Mikhail Teterin <[EMAIL PROTECTED]> writes: > The pam_setenv and pam_putenv are backwards, IMHO. putenv should be > using setenv -- not the other way around. Currently, the setenv takes > NAME and VALUE separately, mallocs a new buffer, sprintfs %s=%s into it, > sends the buffer to putenv, which re-parses it and frees it. > > I think, pam_setenv should be doing the actual "dirty work", with putenv > being a wrapper. This would save some cycles (and, possibly, syscalls > -- from malloc), but, of course, it would not be very significant with > todays hardware, yada, yada...
No, the storage format for environment variables is part of the API and is intended for compatibility with libc. Doing it your way would be backward. > Would you have any other comments about my original post -- why is > pam_setenv causing the segfault somewhere, and is there anything wrong > with my patch? Thanks! I don't know why it crashes, and I haven't looked at the patch. I probably won't have time for it until early next year. In the meantime, merry Christmas and a happy New Year :) DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
