Hello, Looking a little further, the "problem" is that endian.h is not included, because _LINUX_ is not defined. So, does anyone know the right way to handle this? Simply compiling with -D_LINUX_ fixes it; should all the build scripts automatically determine if you're on a linux system and define _LINUX_ if so? Or can we simply grab endian.h, part of the GNU C library, and distribute it with dbmail (Eelco/Roel)?
If there's a clean/right way to handle this, I'll try to put together a patch to handle md5 hash and digest passwords nicely, which after enough testing we'll force IC&S to commit to cvs. :) Thanks, Jesse ---- Original Message ---- From: Dave Logan <dbmail@dbmail.org> To: dbmail@dbmail.org Subject: [Dbmail] sparc endian-ness problems? Sent: Wed, 26 Feb 2003 11:24:34 -0700 (MST) > > OK, I formerly was fingerpointing at libcrypt(o) as the culprit > for the APOP not working. Now I see that it's actually the > fault of just the makemd5(), who calls the md5.h files, who > includes the <endian.h>. > Quick overview again, the hash isn't generated properly, so > the login is not authenticated: > Feb 26 09:33:43 frobnitz dbmail/pop3d[21719]: pop3(): incoming buffer: [apop > [EMAIL PROTECTED] 77b6938dfc1b04468d0d0a0368f75cbc] > Feb 26 09:33:43 frobnitz dbmail/pop3d[21719]: pop3(): command issued :cmd > [apop], value [EMAIL PROTECTED] 77b6938dfc1b04468d0d0a0368f75cbc] > Feb 26 09:33:43 frobnitz dbmail/pop3d[21719]: pop3(): APOP auth, username > [EMAIL PROTECTED], md5_hash [77b6938dfc1b04468d0d0a0368f75cbc] > Feb 26 09:33:43 frobnitz dbmail/pop3d[21719]: auth_md5_validate(): > apop_stamp=[<[EMAIL PROTECTED]>], userpw=[password] > Feb 26 09:33:43 frobnitz dbmail/pop3d[21719]: auth_md5_validate(): > checkstring for md5 [<[EMAIL PROTECTED]>password] -> result > [9c71fcb909c309298119afba7ce6b630] > Feb 26 09:33:43 frobnitz dbmail/pop3d[21719]: auth_md5_validate(): validating > md5_apop_we=[9c71fcb909c309298119afba7ce6b630] > md5_apop_he=[77b6938dfc1b04468d0d0a0368f75cbc] > Feb 26 09:33:43 frobnitz dbmail/pop3d[21719]: auth_md5_validate(): user > [EMAIL PROTECTED] could not be validated > > > If I comment out (in md5.c) the lines that define an emtpy > function for byteReverse(), so it's forced to use the full > function (big endian) it works properly. > So, how do I tell whoever (compiler, the function, ???) that > this is really a big endian sparc64? > > Woefully inadequate in c, > davel > > > -- > Dave Logan > > Entertainment and Education all in One Big Bowl: > http://www.digitalcoven.com/ > > > _______________________________________________ > Dbmail mailing list > Dbmail@dbmail.org > https://mailman.fastxs.nl/mailman/listinfo/dbmail > -- End Original Message -- -- Jesse Norell [EMAIL PROTECTED]
