* jc/clone:
Move "no merge candidate" warning into git-pull
Use preprocessor constants for environment variable names.
Do not create $GIT_DIR/remotes/ directory anymore.
Introduce GIT_TEMPLATE_DIR
Revert "fix testsuite: make sure they use templates freshly built from the source"
fix testsuite: make sure they use templates freshly built from the source
git-clone: lose the traditional 'no-separate-remote' layout
git-clone: lose the artificial "first" fetch refspec
git-pull: refuse default merge without branch.*.merge
git-clone: use wildcard specification for tracking branches
index-pack usage of mmap() is unacceptably slower on many OSes other than Linux
It was reported by Randal L. Schwartz <merlyn@stonehenge.com> that
indexing the Linux repository ~150MB pack takes about an hour on OS x
while it's a minute on Linux. It seems that the OS X mmap()
implementation is more than 2 orders of magnitude slower than the Linux
one.
Linus proposed a patch replacing mmap() with pread() bringing index-pack
performance on OS X in line with the Linux one. The performances on
Linux also improved by a small margin.
Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
gitweb: Show '...' links in "summary" view only if there are more items
Show "..." links in "summary" view to shortlog, heads (if there are
any), and tags (if there are any) only if there are more items to show
than shown already.
This means that "..." link is shown below shortened shortlog if there
are more than 16 commits, "..." link below shortened heads list if
there are more than 16 heads refs (16 branches), "..." link below
shortened tags list if there are more than 16 tags.
Modified patch from Jakub to to apply cleanly to master, also preform
the same "..." link logic to the forks list.
Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Robert Fitzsimons <robfitz@273k.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
Remove COLLISION_CHECK from Makefile since it's not used.
It's rather misleading to have configuration options that don't do
anything. If someone adds collision checking they might also want to
restore this option.
Signed-off-by: Brian Gernhardt <benji@silverinsanity.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
Use preprocessor constants for environment variable names.
We broke the discipline Linus set up to allow compiler help us
avoid typos in environment names in the early days of git over
time. This defines a handful preprocessor constants for
environment variable names used in relatively core parts of the
system.
I've left out variable names specific to subsystems such as HTTP
and SSL as I do not think they are big problems.
* jc/test-clone: (35 commits)
Introduce GIT_TEMPLATE_DIR
Revert "fix testsuite: make sure they use templates freshly built from the source"
fix testsuite: make sure they use templates freshly built from the source
rerere: fix breakage of resolving.
Add config example with respect to branch
Add documentation for show-branch --topics
make git a bit less cryptic on fetch errors
make patch_delta() error cases a bit more verbose
racy-git: documentation updates.
show-ref: fix --exclude-existing
parse-remote::expand_refs_wildcard()
vim syntax: follow recent changes to commit template
show-ref: fix --verify --hash=length
show-ref: fix --quiet --verify
avoid accessing _all_ loose refs in git-show-ref --verify
git-fetch: Avoid reading packed refs over and over again
Teach show-branch how to show ref-log data.
markup fix in svnimport documentation.
Documentation: new option -P for git-svnimport
Fix mis-mark-up in git-merge-file.txt documentation
...
Revert "fix testsuite: make sure they use templates freshly built from the source"
This reverts commit 74d20040cafdced657efbf49795183d209a3a07b.
Version from Johannes to introduce GIT_TEMPLATE_DIR is simpler,
although I unconsciously stayed away from introducing yet another
environment variable.
blame: -b (blame.blankboundary) and --root (blame.showroot)
When blame.blankboundary is set (or -b option is given), commit
object names are blanked out in the "human readable" output
format for boundary commits.
When blame.showroot is not set (or --root is not given), the
root commits are treated as boundary commits. The code still
attributes the lines to them, but with -b their object names are
not shown.
If there are more than one branches to be deleted, failure on
one will no longer stop git-branch to process the next ones.
The command still reports failures by exitting non-zero status.
