# Problem with git-gui and relative directories 4 messages from 2007-08-24 to 2007-08-25. Participants: David Kastrup, Shawn O. Pearce. Thread: https://gitlist.dev/t/9633 ## David Kastrup, 2007-08-24 13:40 Subject: Problem with git-gui and relative directories Message-ID: <868x81vynk.fsf@lola.quinscape.zz> URL: https://gitlist.dev/e/868x81vynk.fsf%40lola.quinscape.zz ``` Doing something like mkdir -p junk/woozle cd junk git-init cd woozle echo "hooray" > plop git add plop git commit -m "x" git gui blame plop errors out with fatal: cannot stat path woozle/plop: No such file or directory Obviously, git-gui is confused about relative paths here. -- David Kastrup ``` ## Shawn O. Pearce, 2007-08-25 03:18 Subject: Re: Problem with git-gui and relative directories Message-ID: <20070825031834.GW27913@spearce.org> URL: https://gitlist.dev/e/20070825031834.GW27913%40spearce.org In-Reply-To: <868x81vynk.fsf@lola.quinscape.zz> ``` David Kastrup wrote: > Doing something like ... > errors out with > > fatal: cannot stat path woozle/plop: No such file or directory > > Obviously, git-gui is confused about relative paths here. Yes, it is. I'm committing this, to be included in 0.8.2: -->8-- git-gui: Correct 'git gui blame' in a subdirectory David Kastrup pointed out that the following sequence was not working as we had intended: $ cd lib $ git gui blame console.tcl fatal: cannot stat path lib/console.tcl: No such file or directory The problem here was we disabled the chdir to the root of the working tree when we are running with a "bare allowed" feature such as blame or browser, but we still kept the prefix we found via `git rev-parse --show-prefix`. This caused us to try and look for the file "console.tcl" within the subdirectory but also include the subdirectory's own path from the root of the working tree. This is unlikely to succeed, unless the user just happened to have a "lib/lib/console.tcl" file in the repository, in which case we would produce the wrong result. In the case of a bare repository we shouldn't get back a value from `rev-parse --show-prefix`, so really $_prefix should only be set to the non-empty string if we are in a working tree and we are in a subdirectory of that working tree. If this is true we really want to always be at the top level of the working tree, as all paths are accessed as though they were relative to the top of the working tree. Converting $_prefix to a ../ sequence is a fairly simple approach to moving up the requisite levels. Signed-off-by: Shawn O. Pearce --- git-gui.sh | 10 +++++++++- 1 files changed, 9 insertions(+), 1 deletions(-) diff --git a/git-gui.sh b/git-gui.sh index 743b7d4..fa30ccc 100755 --- a/git-gui.sh +++ b/git-gui.sh @@ -703,7 +703,15 @@ if {![file isdirectory $_gitdir]} { error_popup "Git directory not found:\n\n$_gitdir" exit 1 } -if {![is_enabled bare]} { +if {$_prefix ne {}} { + regsub -all {[^/]+/} $_prefix ../ cdup + if {[catch {cd $cdup} err]} { + catch {wm withdraw .} + error_popup "Cannot move to top of working directory:\n\n$err" + exit 1 + } + unset cdup +} elseif {![is_enabled bare]} { if {[lindex [file split $_gitdir] end] ne {.git}} { catch {wm withdraw .} error_popup "Cannot use funny .git directory:\n\n$_gitdir" -- 1.5.3.rc6.17.g1911 -- Shawn. ``` ## David Kastrup, 2007-08-25 07:26 Subject: Re: Problem with git-gui and relative directories Message-ID: <85veb4dqhv.fsf@lola.goethe.zz> URL: https://gitlist.dev/e/85veb4dqhv.fsf%40lola.goethe.zz In-Reply-To: <20070825031834.GW27913@spearce.org> ``` "Shawn O. Pearce" writes: > Signed-off-by: Shawn O. Pearce > --- > git-gui.sh | 10 +++++++++- > 1 files changed, 9 insertions(+), 1 deletions(-) > > diff --git a/git-gui.sh b/git-gui.sh > index 743b7d4..fa30ccc 100755 > --- a/git-gui.sh > +++ b/git-gui.sh > @@ -703,7 +703,15 @@ if {![file isdirectory $_gitdir]} { > error_popup "Git directory not found:\n\n$_gitdir" > exit 1 > } > -if {![is_enabled bare]} { > +if {$_prefix ne {}} { > + regsub -all {[^/]+/} $_prefix ../ cdup I don't like this approach. It assumes too much about the file system and cleanliness of paths. It does all of the following: /somedir/ -> /../ /somedir -> /somedir //server/somedir -> //../somedir /somedir//someother -> /..//someother and so on. It can't deal with directory symlinks properly. And the approach does not scale to Windows and other systems with diverging path syntaxes at all. Isn't it possible to move to the workdir root for the purpose of interpreting workdir root relative filenames? -- David Kastrup, Kriemhildstr. 15, 44793 Bochum ``` ## Shawn O. Pearce, 2007-08-25 08:15 Subject: Re: Problem with git-gui and relative directories Message-ID: <20070825081525.GB18160@spearce.org> URL: https://gitlist.dev/e/20070825081525.GB18160%40spearce.org In-Reply-To: <85veb4dqhv.fsf@lola.goethe.zz> ``` David Kastrup wrote: > "Shawn O. Pearce" writes: > > +if {$_prefix ne {}} { > > + regsub -all {[^/]+/} $_prefix ../ cdup > > I don't like this approach. It assumes too much about the file system > and cleanliness of paths. It does all of the following: > > /somedir/ -> /../ > /somedir -> /somedir > //server/somedir -> //../somedir > /somedir//someother -> /..//someother OK, well, uh. Go read the manpage for `git-rev-parse` and focus in particular on the `--show-prefix` option of that command. Its output is what we are using here. Its very well defined within Git, and should have the same definition on all operating systems, including MinGW. And if the MinGW folks did something funny its only a translation of / to \, in which case its easy enough for git-gui to make that decision on its own based on what type of git it is running on. > and so on. It can't deal with directory symlinks properly. Depends on your definition of properly. Here git-gui does exactly what the C programs in core Git do, which is counter to what the shell scripts do. And people prefer the C behavior of going up through the real parents, ignoring the symlinks that were used to get to the directory the user started git-gui within. > And the > approach does not scale to Windows and other systems with diverging > path syntaxes at all. Yes, it does. Because the format of --show-prefix is defined. > Isn't it possible to move to the workdir root for the purpose of > interpreting workdir root relative filenames? Because I have always disagreed with the GIT_WORK_TREE mess. I think its insane to say "I'm here in A, my repository is over there in B and my files are here in C... now go run status!". But apparently its a use case. I guess I could use `git rev-parse --show-cdup` and that would just handle this case. But that's YAFAE (Yet Another Fork And Exec). -- Shawn. ```