Submitter | Sean Farley |
---|---|
Date | Nov. 22, 2013, 8:26 p.m. |
Message ID | <4870c6bcf56d3b86fd74.1385151962@laptop.local> |
Download | mbox | patch |
Permalink | /patch/3093/ |
State | Superseded |
Headers | show |
Comments
kbullock+mercurial@ringworld.org writes: > On 22 Nov 2013, at 2:26 PM, Sean Farley <sean.michael.farley@gmail.com> wrote: > >> # HG changeset patch >> # User Sean Farley <sean.michael.farley@gmail.com> >> # Date 1385072985 18000 >> # Thu Nov 21 17:29:45 2013 -0500 >> # Node ID 4870c6bcf56d3b86fd7456b7075cfd0ba86212b9 >> # Parent 4d30dc9b20a68e8b443da7ebbfa79ea2ea2e5825 >> bash_completion: turn off custom debuggers >> >> Previously, a user could have set ui.debugger=ipdb which forces a control >> (non-printable) character to be prepended to all hg output. This would confuse >> compgen enough to not match the first word in a list becuase it thinks >> \033[?1034h is the first part of the string of the first word. Therefore, we >> force ui.debugger to be unset when calling hg in a subprocess. >> >> diff --git a/contrib/bash_completion b/contrib/bash_completion >> --- a/contrib/bash_completion >> +++ b/contrib/bash_completion >> @@ -54,11 +54,11 @@ >> >> shopt -s extglob >> >> _hg_cmd() >> { >> - HGPLAIN=1 "$hg" "$@" 2>/dev/null >> + HGPLAIN=1 "$hg" --config ui.debugger= "$@" 2>/dev/null > > How about making HGPLAIN disable custom debuggers instead? Yeah, that makes more sense.
Patch
diff --git a/contrib/bash_completion b/contrib/bash_completion --- a/contrib/bash_completion +++ b/contrib/bash_completion @@ -54,11 +54,11 @@ shopt -s extglob _hg_cmd() { - HGPLAIN=1 "$hg" "$@" 2>/dev/null + HGPLAIN=1 "$hg" --config ui.debugger= "$@" 2>/dev/null } _hg_commands() { local commands