Signed-off-by: Quy Tonthat <qtonthat@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
I just twisted it not to check fast-forwardness with the current
branch when you are removing a tracking branch. Most likely,
removal of a tracking branch is not because you are "done with"
it (for a local branch, it usually means "you merged it up"),
but because you are not even interested in it. In other words,
remote tracking branches are more like tags than branches.
fix testsuite: make sure they use templates freshly built from the source
The initial t/trash repository for testing was created properly
but over time we gained many tests that create secondary test
repositories with init-db or clone and they were not careful
enough.
commit e2b70087 botched the RCS merge to git-merge-file conversion.
There is no command called "git merge-file" (yes, we are using safer
variant of Perl's system(3)).
Add a quick paragraph explaining the --topics option for show-branch.
The explanation is an abbreviated version of the commit message from d320a5437f8304cf9ea3ee1898e49d643e005738.
Signed-off-by: Brian Gernhardt <benji@silverinsanity.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
The remote server might not want to tell why it doesn't like us for
security reasons, but let's make the client report such error in a bit
less confusing way. The remote failure remains a mystery, but the local
message might be a bit less so.
[jc: with a gentle wording updates from Andy Parkins]
Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
It is especially important to distinguish between a malloc() failure
from all the other cases. An out of memory condition is much less
worrisome than a compatibility/corruption problem.
Also make test-delta compilable again.
Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
We've removed the workaround for runtime penalty that did not
exist in practice some time ago, but the technical paper that
proposed that change still said "we probably should do so".
* 'jn/web' (early part):
gitweb: Add "next" link to commit view
gitweb: Add title attribute to ref marker with full ref name
gitweb: Do not show difftree for merges in "commit" view
gitweb: SHA-1 in commit log message links to "object" view
gitweb: Hyperlink target of symbolic link in "tree" view (if possible)
gitweb: Add generic git_object subroutine to display object of any type
gitweb: Show target of symbolic link in "tree" view
gitweb: Don't use Content-Encoding: header in git_snapshot
git-fetch: Avoid reading packed refs over and over again
When checking which tags to fetch, the old code used to call
git-show-ref --verify for each remote tag. Since reading even
packed refs is not a cheap operation when there are a lot of
local refs, the code became quite slow.
This fixes it by teaching git-show-ref to filter out existing
refs using a new mode of operation of git-show-ref.
Default GIT_COMMITTER_NAME to login name in recieve-pack.
If GIT_COMMITTER_NAME is not available in receive-pack but reflogs
are enabled we would normally die out with an error message asking
the user to correct their environment settings.
Now that reflogs are enabled by default in (what we guessed to be)
non-bare Git repositories this may cause problems for some users
who don't have their full name in the gecos field and who don't
have access to the remote system to correct the problem.
So rather than die()'ing out in receive-pack when we try to log a
ref change and have no committer name we default to the username,
as obtained from the host's password database.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
When replacing an existing file A with a directory A that has a
file A/B in it in the index, 'update-index --replace --add A/B'
did not properly remove the file to make room for the new
directory.
There was a trivial logic error, most likely a cut & paste one,
dating back to quite early days of git.
When replacing an existing file A with a directory A that has a
file A/B in it in the index, 'git add' did not succeed because
it forgot to pass the allow-replace flag to add_cache_entry().
It might be safer to leave this as an error and require the user
to explicitly remove the existing A first before adding A/B
since it is an unusual case, but doing that automatically is
much easier to use.
update-index: make D/F conflict error a bit more verbose.
When you remove a directory D that has a tracked file D/F out of the
way to create a file D and try to "git update-index --add D", it used
to say "cannot add" which was not very helpful. This issues an extra
error message to explain the situation before the final "fatal" message.
Since D/F conflicts are relatively rare event, extra verbosity would
not make things too noisy.
git-clone: lose the traditional 'no-separate-remote' layout
Finally.
The separate-remote layout is so much more organized than
traditional and easier to work with especially when you need to
deal with remote repositories with multiple branches and/or you
need to deal with more than one remote repositories, and using
traditional layout for new repositories simply does not make
much sense.
