From: Tatjana Azundris Nuernberg Date: May 11 2011 11:50pm Subject: Re: bzr commit into mysql-5.5-bugteam branch (azundris:3229) Bug#21287 Bug#11745920 List-Archive: http://lists.mysql.com/commits/137143 Message-Id: <4DCB20C1.6080507@oracle.com> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: quoted-printable On 2011/05/11 20:48, Davi Arnaut wrote: Hi Davi, Thanks for your input! After reading the docs some more, and running some more tests ... There is absolutely no reason we shouldn't hand in the proper return=20 values from the SSL I/O calls, seeing as we do have the information.=20 This should prevent GIGO. Code fixed to pass in the SSL_*() return=20 values. New patch also removes duplicate report_errors() in=20 viosslfactories.c, and renames the function to ssl_report_errors() to=20 keep the namespace sane. > Why not simply use ERR_peek_error? De facto it's always 0 ("not in error state") for the I/O functions, so=20 we wouldn't get useful error messages for failing I/O. Or well, we do,=20 but we shouldn't. :) I've split the I/O and other error handling now. In = fact I'm not sure the "other errors" ever come up, but at least that way = we're guaranteed an empty error Q *before* calling I/O: this doesn't=20 actually seem necessary at this point, but the docs say we should do=20 this. Call it future-proofing. :) I'm much happier with this patch. Best regards, Tatiana --=20 Tatiana Azundris, Bugs Bunny * IRC/Skype: Azundris * azundris@stripped ORACLE Deutschland B.V. & Co. KG - http://oracle.com Hauptverwaltung: Riesstr. 25, D-80992 M=FCnchen Registergericht: Amtsgericht M=FCnchen, HRA 95603 Komplement=E4rin: ORACLE Deutschland Verwaltung B.V. Rijnzathe 6, 3454PV De Meern, Niederlande Handelsregister der Handelskammer Midden-Niederlande, Nr. 30143697 Geschaftsf=FChrer: J=FCrgen Kunz, Marcel van de Molen, Alexander van der = Ven Oracle is committed to developing practices and products that help protect the environment - http://www.oracle.com/commitment