1
0
Fork 0
mirror of https://github.com/git/git.git synced 2024-11-09 02:33:11 +01:00
git/git-gui/lib/choose_repository.tcl

1039 lines
23 KiB
Tcl
Raw Normal View History

git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
# git-gui Git repository chooser
# Copyright (C) 2007 Shawn Pearce
class choose_repository {
field top
field w
field w_body ; # Widget holding the center content
field w_next ; # Next button
field w_quit ; # Quit button
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
field o_cons ; # Console object (if active)
field w_types ; # List of type buttons in clone
field w_recentlist ; # Listbox containing recent repositories
field w_localpath ; # Entry widget bound to local_path
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
field done 0 ; # Finished picking the repository?
field local_path {} ; # Where this repository is locally
field origin_url {} ; # Where we are cloning from
field origin_name origin ; # What we shall call 'origin'
field clone_type hardlink ; # Type of clone to construct
field readtree_err ; # Error output from read-tree (if any)
field sorted_recent ; # recent repositories (sorted)
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
constructor pick {} {
global M1T M1B
make_toplevel top w
wm title $top [mc "Git Gui"]
if {$top eq {.}} {
menu $w.mbar -tearoff 0
$top configure -menu $w.mbar
set m_repo $w.mbar.repository
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
$w.mbar add cascade \
-label [mc Repository] \
-menu $m_repo
menu $m_repo
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
if {[is_MacOSX]} {
$w.mbar add cascade -label Apple -menu .mbar.apple
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
menu $w.mbar.apple
$w.mbar.apple add command \
-label [mc "About %s" [appname]] \
-command do_about
} else {
$w.mbar add cascade -label [mc Help] -menu $w.mbar.help
menu $w.mbar.help
$w.mbar.help add command \
-label [mc "About %s" [appname]] \
-command do_about
}
wm protocol $top WM_DELETE_WINDOW exit
bind $top <$M1B-q> exit
bind $top <$M1B-Q> exit
bind $top <Key-Escape> exit
} else {
wm geometry $top "+[winfo rootx .]+[winfo rooty .]"
bind $top <Key-Escape> [list destroy $top]
set m_repo {}
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
}
pack [git_logo $w.git_logo] -side left -fill y -padx 10 -pady 10
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
set w_body $w.body
set opts $w_body.options
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
frame $w_body
text $opts \
-cursor $::cursor_ptr \
-relief flat \
-background [$w_body cget -background] \
-wrap none \
-spacing1 5 \
-width 50 \
-height 3
pack $opts -anchor w -fill x
$opts tag conf link_new -foreground blue -underline 1
$opts tag bind link_new <1> [cb _next new]
$opts insert end [mc "Create New Repository"] link_new
$opts insert end "\n"
if {$m_repo ne {}} {
$m_repo add command \
-command [cb _next new] \
-accelerator $M1T-N \
-label [mc "New..."]
bind $top <$M1B-n> [cb _next new]
bind $top <$M1B-N> [cb _next new]
}
$opts tag conf link_clone -foreground blue -underline 1
$opts tag bind link_clone <1> [cb _next clone]
$opts insert end [mc "Clone Existing Repository"] link_clone
$opts insert end "\n"
if {$m_repo ne {}} {
$m_repo add command \
-command [cb _next clone] \
-accelerator $M1T-C \
-label [mc "Clone..."]
bind $top <$M1B-c> [cb _next clone]
bind $top <$M1B-C> [cb _next clone]
}
$opts tag conf link_open -foreground blue -underline 1
$opts tag bind link_open <1> [cb _next open]
$opts insert end [mc "Open Existing Repository"] link_open
$opts insert end "\n"
if {$m_repo ne {}} {
$m_repo add command \
-command [cb _next open] \
-accelerator $M1T-O \
-label [mc "Open..."]
