[p4] Amusing 2006.1 trigger failure
william_ivey at bmc.com
Mon Oct 2 12:00:14 PDT 2006
Just upgraded our server from 2005.1 to 2006.1 yesterday and
it went well, except this morning people couldn't check in.
(Yeah, that's a problem.)
Turned out that a "trick" accepted by triggers in 2005.1 no
len_chk change-submit //dev/... "$HOME/bin/namelen.trgr %change% '%user%@%client%'"
In 2005.1 $HOME was resolved, but that's no longer the case
in 2006.1. Not a big deal. The problem seemed obvious and
replacing $HOME with the actual path fixed it right up. Just
thought I'd toss this out for anyone else who might have done
something like that. (I realize I may be the only user in the
world who thought it was a good idea in the first place, though.)
More information about the perforce-user