David Champion wrote: > Sure, I can try. I failed to note that smtp_fill_account is static, so > you can't see it from global scope. Perhaps try url_parse_ciss instead. > > $ gdb mutt > (gdb) set args -e 'set > smtp_url="smtps://[email protected]:[email protected]/"' [email protected] > (gdb) break url_parse_ciss > (gdb) run > > Interact with mutt; send message and wait for break. You should see > breakpoints, e.g.: > Breakpoint 1, url_parse_ciss (ciss=0xbfffc684, src=0x8276a48 > "smtps://[email protected]:[email protected]/") at url.c:172 > > You may need to enter "cont" a few times to skip past IMAP parses, which > don't interest us at present. Only move on when "src" has a value > beginning with "smtps". > > (gdb) finish > (gdb) print url > $2 = {scheme = U_SMTPS, user = 0x8276a50 "[email protected]", pass = 0x8276a61 > "mypassword", host = 0x8276a69 "smtp.gmail.com", port = 0, > path = 0x8276a78 ""} > > This is what needs to be confirmed. If that appears correctly then the > problem is elsewhere -- SASL, mutt's interaction with SASL, etc. -- and > not with URL parsing.
OK, the problem is not with URL parsing :) > You might also run with debugging on (mutt -d3) and send the output in > ~/.muttdebug0. I don't think a password should appear there, but check > it to be sure. While performing the test with -d3 it has now suddenly decided to start working. So it may have been a wobble with gmail's SASL authentication process server side. Apologies for pinning the tail on the wrong donkey. -- Chris Burdess