bind $top <$M1B-o> [cb _next open]
bind $top <$M1B-O> [cb _next open]
}
$opts conf -state disabled
set sorted_recent [_get_recentrepos]
if {[llength $sorted_recent] > 0} {
if {$m_repo ne {}} {
$m_repo add separator
$m_repo add command \
-state disabled \
-label [mc "Recent Repositories"]
}
label $w_body.space
label $w_body.recentlabel \
-anchor w \
-text [mc "Open Recent Repository:"]
set w_recentlist $w_body.recentlist
text $w_recentlist \
-cursor $::cursor_ptr \
-relief flat \
-background [$w_body.recentlabel cget -background] \
-wrap none \
-width 50 \
-height 10
$w_recentlist tag conf link \
-foreground blue \
-underline 1
set home $::env(HOME)
if {[is_Cygwin]} {
set home [exec cygpath --windows --absolute $home]
}
set home "[file normalize $home]/"
set hlen [string length $home]
foreach p $sorted_recent {
set path $p
if {[string equal -length $hlen $home $p]} {
set p "~/[string range $p $hlen end]"
}
regsub -all "\n" $p "\\n" p
$w_recentlist insert end $p link
$w_recentlist insert end "\n"
if {$m_repo ne {}} {
$m_repo add command \
-command [cb _open_recent_path $path] \
-label " $p"
}
}
$w_recentlist conf -state disabled
$w_recentlist tag bind link <1> [cb _open_recent %x,%y]
pack $w_body.space -anchor w -fill x
pack $w_body.recentlabel -anchor w -fill x
pack $w_recentlist -anchor w -fill x
}
pack $w_body -fill x -padx 10 -pady 10
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
frame $w.buttons
set w_next $w.buttons.next
set w_quit $w.buttons.quit
button $w_quit \
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
-text [mc "Quit"] \
-command exit
pack $w_quit -side right -padx 5
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
pack $w.buttons -side bottom -fill x -padx 10 -pady 10
if {$m_repo ne {}} {
$m_repo add separator
$m_repo add command \
-label [mc Quit] \
-command exit \
-accelerator $M1T-Q
}
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
bind $top <Return> [cb _invoke_next]
bind $top <Visibility> "
[cb _center]
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
grab $top
focus $top
bind $top <Visibility> {}
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
"
wm deiconify $top
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
tkwait variable @done
if {$top eq {.}} {
eval destroy [winfo children $top]
}
}
proc _home {} {
if {[catch {set h $::env(HOME)}]
|| ![file isdirectory $h]} {
set h .
}
return $h
}
method _center {} {
set nx [winfo reqwidth $top]
set ny [winfo reqheight $top]
set rx [expr {([winfo screenwidth $top] - $nx) / 3}]
set ry [expr {([winfo screenheight $top] - $ny) / 3}]
wm geometry $top [format {+%d+%d} $rx $ry]
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
}
method _invoke_next {} {
if {[winfo exists $w_next]} {
uplevel #0 [$w_next cget -command]
}
}
proc _get_recentrepos {} {
set recent [list]
foreach p [get_config gui.recentrepo] {
if {[_is_git [file join $p .git]]} {
lappend recent $p
}
}
return [lsort $recent]
}
proc _unset_recentrepo {p} {
regsub -all -- {([()\[\]{}\.^$+*?\\])} $p {\\\1} p
git config --global --unset gui.recentrepo "^$p\$"
}
proc _append_recentrepos {path} {
set path [file normalize $path]
set recent [get_config gui.recentrepo]
if {[lindex $recent end] eq $path} {
return
}
set i [lsearch $recent $path]
if {$i >= 0} {
_unset_recentrepo $path
set recent [lreplace $recent $i $i]
}
lappend recent $path
git config --global --add gui.recentrepo $path
while {[llength $recent] > 10} {
_unset_recentrepo [lindex $recent 0]
set recent [lrange $recent 1 end]
}
}
method _open_recent {xy} {
set id [lindex [split [$w_recentlist index @$xy] .] 0]
set local_path [lindex $sorted_recent [expr {$id - 1}]]
