Re: libpq5 8.3 breaks 8.2 compatibility with encodings
От | Martin Pitt |
---|---|
Тема | Re: libpq5 8.3 breaks 8.2 compatibility with encodings |
Дата | |
Msg-id | 20071012151446.GI5911@piware.de обсуждение исходный текст |
Ответ на | libpq5 8.3 breaks 8.2 compatibility with encodings (Martin Pitt <martin@piware.de>) |
Ответы |
Re: libpq5 8.3 breaks 8.2 compatibility with encodings
|
Список | pgsql-bugs |
Hi, Martin Pitt [2007-10-12 16:33 +0200]: > I'm currently hunting down the last postgresql-common test case > failure that I see with 8.3beta1. It seems the 8.3 version of libpq > changes some internal encoding lists? Ah, got it. The ordering in pg_enc2name_tbl[] changed, which makes the indices jump around. This was introduced in [1], in particular in those two bits: http://developer.postgresql.org/cvsweb.cgi/pgsql/src/include/mb/pg_wchar.h.diff?r1=1.71;r2=1.72 http://developer.postgresql.org/cvsweb.cgi/pgsql/src/backend/utils/mb/encnames.c.diff?r1=1.32;r2=1.33 With attached patch (which restores the previous ordering) compatibility with 8.2 is restored. This has two drawbacks: * The enum cannot be nicely sorted by internal and client-only encodings until libpq bumps soname again. This is only a cosmetical problem, though. * This patch needs another catalog bump (to "unbump" the one in [1]). That's unfortunate, but the catalog number got bumped in between beta and release in earlier versions, too, so I hope it's not too bad. The pg_enc2name_tbl declaration should probably have a comment saying to never alter the order, but only append new stuff at the end. For encodings which became obsolete (should that happen) there should be an constant like "INVALID" or "DEPRECATED". Thank you! Martin [1] http://archives.postgresql.org/pgsql-committers/2007-04/msg00198.php -- Martin Pitt http://www.piware.de Ubuntu Developer http://www.ubuntu.com Debian Developer http://www.debian.org
Вложения
В списке pgsql-bugs по дате отправления: