| Summary: | app-shells/bash-3.2: history problems in vi mode | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Daniel Gryniewicz (RETIRED) <dang> |
| Component: | [OLD] Core system | Assignee: | Gentoo's Team for Core System packages <base-system> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | ||
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
|
Description
Daniel Gryniewicz (RETIRED)
2007-03-22 17:22:28 UTC
Jakub: What does this have to do with vim? It's a bash issue, and shows up if vim isn't even installed. Uh, yeah... this is totally not related to vi/vim. It's a set of key bindings for readline. So, this is a bash/readline problem. Re-assigning to base-system. Daniel, do you have colour prompt? Then this could be related bug 155369 ... I do, indeed, have a color prompt. My prompt is: PS1="^[[32;1m[\t \h] ^[[34;1m\W^[[0m> " However, I have a ~amd64 system, so I have those fixes versions already. I have sys-libs/readline-5.2_p2, and I still had the problem with app-shells/bash-3.2_p15 (although I don't appear to have put 3.2_p10-r1 into my package.mask, so maybe that one worked? I'll emerge it and try...) Nope, 3.2_p10-r1 still exhibits the same problem... 3.2_p17 still has the problem. Daniel, could you please describe the problem another time? Just explain steps you do and what goes wrong? Personally, I do not see anything unusual. Sure. Here's what I do: [15:19:55 athena] ~> ls evolution-2.11.1-more-64-bit.patch evolution-2.9.91-cal-reply.patch (That's all one one line). It doesn't matter what the command is; any command that results in more than 80 char on the screen, ie that causes the line to scroll with "<" on the left side, is sufficient. Then, after the command is run, I hit <esc>k to go up one line into history. The cursor is somewhere in the middle of the line, but it think's it's at the beginning. The line does not redraw correctly, even with <ctrl>-l. As I move the cursor right and left, it updates the character under the cursor, but it reverts to the wrong one after any kind of redraw. This makes command line history completely useless to use. If this isn't clear enough, I can try to make a screencast of the problem... My .profile sets these related settings: shopt -s histappend export HISTCONTROL="erasedups" PS1="^[[32;1m[\t \h] ^[[34;1m\W^[[0m> " set -o vi I have bash-completion enabled. My .inputrc contains this: set horizontal-scroll-mode On set editing-mode vi set completion-ignore-case On bash 2 is fine; bash 3.0.x is fine. bash 3.1 had the problem briefly (p8 and p10 failed) but 3.1_p17 is fine, and what I'm using now. bash 3.2.x all have the problem. I've tried p9 p9-r1 p10 p10-r1 p15 p15-r1 p17, all of which fail. I have readline-5.2_p4. Portage 2.1.2.5 (default-linux/amd64/2006.1/desktop, gcc-4.1.2, glibc-2.5-r2, 2.6.21-ck1-rc2 x86_64) ================================================================= System uname: 2.6.21-ck1-rc2 x86_64 Intel(R) Core(TM)2 CPU T7200 @ 2.00GHz Gentoo Base System release 1.12.10 Timestamp of tree: Unknown ccache version 2.4 [enabled] dev-lang/python: 2.4.4 dev-python/pycrypto: 2.0.1-r5 dev-util/ccache: 2.4-r7 sys-apps/sandbox: 1.2.18.1 sys-devel/autoconf: 2.13, 2.61 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10 sys-devel/binutils: 2.17 sys-devel/gcc-config: 1.3.16 sys-devel/libtool: 1.5.23b virtual/os-headers: 2.6.21 ACCEPT_KEYWORDS="amd64 ~amd64" AUTOCLEAN="yes" CBUILD="x86_64-pc-linux-gnu" CFLAGS="-O2 -march=nocona -pipe -ggdb -Wstrict-aliasing -Wreturn-type -Wimplicit -Wformat" CHOST="x86_64-pc-linux-gnu" CONFIG_PROTECT="/etc /usr/share/X11/xkb" CONFIG_PROTECT_MASK="/etc/env.d /etc/gconf /etc/rc.d /etc/revdep-rebuild /etc/terminfo /etc/texmf/web2c /etc/wget" CXXFLAGS="-O2 -march=nocona -pipe -ggdb -Wstrict-aliasing -Wreturn-type -Wimplicit -Wformat" DISTDIR="/home/portage/distfiles" FEATURES="ccache collision-protect cvs distlocks installsources metadata-transfer multilib-strict sandbox sfperms sign splitdebug strict" GENTOO_MIRRORS="http://distfiles.gentoo.org http://distro.ibiblio.org/pub/linux/distributions/gentoo" LANG="en_US.utf8" LDFLAGS="-Wl,--as-needed" LINGUAS="en en_US pq" MAKEOPTS="-j5" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --delete-after --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages --filter=H_**/files/digest-*" PORTAGE_TMPDIR="/home/portage" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/home/portage/overlays/gnome /home/portage/overlays/uml-overlay /home/portage/overlays/dang.overlays/user /home/portage/overlays/dang.overlays/maintainer /home/portage/overlays/sunrise /home/portage/overlays/gentopia /home/portage/overlays/local" SYNC="rsync://dang.nexthop.com/gentoo-portage" USE="X aac aalib accessibility acpi alsa amd64 apache2 artworkextra avahi avi bash-completion berkdb bitmap-fonts branding bzip2 cairo cdr cli cracklib crypt cscope cups db2 dbus dga divx4linux dri dvd dvdr eds emboss encode ethereal evo exif fam firefox flac font-server foomaticdb gd gdbm gif gimpprint gnome gnutls gphoto2 gstreamer gtk gtk2 guile hal howl iconv imagemagick imap imlib ipv6 isdnlog jabber jpeg keyring ldap libg++ libnotify libwww logrotate lzo lzw mad matroska midi mikmod mime mmap mng mono mozilla mp3 mpeg mpi ncurses nls nptl nptlonly offensive ogg opengl oscar oss pam pcre pda pdflib perl png ppds pppd python qemu-fast quicktime readline reflection rtc samba sdl session sharedmem slang slp smime soap softmmu speex spell spl sqlite sse3 ssl svg tcpd theora tiff tracker truetype truetype-fonts type1-fonts unicode usb vorbis wifi wmf xface xml xml2 xorg xsl xulrunner xv xvid zlib" ALSA_CARDS="hda-intel" ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa lfloat linear meter mulaw multi null plug rate route share shm softvol" ELIBC="glibc" INPUT_DEVICES="evdev keyboard mouse synaptics" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" LINGUAS="en en_US pq" USERLAND="GNU" VIDEO_CARDS="i810 vesa" Unset: CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LC_ALL, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS Daniel, can you reproduce your problem with horizontal-scroll-mode off? Also do you have \[ \] around color specifiers in your PS1? Try this (I've substituted ^[ with \033 (so you can just cut and paste) and putted \[ \] around colors) : PS1="\[\033[32;1m\][\t \h] \[\033[34;1m\]\W\[\033[0m\]> " horizontal-scroll-mode seems to make no difference; with it on or off, all of my investigation below was the same.
In addition, I tried both with and without vi history mode, and I had similar (although not identical, because emacs mode positions the cursor at the end of the line, while vi mode positions it at the beginning of the line) problems with both.
No, actual information.
That prompt is actually worse. I get the same problem (both with horizontal-scroll-mode on and off), but the cursor is over the first : in the time, not on the end of the line like with my normal prompt:
[12:04:54 athena] ~>
^
Cursor is here.
It does, however, seem to be prompt related, because the default prompt doesn't cause this problem, just my prompt.
Okay, a litte more investigation. My original prompt, with \t changed to \A still causes the problem. The stock gentoo prompt, with \u changed to \t causes the problem; however, with \u changed to \A, there is no problem. My original prompt with \t changed to \A *does* cause the problem.
Here's a prompt that works:
PS1='\[\033[01;32m\]\A \h\[\033[01;34m\] \w \[\033[00m\]> '
Here's one that fails:
PS1='\[\033[01;32m\]\t \h\[\033[01;34m\] \w \[\033[00m\]> '
Just for fun, I tried \D. Using seconds fails:
PS1='\[\033[01;32m\]\D{%H:%M%S} \h\[\033[01;34m\] \w \[\033[00m\]> '
But using no seconds succeeds:
PS1='\[\033[01;32m\]\D{%H:%M} \h\[\033[01;34m\] \w \[\033[00m\]> '
Interestingly, by accident, I typoed the prompt, and this works fine:
PS1='\[\033[01;32m\]\D{%H:%M%S} \h\[\033[01;34m\] \W \[\033[00m\]> '
(note the lack of : between %M and %S) Adding that : back in causes failure.
For now, I'm using the prompt that works above, but I'd like to have my seconds back on the display.
One final thing: I'm using gnome-terminal as my terminal, with vte-0.16.3 as it's renderer. I tried several combinations with xterm, as well, and xterm fails for any prompt I tried with time in it, even the ones that worked with gnome-terminal, so there is some terminal component involved.
Update: I was wrong; the prompts marked "working" don't actually work all the time. They work fine in my homedir, where the path is "~", but as soon as I change directories, they start failing again. So, it's back to 3.1_p17 for me. I suddenly need a solution to this; sys-apps/portage-2.1.3.15 needs >=app-shells/bash-3.2_p17, and both p17 and p17-r1 are broken. I'd highly prefer not to lose my timestamps in my prompts. I've found a version of the prompt that works. Basically, you have to set the color before and after the time, or the display is messed up. Here's my working prompt:
COLOR1=34
COLOR2=31
PCOLOR1="\[\033[1;${COLOR1}m\]"
PCOLOR2="\[\033[1;${COLOR2}m\]"
PCOLORN="\[\033[0m\]"
PS1="${PCOLOR1}[${PCOLOR1}\A${PCOLOR1}\
${PCOLOR1} ${PCOLOR1}\h${PCOLOR1}]\
${PCOLOR2}\W\
${PCOLORN}>${PCOLORN} "
still a problem with 4.2 ? No longer a problem with 4.2. This can be closed, as far as I'm concerned. |