_do_open2 $this
}
method _open_recent_path {p} {
set local_path $p
_do_open2 $this
}
method _next {action} {
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
destroy $w_body
if {![winfo exists $w_next]} {
button $w_next -default active
pack $w_next -side right -padx 5 -before $w_quit
}
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
_do_$action $this
}
method _write_local_path {args} {
if {$local_path eq {}} {
$w_next conf -state disabled
} else {
$w_next conf -state normal
}
}
method _git_init {} {
if {[catch {file mkdir $local_path} err]} {
error_popup [strcat \
[mc "Failed to create repository %s:" $local_path] \
"\n\n$err"]
return 0
}
if {[catch {cd $local_path} err]} {
error_popup [strcat \
[mc "Failed to create repository %s:" $local_path] \
"\n\n$err"]
return 0
}
if {[catch {git init} err]} {
error_popup [strcat \
[mc "Failed to create repository %s:" $local_path] \
"\n\n$err"]
return 0
}
_append_recentrepos [pwd]
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
set ::_gitdir .git
set ::_prefix {}
return 1
}
proc _is_git {path} {
if {[file exists [file join $path HEAD]]
&& [file exists [file join $path objects]]
&& [file exists [file join $path config]]} {
return 1
}
if {[is_Cygwin]} {
if {[file exists [file join $path HEAD]]
&& [file exists [file join $path objects.lnk]]
&& [file exists [file join $path config.lnk]]} {
return 1
}
}
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
return 0
}
proc _objdir {path} {
set objdir [file join $path .git objects]
if {[file isdirectory $objdir]} {
return $objdir
}
set objdir [file join $path objects]
if {[file isdirectory $objdir]} {
return $objdir
}
if {[is_Cygwin]} {
set objdir [file join $path .git objects.lnk]
if {[file isfile $objdir]} {
return [win32_read_lnk $objdir]
}
set objdir [file join $path objects.lnk]
if {[file isfile $objdir]} {
return [win32_read_lnk $objdir]
}
}
return {}
}
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
######################################################################
##
## Create New Repository
method _do_new {} {
$w_next conf \
-state disabled \
-command [cb _do_new2] \
-text [mc "Create"]
frame $w_body
label $w_body.h \
-font font_uibold \
-text [mc "Create New Repository"]
pack $w_body.h -side top -fill x -pady 10
pack $w_body -fill x -padx 10
frame $w_body.where
label $w_body.where.l -text [mc "Directory:"]
entry $w_body.where.t \
-textvariable @local_path \
-font font_diff \
-width 50
button $w_body.where.b \
-text [mc "Browse"] \
-command [cb _new_local_path]
set w_localpath $w_body.where.t
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
grid $w_body.where.l $w_body.where.t $w_body.where.b -sticky ew
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
pack $w_body.where -fill x
trace add variable @local_path write [cb _write_local_path]
bind $w_body.h <Destroy> [list trace remove variable @local_path write [cb _write_local_path]]
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
update
focus $w_body.where.t
}
method _new_local_path {} {
if {$local_path ne {}} {
set p [file dirname $local_path]
} else {
set p [_home]
}
set p [tk_chooseDirectory \
-initialdir $p \
-parent $top \
-title [mc "Git Repository"] \
-mustexist false]
if {$p eq {}} return
set p [file normalize $p]
if {![_new_ok $p]} {
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
return
}
set local_path $p
$w_localpath icursor end
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
}
method _do_new2 {} {
if {![_new_ok $local_path]} {
return
}
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
if {![_git_init $this]} {
return
}
set done 1
}
proc _new_ok {p} {
if {[file isdirectory $p]} {
if {[_is_git [file join $p .git]]} {
error_popup [mc "Directory %s already exists." $p]
return 0
}
} elseif {[file exists $p]} {
error_popup [mc "File %s already exists." $p]
return 0
}
return 1
}
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
######################################################################
##
## Clone Existing Repository
method _do_clone {} {
$w_next conf \
-state disabled \
-command [cb _do_clone2] \
-text [mc "Clone"]
frame $w_body
label $w_body.h \
-font font_uibold \
-text [mc "Clone Existing Repository"]
pack $w_body.h -side top -fill x -pady 10
pack $w_body -fill x -padx 10
set args $w_body.args
frame $w_body.args
pack $args -fill both
label $args.origin_l -text [mc "URL:"]
entry $args.origin_t \
-textvariable @origin_url \
-font font_diff \
-width 50
button $args.origin_b \
-text [mc "Browse"] \
-command [cb _open_origin]
grid $args.origin_l $args.origin_t $args.origin_b -sticky ew
label $args.where_l -text [mc "Directory:"]
entry $args.where_t \
-textvariable @local_path \
-font font_diff \
-width 50
button $args.where_b \
-text [mc "Browse"] \
-command [cb _new_local_path]
grid $args.where_l $args.where_t $args.where_b -sticky ew
set w_localpath $args.where_t
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
label $args.type_l -text [mc "Clone Type:"]
frame $args.type_f
set w_types [list]
lappend w_types [radiobutton $args.type_f.hardlink \
-state disabled \
-anchor w \
-text [mc "Standard (Fast, Semi-Redundant, Hardlinks)"] \
-variable @clone_type \
-value hardlink]
lappend w_types [radiobutton $args.type_f.full \
-state disabled \
-anchor w \
-text [mc "Full Copy (Slower, Redundant Backup)"] \
-variable @clone_type \
-value full]
lappend w_types [radiobutton $args.type_f.shared \
-state disabled \
-anchor w \
-text [mc "Shared (Fastest, Not Recommended, No Backup)"] \
-variable @clone_type \
-value shared]
foreach r $w_types {
pack $r -anchor w
}
grid $args.type_l $args.type_f -sticky new
grid columnconfigure $args 1 -weight 1
trace add variable @local_path write [cb _update_clone]
trace add variable @origin_url write [cb _update_clone]
bind $w_body.h <Destroy> "
[list trace remove variable @local_path write [cb _update_clone]]
[list trace remove variable @origin_url write [cb _update_clone]]
"
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
update
focus $args.origin_t
}
method _open_origin {} {
if {$origin_url ne {} && [file isdirectory $origin_url]} {
set p $origin_url
} else {
set p [_home]
}
set p [tk_chooseDirectory \
-initialdir $p \
-parent $top \
-title [mc "Git Repository"] \
-mustexist true]
if {$p eq {}} return
set p [file normalize $p]
if {![_is_git [file join $p .git]] && ![_is_git $p]} {
error_popup [mc "Not a Git repository: %s" [file tail $p]]
return
}
set origin_url $p
}
method _update_clone {args} {
if {$local_path ne {} && $origin_url ne {}} {
$w_next conf -state normal
} else {
$w_next conf -state disabled
}
if {$origin_url ne {} &&
( [_is_git [file join $origin_url .git]]
|| [_is_git $origin_url])} {
set e normal
if {[[lindex $w_types 0] cget -state] eq {disabled}} {
set clone_type hardlink
}
} else {
set e disabled
set clone_type full
}
foreach r $w_types {
$r conf -state $e
}
}
method _do_clone2 {} {
if {[file isdirectory $origin_url]} {
set origin_url [file normalize $origin_url]
}
if {$clone_type eq {hardlink} && ![file isdirectory $origin_url]} {
error_popup [mc "Standard only available for local repository."]
return
}
if {$clone_type eq {shared} && ![file isdirectory $origin_url]} {
error_popup [mc "Shared only available for local repository."]
