specific file: | | >> %racket: !ia64 # [ANAIS] incompatible with ia64's stack layout > [...] >> Presumably, this should only be matching whole-word occurances, and >> should never match anything in a comment? > > fwiw, the following appears to at least be less crazy than the > current match: > > - $message = shell_exec(sprintf("egrep \"%s\" %s", + $message > = shell_exec(sprintf("egrep \"^%%?(%s):\" %s", > > This may not match binary package P-a-s entries correctly, but then > the current version doesn't either unless the binary contains the > source package name as a substring... >
I've applied Adam's suggestion, for now. In order to properly fix this issue, pgstatus needs to have the src<->binary associations so that it knows what it could look for in P-A-S files. But I'm not sure it is worth the trouble :/ as it means parsing big files. (that's why I didn't dare to fix the grep line earlier). Besides, nowadays, we tend to get that information out of P-A-S files and let maintainers deal with the architecture field of their packages. I'm not closing this bug as one may still implement a correct matching some day. Regards, -- Mehdi Dogguy مهدي الدڤي http://dogguy.org/ -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
