Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email email@example.com
Subject: pam_get_user() query
From: David Wragg (dpwdoc.ic.ac.uk)
Date: Tue Feb 15 2000 - 14:23:18 CST
- Next message: Steve Langasek: "Re: Large password file"
- Previous message: werner maes: "Large password file"
- Next in thread: Steve Langasek: "Re: pam_get_user() query"
- Reply: Steve Langasek: "Re: pam_get_user() query"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
The Module Writers' Guide says that the storage for the username given
by pam_get_user() will be freed at pam_end(), unsurprisingly, or at
the next call to pam_get_user(), which is surprising to me. Assuming
that the username cannot change between pam_start() and pam_end(), why
wouldn't pam_get_user() just return a pointer to the same storage on
every call? (Which is how it works in the current Linux-PAM source).
Am I missing something, or does this provision come from some
preexisting PAM spec? If not, I'd like to suggest a change to the
relevant paragraph, so that the module I'm working on can rely on the
username asociated with a pam_handle_t being fixed.
-- To unsubscribe: mail -s unsubscribe pam-list-requestredhat.com < /dev/null