return
}
if {$clone_type eq {hardlink} || $clone_type eq {shared}} {
set objdir [_objdir $origin_url]
if {$objdir eq {}} {
error_popup [mc "Not a Git repository: %s" [file tail $origin_url]]
return
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
}
}
set giturl $origin_url
if {[is_Cygwin] && [file isdirectory $giturl]} {
set giturl [exec cygpath --unix --absolute $giturl]
if {$clone_type eq {shared}} {
set objdir [exec cygpath --unix --absolute $objdir]
}
}
if {[file exists $local_path]} {
error_popup [mc "Location %s already exists." $local_path]
return
}
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
if {![_git_init $this]} return
set local_path [pwd]
if {[catch {
git config remote.$origin_name.url $giturl
git config remote.$origin_name.fetch +refs/heads/*:refs/remotes/$origin_name/*
} err]} {
error_popup [strcat [mc "Failed to configure origin"] "\n\n$err"]
return
}
destroy $w_body $w_next
switch -exact -- $clone_type {
hardlink {
set o_cons [status_bar::two_line $w_body]
pack $w_body -fill x -padx 10 -pady 10
$o_cons start \
[mc "Counting objects"] \
[mc "buckets"]
update
if {[file exists [file join $objdir info alternates]]} {
set pwd [pwd]
if {[catch {
file mkdir [gitdir objects info]
set f_in [open [file join $objdir info alternates] r]
set f_cp [open [gitdir objects info alternates] w]
fconfigure $f_in -translation binary -encoding binary
fconfigure $f_cp -translation binary -encoding binary
cd $objdir
while {[gets $f_in line] >= 0} {
if {[is_Cygwin]} {
puts $f_cp [exec cygpath --unix --absolute $line]
} else {
puts $f_cp [file normalize $line]
}
}
close $f_in
close $f_cp
cd $pwd
} err]} {
catch {cd $pwd}
_clone_failed $this [mc "Unable to copy objects/info/alternates: %s" $err]
return
}
}
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
set tolink [list]
set buckets [glob \
-tails \
-nocomplain \
-directory [file join $objdir] ??]
set bcnt [expr {[llength $buckets] + 2}]
set bcur 1
$o_cons update $bcur $bcnt
update
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
file mkdir [file join .git objects pack]
foreach i [glob -tails -nocomplain \
-directory [file join $objdir pack] *] {
lappend tolink [file join pack $i]
}
$o_cons update [incr bcur] $bcnt
update
foreach i $buckets {
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
file mkdir [file join .git objects $i]
foreach j [glob -tails -nocomplain \
-directory [file join $objdir $i] *] {
lappend tolink [file join $i $j]
}
$o_cons update [incr bcur] $bcnt
update
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
}
$o_cons stop
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
if {$tolink eq {}} {
info_popup [strcat \
[mc "Nothing to clone from %s." $origin_url] \
"\n" \
[mc "The 'master' branch has not been initialized."] \
]
destroy $w_body
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
set done 1
return
}
set i [lindex $tolink 0]
if {[catch {
file link -hard \
[file join .git objects $i] \
[file join $objdir $i]
} err]} {
info_popup [mc "Hardlinks are unavailable. Falling back to copying."]
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
set i [_copy_files $this $objdir $tolink]
} else {
set i [_link_files $this $objdir [lrange $tolink 1 end]]
}
if {!$i} return
destroy $w_body
}
full {
set o_cons [console::embed \
$w_body \
