Le 29/07/2010 04:50, Tommi Vainikainen a écrit : > In libpcre3 8.02 pcre_config option MATCH_LIMIT and > MATCH_LIMIT_RECURSION take a long integer pointer as where parameter, > but instead in older pcre those take a integer pointer. (see pcreapi.3 > function pcre_config and parameter MATCH_LIMIT, and implementation in > pcre_config.c.)
...so this is an ABI breakage. And looking at pcre3 changelog suggests that it was the previous version which was actually buggy. I guess anything using pcre_config in a similar way has the same problem. I've checked all reverse dependencies of libpcre3-dev. I looks like pcre is embedded in many other projects' sources: emboss, gtamsanalyzer.app, php5, tin, freemat, cvsnt, mongodb, nmap, hypermail and apache2. I guess they are not used, since they depend on libpcre3-dev (but I didn't check). So it looks like pcre-ocaml (and most likely its reverse-depends) was the only package affected by that. The proper way would be to do a transition... but in this case, it looks like too much. Maybe add Breaks in libpcre3 to prevent a partial upgrades from lenny from happening would do the trick. Or we could just forget about this bug. > And this is fixed in pcre-ocaml package version 6 in testing and > unstable, and only causes problem with stable's pcre-ocaml version 5 > packages. It was fixed in 6.0.1, by the way. Cheers, -- Stéphane -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
