Description of problem: Commands longer than screen width don't wrap properly Version-Release number of selected component (if applicable): bash-3.0-31 How reproducible: Always. Steps to Reproduce: At a bash prompt 1.type the letter 'x' until the end of the line is reached. 2.continue typing the letter 'x'. 3.set -o vi only changes the problem from one incorrect behaviour to another incorrect behaviour. Actual results: Without set -o vi: The last character overwrites the (previous) final character of the line on the screen. With set -o vi: The line wraps around on the screen onto the same line already filled with the first part of the command. Expected results: The line should continue on the line below the first one. Additional info: This problem does not occur with bash-3.0-18 on FC3. I am ssh'd into the FC4 box from a FC3 box which is running screen. I am ssh'd into the FC3 box from a W2K (5.00.2195 SP4) box which is running Putty 0.58.
Please paste in the output of 'set' so I can see your prompt settings.
subset of the output of 'set' follows: BASH=/bin/bash BASH_ARGC=() BASH_ARGV=() BASH_LINENO=() BASH_SOURCE=() BASH_VERSINFO=([0]="3" [1]="00" [2]="16" [3]="1" [4]="release" [5]="i386-redhat- linux-gnu") BASH_VERSION='3.00.16(1)-release' COLUMNS=133 HOSTTYPE=i386 IFS=$' \t\n' INPUTRC=/etc/inputrc LANG=en_US.UTF-8 LINES=50 MACHTYPE=i386-redhat-linux-gnu PROMPT_COMMAND='echo -ne "\033_${USER}@${HOSTNAME%%.*}:${PWD/#$HOME/~}\033\\"' PS1='[\u@\h \W]\$ ' PS2='> ' PS4='+ ' SHELL=/bin/bash SHELLOPTS=braceexpand:hashall:histexpand:history:interactive-comments:monitor:vi TERM=screen
Additional data point: I also see this behaviour going from SuSE 9.3 running in Konsole ssh'd to FC3 running the same screen, ssh'd to FC4. However . . . I _don't_ see the behaviour going directly from the SuSE 9.3 box to the FC4 box. Looks like it's some kind of bizarre interaction between screen on FC3 and bash on FC4. If I eliminate either of those, the behaviour goes away. screen on FC3 is screen-4.0-2.5 Now I'm not sure which to file this bugreport against.
OK, this gets even more bizarre. I re-did (as best as I knew) the things leading up to this problem in another session, and it didn't happen again. It was still happening in the first session though. When I backed out of ssh+screen, and went back into it, I can't reproduce the problem. At this point since I cannot reliably reproduce the problem it would probably be OK to close this bug.
I just want to put one final comment on this. This is most likely some kind of screen issue. The session that had this problem was a long-lived screen session that had been attached, detached, and reattached from over many days, from tty's with a variety of screen sizes. I would surmise that somehow screen lost track of the tty size for the window that was attached to the FC4 box.