> From: Jeremy Farrell > Sent: Wednesday, January 29, 2014 1:39 AM > > > From: Dr. Stephen Henson [mailto:st...@openssl.org] > > Sent: Wednesday, January 29, 2014 12:50 AM > > To: openssl-users@openssl.org > > > > On Tue, Jan 28, 2014, Jeremy Farrell wrote: > > > > > > > > Ugh. Thanks for checking Steve, that's rather different from the > > > understanding I'd built up. I suggest a quick fix to improve the > > > documentation would be simply to delete the sentence "If it has > been > > > allocated by OPENSSL_malloc(), *flags&ERR_TXT_MALLOCED is true". > > > At the moment, that appears to be giving a hint that the caller > must > > > free it, whereas it's actually an internal detail of no use to the > > > caller and rather dangerous for him to know. > > > > > > If I remember correctly, few (if any) bits of code malloc the > string > > > at the moment, so it's not currently a big issue in practice. > > > > Actually just about everything mallocs the extra data string. It is > > almost > > exclusively used used when additional data is added to an error code. > > > > The code in question is ERR_add_error_data which calls > > ERR_add_error_vdata > > which concatenates strings into a OPENSSL_malloc'ed buffer and then > > sets it > > using ERR_set_error_data which includes ERR_TXT_MALLOCED. > > > > Steve. > > Man, my memory's getting worse than a ... a ... what d'you call it.
And thanks for the quick doc fixes Steve, should prevent others making my mistake. ______________________________________________________________________ OpenSSL Project http://www.openssl.org User Support Mailing List openssl-users@openssl.org Automated List Manager majord...@openssl.org