[mc "Cloning from %s" $origin_url]]
pack $w_body -fill both -expand 1 -padx 10
$o_cons exec \
[list git fetch --no-tags -k $origin_name] \
[cb _do_clone_tags]
}
shared {
set fd [open [gitdir objects info alternates] w]
fconfigure $fd -translation binary
puts $fd $objdir
close $fd
}
}
if {$clone_type eq {hardlink} || $clone_type eq {shared}} {
if {![_clone_refs $this]} return
set pwd [pwd]
if {[catch {
cd $origin_url
set HEAD [git rev-parse --verify HEAD^0]
} err]} {
_clone_failed $this [mc "Not a Git repository: %s" [file tail $origin_url]]
return 0
}
cd $pwd
_do_clone_checkout $this $HEAD
}
}
method _copy_files {objdir tocopy} {
$o_cons start \
[mc "Copying objects"] \
[mc "KiB"]
set tot 0
set cmp 0
foreach p $tocopy {
incr tot [file size [file join $objdir $p]]
}
foreach p $tocopy {
if {[catch {
set f_in [open [file join $objdir $p] r]
set f_cp [open [file join .git objects $p] w]
fconfigure $f_in -translation binary -encoding binary
fconfigure $f_cp -translation binary -encoding binary
while {![eof $f_in]} {
incr cmp [fcopy $f_in $f_cp -size 16384]
$o_cons update \
[expr {$cmp / 1024}] \
[expr {$tot / 1024}]
update
}
close $f_in
close $f_cp
} err]} {
_clone_failed $this [mc "Unable to copy object: %s" $err]
return 0
}
}
return 1
}
method _link_files {objdir tolink} {
set total [llength $tolink]
$o_cons start \
[mc "Linking objects"] \
[mc "objects"]
for {set i 0} {$i < $total} {} {
set p [lindex $tolink $i]
if {[catch {
file link -hard \
[file join .git objects $p] \
[file join $objdir $p]
} err]} {
_clone_failed $this [mc "Unable to hardlink object: %s" $err]
return 0
}
incr i
if {$i % 5 == 0} {
$o_cons update $i $total
update
}
}
return 1
}
method _clone_refs {} {
set pwd [pwd]
if {[catch {cd $origin_url} err]} {
error_popup [mc "Not a Git repository: %s" [file tail $origin_url]]
return 0
}
set fd_in [git_read for-each-ref \
--tcl \
{--format=list %(refname) %(objectname) %(*objectname)}]
cd $pwd
set fd [open [gitdir packed-refs] w]
fconfigure $fd -translation binary
puts $fd "# pack-refs with: peeled"
while {[gets $fd_in line] >= 0} {
set line [eval $line]
set refn [lindex $line 0]
set robj [lindex $line 1]
set tobj [lindex $line 2]
if {[regsub ^refs/heads/ $refn \
"refs/remotes/$origin_name/" refn]} {
puts $fd "$robj $refn"
} elseif {[string match refs/tags/* $refn]} {
puts $fd "$robj $refn"
if {$tobj ne {}} {
puts $fd "^$tobj"
}
}
}
close $fd_in
close $fd
return 1
}
method _do_clone_tags {ok} {
if {$ok} {
$o_cons exec \
[list git fetch --tags -k $origin_name] \
[cb _do_clone_HEAD]
} else {
$o_cons done $ok
_clone_failed $this [mc "Cannot fetch branches and objects. See console output for details."]
}
}
method _do_clone_HEAD {ok} {
if {$ok} {
$o_cons exec \
[list git fetch $origin_name HEAD] \
[cb _do_clone_full_end]
} else {
$o_cons done $ok
_clone_failed $this [mc "Cannot fetch tags. See console output for details."]
}
}
method _do_clone_full_end {ok} {
$o_cons done $ok
if {$ok} {
destroy $w_body
set HEAD {}
if {[file exists [gitdir FETCH_HEAD]]} {
set fd [open [gitdir FETCH_HEAD] r]
while {[gets $fd line] >= 0} {
if {[regexp "^(.{40})\t\t" $line line HEAD]} {
break
}
}
close $fd
}
catch {git pack-refs}
_do_clone_checkout $this $HEAD
} else {
_clone_failed $this [mc "Cannot determine HEAD. See console output for details."]
