The show-all-if-ambiguous setting is now enabled in /etc/inputrc.
Previously the setting was set to off, accompanied by the comment 'Show
all instead of beeping first'. The documentation for this setting says:
This alters the default behavior of the completion functions. If set
to ‘on’, words which have more than one possible completion cause the
matches to be listed immediately instead of ringing the bell. The
default value is ‘off’.
Since the setting is off by default and setting it to off does the
opposite of what the comment says, it looks like the intent was to
enable the setting and that it's disabled by mistake. The commit
message in ed01c26 ('Kills the annoying bell') where this setting was
originally added also suggests that the setting was intended to be
enabled.
We made this decision a long time ago in https://github.com/github/msysgit/pull/24 but we only changed the default `.gitattributes` and not the one we use when we actually create a repository with GHfW (`.gitattributes.suggested`)
Reasons
- people don't like seeing an "origin" remote immediately after
initializing a repository
- people don't like fetching PRs from the command line
Fix for issue #128 where the prompt causes the console to beep if there are
non-ASCII characters in the path. This patch replaces all non-ASCII characters
with a question mark when expanding the prompt
Signed-off-by: Anselm Kruis <a.kruis@science-computing.de>
Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
Piping the result of the `ls` command (alias) to a file will contain escape characters because the `--color` setting for `ls` is defaulting to `--color=always` instead of `--color=auto` (auto means, it will only produce color control codes on a real tty)
We've been setting the fetch refspec in the repo config, so having the
fetch refspecs in the system config causes trouble. We might put this
back after we decide to unset the repo's config, but for now we'll fetch
PRs explicitly when cloning to a branch.
In effect, this reverts the change of the escape sequence from
\033]0;
to
\033[0m
due to an incorrect understanding of their purposes in 4e4e3f3.
Conflicts:
etc/profile
In 4e4e3f3, the 0; and bell etc. were incorrectly interpreted as having
to do with colors. In fact, it is responsible for setting the window
title. [1]
[1] http://www.faqs.org/docs/Linux-mini/Xterm-Title.html#ss3.1
Signed-off-by: Tay Ray Chuan <rctay89@gmail.com>
This reverts commit 2db3aa7c41.
An end user cannot unset configuration options which means that it is not
possible to disable this feature without administrative permissions. This
should only be set in the global or repository configuration files unless
the user chooses to enable this option upon installation.
Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
The last change also printed the current working directory twice - once
after the MSYSTEM and once after user@host. Reduced to a single instance.
Acked-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
The previous "case" check already is case-sensitive, so there is no point
in making grep case-insensitive.
Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com>
Using awk to strip any Cygwin path creates a null-byte at the end of the
string. This makes the last item in the path unresolvable and causes child
processes that are not dependent on msys-1.0.dll to inherit POSIX rather
than Windows paths. Fix this by not using awk at all.
Reported-by: John Stevenson <john-stevenson@blueyonder.co.uk>
Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com>
Using awk to strip the Cygwin paths creates a null-byte at the end of the string. This makes the last item in the path unresolvable and causes child processes that are not dependent on msys-1.0.dll to inherit POSIX rather than Windows paths. This routine avoids awk and includes code from nicerobot on StackOverflow: http://stackoverflow.com/a/370255/1127485
Signed-off-by: johnstevenson <john-stevenson@blueyonder.co.uk>
In git commit af31a456 the git prompt code was moved out into another
file to permit lazy loading of bash functions. To maintain our git
prompt we must source both files now.
Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
`\[` and `\]` are used for enclosing nonprinting characters in bash PS1.
Enclosing printable characters inside those bracket pairs results in bad
behaviour when the PWD contains characters from an extended character
set (not strictly in ASCII).
Therefore, instead of enclosing big chunks of PS1 inside `\[`...`\]`
pairs, we only put them around the small, non-printable character parts.
Signed-off-by: Benoit Mortgat <mortgat@gmail.com>
This reverts commit f08322cab3
This check no longer works as since 60527eb binaries may be modified on
installation to speed up loading which changes the MD5 checksum on the
users system.
Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
These were added in ef3e533f2e, but they
are hurting, not helping. Consider the following patterns:
foo/
**/foo/
foo/**
The first will match all directories named "foo" (and anything inside
them) no matter where they reside in the repository.
The second will match directories named "foo" (and anything inside
them), but only if they are below the .gitignore in the directory
structure.
The third will match anything inside all directories named "foo".
Because Git only tracks files, not directories, the first and third
patterns are essentially equivalent. Since the first is simpler, we
should prefer it.
The second pattern can be useful in some cases, but as far as I can tell
it is not actually desired in any of the cases where we are currently
using it.
So, I've replaced all instances of patterns 2 and 3 above with pattern
1.
Brendan Forster
yuvale
Phil Haack
Mads Kristensen
Henrique Pereira
Drew Miller
Paul Betts
Paul Betts
Bernd
Anselm Kruis
Karsten Blees
Pat Thoyts
Tay Ray Chuan
Johannes Schindelin
Sebastian Schuberth
John Stevenson
Benoit Mortgat
Adam Roben
Tim Clem