path: root/diff-lib.c
AgeCommit message (Collapse)Author
2011-08-01Merge branch 'jc/maint-reset-unmerged-path'Junio C Hamano
* jc/maint-reset-unmerged-path: reset [<commit>] paths...: do not mishandle unmerged paths
2011-07-14reset [<commit>] paths...: do not mishandle unmerged pathsJunio C Hamano
Because "diff --cached HEAD" showed an incorrect blob object name on the LHS of the diff, we ended up updating the index entry with bogus value, not what we read from the tree. Noticed by John Nowak. Signed-off-by: Junio C Hamano <>
2011-06-06Merge branch 'jk/diff-not-so-quick'Junio C Hamano
* jk/diff-not-so-quick: diff: futureproof "stop feeding the backend early" logic diff_tree: disable QUICK optimization with diff filter Conflicts: diff.c
2011-05-31diff-index --quiet: learn the "stop feeding the backend early" logicJunio C Hamano
A negative return from the unpack callback function usually means unpack failed for the entry and signals the unpack_trees() machinery to fail the entire merge operation, immediately and there is no other way for the callback to tell the machinery to exit early without reporting an error. This is what we usually want to make a merge all-or-nothing operation, but the machinery is also used for diff-index codepath by using a custom unpack callback function. And we do sometimes want to exit early without failing, namely when we are under --quiet and can short-cut the diff upon finding the first difference. Add "exiting_early" field to unpack_trees_options structure, to signal the unpack_trees() machinery that the negative return value is not signaling an error but an early return from the unpack_trees() machinery. As this by definition hasn't unpacked everything, discard the resulting index just like the failure codepath. Signed-off-by: Junio C Hamano <>
2011-05-31Merge remote-tracking branch 'ko/maint' into jc/diff-index-quick-exit-earlyJunio C Hamano
* ko/maint: (4352 commits) separate parens by a space to avoid confusing some shells Documentation/technical/api-diff.txt: correct name of diff_unmerge() read_gitfile_gently: use ssize_t to hold read result remove tests of always-false condition rerere.c: diagnose a corrupt MERGE_RR when hitting EOF between TAB and '\0' Git init/clone: remove short option -L and document --separate-git-dir do not read beyond end of malloc'd buffer git-svn: Fix git svn log --show-commit Git provide a copy of the LGPLv2.1 test core.gitproxy configuration copy_gecos: fix not adding nlen to len when processing "&" Update draft release notes to Documentation/git-fsck.txt: fix typo: unreadable -> unreachable send-pack: avoid deadlock on git:// push with failed pack-objects connect: let callers know if connection is a socket connect: treat generic proxy processes like ssh processes sideband_demux(): fix decl-after-stmt t3503: test cherry picking and reverting root commits ... Conflicts: diff.c
2011-05-31diff: futureproof "stop feeding the backend early" logicJunio C Hamano
Refactor the "do not stop feeding the backend early" logic into a small helper function and use it in both run_diff_files() and diff_tree() that has the stop-early optimization. We may later add other types of diffcore transformation that require to look at the whole result like diff-filter does, and having the logic in a single place is essential for longer term maintainability. Signed-off-by: Junio C Hamano <>
2011-05-13Merge branch 'jc/fix-diff-files-unmerged' into maintJunio C Hamano
* jc/fix-diff-files-unmerged: diff-files: show unmerged entries correctly diff: remove often unused parameters from diff_unmerge() diff.c: return filepair from diff_unmerge() test: use $_z40 from test-lib
2011-05-06Merge branch 'jc/fix-diff-files-unmerged'Junio C Hamano
* jc/fix-diff-files-unmerged: diff-files: show unmerged entries correctly diff: remove often unused parameters from diff_unmerge() diff.c: return filepair from diff_unmerge() test: use $_z40 from test-lib
2011-04-24diff-files: show unmerged entries correctlyJunio C Hamano
Earlier, e9c8409 (diff-index --cached --raw: show tree entry on the LHS for unmerged entries., 2007-01-05) taught the command to show the object name and the mode from the entry coming from the tree side when comparing a tree with an unmerged index. This is a belated companion patch that teaches diff-files to show the mode from the entry coming from the working tree side, when comparing an unmerged index and the working tree. Signed-off-by: Junio C Hamano <>
2011-04-24diff: remove often unused parameters from diff_unmerge()Junio C Hamano