}
}
method _clone_failed {{why {}}} {
if {[catch {file delete -force $local_path} err]} {
set why [strcat \
$why \
"\n\n" \
[mc "Unable to cleanup %s" $local_path] \
"\n\n" \
$err]
}
if {$why ne {}} {
update
error_popup [strcat [mc "Clone failed."] "\n" $why]
}
}
method _do_clone_checkout {HEAD} {
if {$HEAD eq {}} {
info_popup [strcat \
[mc "No default branch obtained."] \
"\n" \
[mc "The 'master' branch has not been initialized."] \
]
set done 1
return
}
if {[catch {
git update-ref HEAD $HEAD^0
} err]} {
info_popup [strcat \
[mc "Cannot resolve %s as a commit." $HEAD^0] \
"\n $err" \
"\n" \
[mc "The 'master' branch has not been initialized."] \
]
set done 1
return
}
set o_cons [status_bar::two_line $w_body]
pack $w_body -fill x -padx 10 -pady 10
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
$o_cons start \
[mc "Creating working directory"] \
[mc "files"]
set readtree_err {}
set fd [git_read --stderr read-tree \
-m \
-u \
-v \
HEAD \
HEAD \
]
fconfigure $fd -blocking 0 -translation binary
fileevent $fd readable [cb _readtree_wait $fd]
}
method _readtree_wait {fd} {
set buf [read $fd]
$o_cons update_meter $buf
append readtree_err $buf
fconfigure $fd -blocking 1
if {![eof $fd]} {
fconfigure $fd -blocking 0
return
}
if {[catch {close $fd}]} {
set err $readtree_err
regsub {^fatal: } $err {} err
error_popup [strcat \
[mc "Initial file checkout failed."] \
"\n\n$err"]
return
}
set done 1
}
######################################################################
##
## Open Existing Repository
method _do_open {} {
$w_next conf \
-state disabled \
-command [cb _do_open2] \
-text [mc "Open"]
frame $w_body
label $w_body.h \
-font font_uibold \
-text [mc "Open Existing Repository"]
pack $w_body.h -side top -fill x -pady 10
pack $w_body -fill x -padx 10
frame $w_body.where
label $w_body.where.l -text [mc "Repository:"]
entry $w_body.where.t \
-textvariable @local_path \
-font font_diff \
-width 50
button $w_body.where.b \
-text [mc "Browse"] \
-command [cb _open_local_path]
grid $w_body.where.l $w_body.where.t $w_body.where.b -sticky ew
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
pack $w_body.where -fill x
trace add variable @local_path write [cb _write_local_path]
bind $w_body.h <Destroy> [list trace remove variable @local_path write [cb _write_local_path]]
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
update
focus $w_body.where.t
}
method _open_local_path {} {
if {$local_path ne {}} {
set p $local_path
} else {
set p [_home]
}
set p [tk_chooseDirectory \
-initialdir $p \
-parent $top \
-title [mc "Git Repository"] \
-mustexist true]
if {$p eq {}} return
set p [file normalize $p]
if {![_is_git [file join $p .git]]} {
error_popup [mc "Not a Git repository: %s" [file tail $p]]
return
}
set local_path $p
}
method _do_open2 {} {
if {![_is_git [file join $local_path .git]]} {
error_popup [mc "Not a Git repository: %s" [file tail $local_path]]
return
}
if {[catch {cd $local_path} err]} {
error_popup [strcat \
[mc "Failed to open repository %s:" $local_path] \
"\n\n$err"]
return
}
_append_recentrepos [pwd]
git-gui: Allow users to choose/create/clone a repository If we are started outside of a git repository than it is likely the user started us from some sort of desktop shortcut icon in the operating system. In such a case the user is expecting us to prompt them to locate the git repository they want to work on, or to help them make a new repository, or to clone one from an existing location. This is a very simple wizard that offers the user one of these three choices. When we clone a repository we always use the name `master` in the local repository, even if the remote side does not appear to point to that name. I chose this as a policy decision. Much of the Git documentation talks about `master` being the default branch in a repository and that's what git-init does too. If the remote side doesn't call its default branch `master` most users just don't care, they just want to use Git the way the documentation describes. Rather than relying on the git-clone Porcelain that ships with git we build the new repository ourselves and then obtain content by git-fetch. This technique simplifies the entire clone process to roughly: `git init && git fetch && git pull`. Today we use three passes with git-fetch; the first pass gets us the bulk of the objects and the branches, the second pass gets us the tags, and the final pass gets us the current value of HEAD to initialize the default branch. If the source repository is on the local disk we try to use a hardlink to connect the objects into the new clone as this can be many times faster than copying the objects or packing them and passing the data through a pipe to index-pack. Unlike git-clone we stick to pure Tcl [file link -hard] operation thus avoiding the need to fork a cpio process to setup the hardlinks. If hardlinks do not appear to be supported (e.g. filesystem doesn't allow them or we are crossing filesystem boundaries) we use file copying instead. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-22 09:47:43 +02:00
set ::_gitdir .git
set ::_prefix {}
set done 1
}
}