Red Hat Bugzilla – Bug 128375
Last modified: 2015-01-07 19:08:18 EST
Description of problem:
Details are sketchy, and not even really sure that vte is at fault,
but with vte-0.11.10-5.1 and gnome-terminal-2.6.0-4, I'm seeing
corruption in a gnome-terminal when running mutt. I've not seen the
corruption outside of mutt, so mutt might be at fault. Normally
everything is alright, but from time to time, it appears that a
portion of the screen won't update, leaving dangling text.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
Created attachment 102135 [details]
Created attachment 102136 [details]
Created attachment 102137 [details]
Screen after <ctrl>-L to refresh
Created attachment 102138 [details]
either TERM needs to be set to "gnome" or gnome-terminal needs to fix
its xterm emulation.
*** Bug 134442 has been marked as a duplicate of this bug. ***
Indeed setting TERM=gnome does resolve the problem. Would be really,
really nice if this were a commandline option (hint, hint). Leaving
the bug open, as the xterm emulation doesn't has some warts.
If you use a version of vte >= 0.11.11-7 (from rawhide) does it fix
Nope, updating to 0.11.11-10 with both gnome-terminal-2.7.3-1 and
gnome-terminal-2.8.0-2 still gives me the same behavior.
*** Bug 138548 has been marked as a duplicate of this bug. ***
For the record, I don't believe having a separate TERM=gnome setting
is the best alternative. TERM=xterm is well supported on many legacy
systems that one may still need to work with, whereas these systems
won't know anything about terminal type gnome.
I can't see this as anything but bugs in gnome-terminal that really
ought to be fixed.
We all agree with you. It just hasn't happened yet. Bill was giving
you a workaround in the interim.
*** Bug 141180 has been marked as a duplicate of this bug. ***
FWIW, I reverted to ncurses-5.4-5, which fixed the problem for me.
Is this still an issue in RHEL4?
I saw it in FC-3 with all the updates.
Even the vte-0.11.12 update?
Downgrading ncurses to 5.4-5 (from ncurses 5.4-13 / FC3) fixed the problem for
me as well. Nothing I did with vte worked. This is probably not a vte problem.
With no definite details, there is probably nothing we can do here. Please open
specific issues that you are able to reproduce at the upstream GNOME Bugzilla.
I hit the same problem and spent some too much time trying to find out
what gnome-terminal can be doing for me. Eventually it downed on me that
a terminfo entry may be responsible for that mess (unless the problem really
is that gnome-terminal is not doing correctly xterm stuff as doing the same
in xterm window does not have issues).
It is not necessary to drop the whole ncurses package to 5.4-5 level.
It is enough just to replace /usr/share/terminfo/x/xterm with the old one
and the problem disappears. 'infocmp' shows a long list of differences between
these two descriptions. Function key assignments are not likely the issue
here but even without those there is still enough left.
BTW - using here an xterm description from ncurses-5.4-17 (FC4) gives the
same troubles. That is not surprising as infocmp for those two shows only
this (names are my own so I can identify files):
comparing xterm-5.4-13 to xterm-5.4-17.
kbs: '^H', '\177'.
I should add that replacing /usr/share/terminfo/x/xterm provides a bit
of a relief but it is not a solution. After logging on a remove machine
/usr/share/terminfo/x/xterm from there seems to be used and not a local one.
I am afraid that leaving that CLOSED UPSTREAM is not of much help.
"Upstream" will tell you "just use a newer version" and I just checked
that at least gnome-terminal-2.13.93-1 from the current "rawhide" does
not have that problem. OTOH systems using gnome-terminal-2.7.3 will
be around for quite a while and just recompiling something newer is not
a realistic option. Even gnome-terminal-2.10.x needs various updated
libraries or it will not link.
BTW - my suspicions would be first on 'setf' and 'setb' string handling.
Both 'gnome' and an old 'xterm' have these set to NULL.
I have to agree with Michal. I expect a major bug (my manifestation is screen
corruption using ispell) in a critical component (the default terminal) of a
supposedly supported platform to either have an open bug against it or have a
Since the bug is closed, what is the workable solution?