Internally we still have code for 1:1 mappings to create a bare
clone; that is a good thing and will not go away.
git-clone: lose the artificial "first" fetch refspec
Now we lost the "first refspec is the one that is merged by default"
rule, there is no reason for clone to list the remote primary branch
in the config file explicitly anymore.
We still need it for the traditional layout for other reasons,
though.
git-pull: refuse default merge without branch.*.merge
Everybody hated the pull behaviour of merging the first branch
listed on remotes/* file (or remote.*.fetch config) into the
current branch. This finally corrects that UI wart by
forbidding "git pull" without an explicit branch name on the
command line or branch.$current.merge for the current branch.
The matching change to git-clone was made to prepare the default
branch.*.merge entry for the primary branch some time ago.
git-clone: use wildcard specification for tracking branches
This stops enumerating the set of branches found on the remote
side when a clone was made in the configuration file. Instead,
a single entry that maps each remote branch to the local
tracking branch for the remote under the same name is created.
Doing it this way not only shortens the configuration file, but
automatically adjusts to a new branch added on the remote side
after the clone is made.
Unfortunately this cannot be done for the traditional layout,
where we always need to special case the 'master' to 'origin'
mapping within the local branch namespace. But that is Ok; it
will be going away before v1.5.0.
We could also lose the "primary branch" mapping at the
beginning, but that has to wait until we implement the "forbid
'git pull' when we do not have branch.$current.merge for the
current branch" policy we earlier discussed. That should also
be in v1.5.0
Add a kind of "next" view in the bottom part of navigation bar for
"commit" view, similar to what was added for "commitdiff" view in
commit 151602df00b8e5c5b4a8193f59a94b85f9b5aebc
'gitweb: Add "next" link to commitdiff view'
For "commit" view for single parent commit:
(parent: _commit_)
For "commit" view for merge (multi-parent) commit:
(merge: _commit_ _commit_ ...)
For "commit" view for root (parentless) commit
(initial)
where _link_ denotes hyperlink. SHA1 of commit is shortened
to 7 characters on display.
While at it, remove leftovers from commit cae1862a by Petr Baudis:
'gitweb: More per-view navigation bar links'
namely the "blame" link if there exist $file_name and commit has a
parent; it was added in git_commit probably by mistake. The rest
of what mentioned commit added for git_commit was removed in
commit 6e0e92fda893311ff5af91836e5007bf6bbd4a21 by Luben Tuikov:
'gitweb: Do not print "log" and "shortlog" redundantly in commit view'
(which should have probably removed also this "blame" link removed now).
Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
gitweb: Add title attribute to ref marker with full ref name
Add title attribute, which will be shown as popup on mouseover in
graphical web browsers, with full name of ref, including part (type)
removed from the name of ref itself. This is useful to see that this
strange ref is StGIT ref, or it is remote branch, or it is lightweigh
tag (with branch-like name).
Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
gitweb: Do not show difftree for merges in "commit" view
Do not show difftree against first parent for merges (commits with
more than one parent) in "commit" view, because it usually is
misleading. git-show and git-whatchanged doesn't show diff for merges
either.
Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
Align section headers of 'git status' to new 'git add'.
Now that 'git add' is considered a first-class UI for 'update-index'
and that the 'git add' documentation states "Even modified files
must be added to the set of changes about to be committed" we should
make the output of 'git status' align with that documentation and
common usage.
So now we see a status output such as:
# Added but not yet committed:
# (will commit)
#
# new file: x
#
# Changed but not added:
# (use "git add file1 file2" to include for commit)
#
# modified: x
#
# Untracked files:
# (use "git add" on files to include for commit)
#
# y
which just reads better in the context of using 'git add' to
manipulate a commit (and not a checkin, whatever the heck that is).
We also now support 'color.status.added' as an alias for the existing
'color.status.updated', as this alias more closely aligns with the
current output and documentation.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Suggest use of "git add file1 file2" when there is nothing to commit.
