Code review comment for lp:~spiv/bzr/no-sigwinch-583941
Revision history for this message
| Andrew Bennetts (spiv) wrote | # |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
73416ba
(Get the code!)
vila and I have had an extensive discussion on IRC. I'll try to write a comprehensive summary soon, but right now it's dinner time, so here's the short version:
1) I'll rework this patch for 2.1 to preserve the 2.1 behaviour as much as possible without using SIGWINCH (i.e. use COLUMNS until we observe the size from TIOCGWINSZ change), to minimise risk of regressions in 2.1.
2) I'll submit another patch for trunk that I think will be a better implementation of terminal_width: try $BZR_COLUMNS, then TIOCGWINSZ, then $COLUMNS in turn, until we find one that is set to a non-zero value. If none are set non-zero, return None. This patch will try to explicitly document the rationale and problematic cases to take into account, because the literature around determining terminal sizes is pretty confusing.
I'm popping this branch back to work-in-progress for now, but expect a resubmission soon.