e9c8409 (diff-index --cached --raw: show tree entry on the LHS for unmerged entries., 2007-01-05) added a <mode, object name> pair as parameters to this function, to store them in the pre-image side of an unmerged file pair. Now the function is fixed to return the filepair it queued, we can make the caller on the special case codepath to do so. Signed-off-by: Junio C Hamano <>
2011-03-23Merge branch 'jc/maint-diff-q-filter'Junio C Hamano
* jc/maint-diff-q-filter: diff --quiet: disable optimization when --diff-filter=X is used
2011-03-16diff --quiet: disable optimization when --diff-filter=X is usedJunio C Hamano
The code notices that the caller does not want any detail of the changes and only wants to know if there is a change or not by specifying --quiet. And it breaks out of the loop when it knows it already found any change. When you have a post-process filter (e.g. --diff-filter), however, the path we found to be different in the previous round and set HAS_CHANGES bit may end up being uninteresting, and there may be no output at the end. The optimization needs to be disabled for such case. Note that the f245194 (diff: change semantics of "ignore whitespace" options, 2009-05-22) already disables this optimization by refraining from setting HAS_CHANGES when post-process filters that need to inspect the contents of the files (e.g. -S, -w) in diff_change() function. Signed-off-by: Junio C Hamano <>
2011-02-03Convert ce_path_match() to use struct pathspecNguyễn Thái Ngọc Duy
Signed-off-by: Nguyễn Thái Ngọc Duy <> Signed-off-by: Junio C Hamano <>
2011-02-03struct rev_info: convert prune_data to struct pathspecNguyễn Thái Ngọc Duy
Signed-off-by: Nguyễn Thái Ngọc Duy <> Signed-off-by: Junio C Hamano <>
2011-02-03Convert struct diff_options to use struct pathspecNguyễn Thái Ngọc Duy
Signed-off-by: Nguyễn Thái Ngọc Duy <> Signed-off-by: Junio C Hamano <>
2010-08-09Submodules: Add the new "ignore" config option for diff and statusJens Lehmann
The new "ignore" config option controls the default behavior for "git status" and the diff family. It specifies under what circumstances they consider submodules as modified and can be set separately for each submodule. The command line option "--ignore-submodules=" has been extended to accept the new parameter "none" for both status and diff. Users that chose submodules to get rid of long work tree scanning times might want to set the "dirty" option for those submodules. This brings back the pre 1.7.0 behavior, where submodule work trees were never scanned for modifications. By using "--ignore-submodules=none" on the command line the status and diff commands can be told to do a full scan. This option can be set to the following values (which have the same name and meaning as for the "--ignore-submodules" option of status and diff): "all": All changes to the submodule will be ignored. "dirty": Only differences of the commit recorded in the superproject and the submodules HEAD will be considered modifications, all changes to the work tree of the submodule will be ignored. When using this value, the submodule will not be scanned for work tree changes at all, leading to a performance benefit on large submodules. "untracked": Only untracked files in the submodules work tree are ignored, a changed HEAD and/or modified files in the submodule will mark it as modified. "none" (which is the default): Either untracked or modified files in a submodules work tree or a difference between the subdmodules HEAD and the commit recorded in the superproject will make it show up as changed. This value is added as a new parameter for the "--ignore-submodules" option of the diff family and "git status" so the user can override the settings in the configuration. Signed-off-by: Jens Lehmann <> Signed-off-by: Junio C Hamano <>
2010-06-11Add optional parameters to the diff option "--ignore-submodules"Jens Lehmann
In some use cases it is not desirable that the diff family considers submodules that only contain untracked content as dirty. This may happen e.g. when the submodule is not under the developers control and not all build generated files have been added to .gitignore by the upstream developers. Using the "untracked" parameter for the "--ignore-submodules" option disables checking for untracked content and lets git diff report them as changed only when they have new commits or modified content. Sometimes it is not wanted to have submodules show up as changed when they just contain changes to their work tree. An example for that are scripts which just want to check for submodule commits while ignoring any changes to the work tree. Also users having large submodules known not to change might want to use this option, as the - sometimes substantial - time it takes to scan the submodule work tree(s) is saved. Signed-off-by: Jens Lehmann <> Signed-off-by: Junio C Hamano <>