If a user modifies files and runs 'git commit' (without the very
useful -a option) and they have not yet updated the index they
are probably coming from another SCM-like tool which would perform
the same as 'git commit -a' in this case. Showing the user their
current status and a final line of "nothing to commit" is not very
reassuring, as the user might believe that Git did not recognize
their files were modified.
Instead we can suggest as part of the 'nothing to commit' message
that the user invoke 'git add' to add files to their next commit.
Suggested by Andy Parkins' Git 'niggles' list
(<200612132237.10051.andyparkins@gmail.com>).
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Make git-diff documentation use [--] when it should.
Two of the cases has "[--] [<path>...]" and two had "-- [<path>...]".
Not terribly consistent and potentially confusing. Also add "[--]" to
the synopsis so that it's obvious you can use it from the very
beginning.
Signed-off-by: Brian Gernhardt <benji@silverinsanity.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
For multivars, the "git-repo-config name value ^$" is useful but
nonintuitive and troublesome to do repeatedly (since the value is not
at the end of the command line). This commit simply adds an --add
option that adds a new value to a multivar. Particularly useful for
tracking a new branch on a remote:
Now that 'git show' accepts ref:path as an argument to specify a
tree or blob we should use the same completion logic as we support
for cat-file's object identifier.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Enable reflogs by default in any repository with a working directory.
New and experienced Git users alike are finding out too late that
they forgot to enable reflogs in the current repository, and cannot
use the information stored within it to recover from an incorrectly
entered command such as `git reset --hard HEAD^^^` when they really
meant HEAD^^ (aka HEAD~2).
So enable reflogs by default in all future versions of Git, unless
the user specifically disables it with:
[core]
logAllRefUpdates = false
in their .git/config or ~/.gitconfig.
We only enable reflogs in repositories that have a working directory
associated with them, as shared/bare repositories do not have
an easy means to prune away old log entries, or may fail logging
entirely if the user's gecos information is not valid during a push.
This heuristic was suggested on the mailing list by Junio.
Documentation was also updated to indicate the new default behavior.
We probably should start to teach usuing the reflog to recover
from mistakes in some of the tutorial material, as new users are
likely to make a few along the way and will feel better knowing
they can recover from them quickly and easily, without fsck-objects'
lost+found features.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Provide more meaningful output from 'git init-db'.
Back in the old days of Git when people messed around with their
GIT_DIR environment variable more often it was nice to know whether
or not git-init-db created a .git directory or used GIT_DIR.
As most users at that time were rather technical UNIXy folk the
message "defaulting to local storage area" made sense to some and
seemed reasonable.
But it doesn't really convey any meaning to the new Git user,
as they don't know what a 'local storage area is' nor do they
know enough about Git to care. It also really doesn't tell the
experienced Git user a whole lot about the command they just ran,
especially if they might be reinitializing an existing repository
(e.g. to update hooks).
So now we print out what we did ("Initialized empty" or
"Reinitialized existing"), what type of repository ("" or "shared"),
and what location the repository will be in ("$GIT_DIR").
Suggested in part by Andy Parkins in his Git 'niggles' list
(<200612132237.10051.andyparkins@gmail.com>).
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
make commit message a little more consistent and conforting
It is nicer to let the user know when a commit succeeded all the time,
not only the first time. Also the commit sha1 is much more useful than
the tree sha1 in this case.
This patch also introduces a -q switch to supress this message as well
as the summary of created/deleted files.
Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Avoid accessing a slow working copy during diffcore operations.
The Cygwin folks have done a fine job at creating a POSIX layer
on Windows That Just Works(tm). However it comes with a penalty;
accessing files in the working tree by way of stat/open/mmap can
be slower for diffcore than inflating the data from a blob which
is stored in a packfile.
This performance problem is especially an issue in merge-recursive
when dealing with nearly 7000 added files, as we are loading
each file's content from the working directory to perform rename
detection. I have literally seen (and sadly watched) paint dry in
less time than it takes for merge-recursive to finish such a merge.
