OK, what stops you from creating your own implementation table and fill that with whatever you want, and give that as an argument to ERR_set_implementation(). I know, it means you have to look in crypto/err/err.c for each version to see if there's been a change to ERR_FNS. Guess what? It sounds like you must fiddle with that file eaither way...
-- Richard Levitte \ Tunnlandsvägen 3 \ [EMAIL PROTECTED] [EMAIL PROTECTED] \ S-168 36 BROMMA \ T: +46-8-26 52 47 \ SWEDEN \ or +46-708-26 53 44 Procurator Odiosus Ex Infernis -- [EMAIL PROTECTED] Member of the OpenSSL development team: http://www.openssl.org/ Unsolicited commercial email is subject to an archival fee of $400. See <http://www.stacken.kth.se/~levitte/mail/> for more info. In message <[EMAIL PROTECTED]> on Fri, 4 Jul 2003 20:02:15 +0200 (METDST), "Frédéric Giudicelli via RT" <[EMAIL PROTECTED]> said: rt> Because, I could "stub" the default implementation, and if the error rt> handling has been disabled, then I just don't call the default rt> implementation function. rt> rt> Frédéric Giudicelli rt> http://www.newpki.org rt> rt> rt> ----- Original Message ----- rt> From: "Richard Levitte - VMS Whacker" <[EMAIL PROTECTED]> rt> To: <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> rt> Cc: <[EMAIL PROTECTED]> rt> Sent: Friday, July 04, 2003 1:52 PM rt> Subject: Re: [openssl.org #629] Custom error handling rt> rt> rt> > In message <[EMAIL PROTECTED]> on Fri, 4 Jul 2003 rt> 00:12:24 +0200, Frédéric Giudicelli <[EMAIL PROTECTED]> said: rt> > rt> > groups> The problem is the following, yes your code rt> (ERR_pop_to_mark/ERR_set_mark) rt> > groups> is fine when a child function is adding a new error, however, what rt> happends rt> > groups> when it calls ERR_clear_error ? In my implementation I need the rt> error stack rt> > groups> to be preserved even if a child function calls ERR_clear_error. rt> > groups> rt> > groups> That's why if you happended to remove the "if (!err_fns)" test in rt> > groups> ERR_set_implementation, I would be more than happy. rt> > rt> > I'm sorry, but in what way does that prevent the error stack to be rt> > cleared? Maybe a better thing would be to have a flag that inhibits rt> > clearing the error stack... I'll ponder over this issue. rt> > rt> > -- rt> > Richard Levitte \ Tunnlandsvägen 3 \ [EMAIL PROTECTED] rt> > [EMAIL PROTECTED] \ S-168 36 BROMMA \ T: +46-8-26 52 47 rt> > \ SWEDEN \ or +46-708-26 53 44 rt> > Procurator Odiosus Ex Infernis -- [EMAIL PROTECTED] rt> > Member of the OpenSSL development team: http://www.openssl.org/ rt> > rt> > Unsolicited commercial email is subject to an archival fee of $400. rt> > See <http://www.stacken.kth.se/~levitte/mail/> for more info. rt> > ______________________________________________________________________ rt> > OpenSSL Project http://www.openssl.org rt> > Development Mailing List [EMAIL PROTECTED] rt> > Automated List Manager [EMAIL PROTECTED] rt> > rt> rt> ______________________________________________________________________ rt> OpenSSL Project http://www.openssl.org rt> Development Mailing List [EMAIL PROTECTED] rt> Automated List Manager [EMAIL PROTECTED] ______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]