On the other hand this very same merge runs very fast on Solaris.
If Git is compiled with NO_FAST_WORKING_DIRECTORY set then we will
avoid looking at the working directory when the blob in question
is available within a packfile and the caller doesn't need the data
unpacked into a temporary file.
We don't use loose objects as they have the same open/mmap/close
costs as the working directory file access, but have the additional
CPU overhead of needing to inflate the content before use. So it
is still faster to use the working tree file over the loose object.
If the caller needs the file data unpacked into a temporary file
its likely because they are going to call an external diff program,
passing the file as a parameter. In this case reusing the working
tree file will be faster as we don't need to inflate the data and
write it out to a temporary file.
The NO_FAST_WORKING_DIRECTORY feature is enabled by default on
Cygwin, as that is the platform which currently appears to benefit
the most from this option.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Sometimes it is asked on the list how to revert selected path in
the index from a tree, most often HEAD, without affecting the
files in the working tree. A similar operation that also
affects the working tree files has been available in the form of
"git checkout <tree> -- <paths>...".
By definition --soft would never affect either the index nor the
working tree files, and --hard is the way to make the working
tree files as close to pristine, so this new option is available
only for the default --mixed case.
Bypass expensive content comparsion during rename detection.
When comparing file contents during the second loop through a rename
detection attempt we can skip the expensive byte-by-byte comparsion
if both source and destination files have valid SHA1 values. This
improves performance by avoiding either an expensive open/mmap to
read the working tree copy, or an expensive inflate of a blob object.
Unfortunately we still have to at least initialize the sizes of the
source and destination files even if the SHA1 values don't match.
Failing to initialize the sizes causes a number of test cases to fail
and start reporting different copy/rename behavior than was expected.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Porcelain documentation should talk in terms of end-user workflow, not
in terms of implementation details. Do not suggest update-index, but
git-add instead. Explain differences among 0-, 1- and 2-tree cases
not as differences of number of trees given to the command, but say
why user would want to give these number of trees to the command in
what situation.
git-blame: show lines attributed to boundary commits differently.
When blaming with revision ranges, often many lines are attributed
to different commits at the boundary, but they are not interesting
for the purpose of finding project history during that revision range.
This outputs the lines blamed on boundary commits differently. When
showing "human format" output, their SHA-1 are shown with '^' prefixed.
In "porcelain format", the commit will be shown with an extra attribute
line "boundary".
* jc/read-tree-ignore:
read-tree: document --exclude-per-directory
Loosen "working file will be lost" check in Porcelain-ish
read-tree: further loosen "working file will be lost" check.
* np/addcommit:
git-commit: allow --only to lose what was staged earlier.
Documentation/git-commit: rewrite to make it more end-user friendly.
make 'git add' a first class user friendly interface to the index
* lh/branch-rename:
git-branch: let caller specify logmsg
rename_ref: use lstat(2) when testing for symlink
git-branch: add options and tests for branch renaming
Showing local and remote branches in green and red was simply
overkill, as all we wanted was to make it easy to tell them
apart (local ones can be built on top by committing, but the
remote tracking ones can't).
Use plain coloring for local branches and paint remotes in red.
* js/merge:
merge-recursive: add/add really is modify/modify with an empty base
Get rid of the dependency on RCS' merge program
merge-file: support -p and -q; fix compile warnings
Add builtin merge-file, a minimal replacement for RCS merge
xdl_merge(): fix and simplify conflict handling
xdl_merge(): fix thinko
xdl_merge(): fix an off-by-one bug
merge-recursive: use xdl_merge().
xmerge: make return value from xdl_merge() more usable.
xdiff: add xdl_merge()
"git push $URL HEAD~6" created a bogus ref HEAD~6 immediately
under $GIT_DIR of the remote repository. While we should keep
refspecs that have arbitrary extended SHA-1 expression on the
source side working (e.g. "HEAD~6:refs/tags/yesterday"), we
should not create bogus ref on the other end.