2010-03-24Merge branch 'jl/submodule-diff-dirtiness'Junio C Hamano
* jl/submodule-diff-dirtiness: git status: ignoring untracked files must apply to submodules too git status: Fix false positive "new commits" output for dirty submodules Refactor dirty submodule detection in diff-lib.c git status: Show detailed dirty status of submodules in long format git diff --submodule: Show detailed dirty status of submodules
2010-03-14git status: ignoring untracked files must apply to submodules tooJens Lehmann
Since 1.7.0 submodules are considered dirty when they contain untracked files. But when git status is called with the "-uno" option, the user asked to ignore untracked files, so they must be ignored in submodules too. To achieve this, the new flag DIFF_OPT_IGNORE_UNTRACKED_IN_SUBMODULES is introduced. Signed-off-by: Jens Lehmann <> Signed-off-by: Junio C Hamano <>
2010-03-13git status: Fix false positive "new commits" output for dirty submodulesJens Lehmann
Testing if the output "new commits" should appear in the long format of "git status" is done by comparing the hashes of the diffpair. This always resulted in printing "new commits" for submodules that contained untracked or modified content, even if they did not contain new commits. The reason was that match_stat_with_submodule() did set the "changed" flag for dirty submodules, resulting in two->sha1 being set to the null_sha1 at the call sites, which indicates that new commits are present. This is changed so that when no new commits are present, the same object names are in the sha1 field for both sides of the filepair, and the working tree side will have the "dirty_submodule" flag set when appropriate. For a submodule to be seen as modified even when it just has a dirty work tree, some conditions had to be extended to also check for the "dirty_submodule" flag. Unfortunately the test case that should have found this bug had been changed incorrectly too. It is fixed and extended to test for other combinations too. Signed-off-by: Jens Lehmann <> Signed-off-by: Junio C Hamano <>
2010-03-13Refactor dirty submodule detection in diff-lib.cJens Lehmann
Moving duplicated code into the new function match_stat_with_submodule(). Replacing the implicit activation of detailed checks for the dirtiness of submodules when DIFF_FORMAT_PATCH was selected with explicitly setting the recently added DIFF_OPT_DIRTY_SUBMODULES option in diff_setup_done(). Signed-off-by: Jens Lehmann <> Signed-off-by: Junio C Hamano <>
2010-03-09revision: introduce setup_revision_optJunio C Hamano
So far the last parameter to setup_revisions() was to specify the default ref when the command line did not give any (typically "HEAD"). This changes it to take a pointer to a structure so that we can add other information without touching too many codepaths in later patches. There is no functionality change. Signed-off-by: Junio C Hamano <>
2010-03-08git status: Show detailed dirty status of submodules in long formatJens Lehmann
Since 1.7.0 there are three reasons a submodule is considered modified against the work tree: It contains new commits, modified content or untracked content. Lets show all reasons in the long format of git status, so the user can better asses the nature of the modification. This change does not affect the short and porcelain formats. Two new members are added to "struct wt_status_change_data" to store the information gathered by run_diff_files(). wt-status.c uses the new flag DIFF_OPT_DIRTY_SUBMODULES to tell diff-lib.c it wants to get detailed dirty information about submodules. A hint line for submodules is printed in the dirty header when dirty submodules are present. Signed-off-by: Jens Lehmann <> Signed-off-by: Junio C Hamano <>
2010-03-05git diff --submodule: Show detailed dirty status of submodulesJens Lehmann
When encountering a dirty submodule while doing "git diff --submodule" print an extra line for new untracked content and another for modified but already tracked content. And if the HEAD of the submodule is equal to the ref diffed against in the superproject, drop the output which would just show the same SHA1s and no commit message headlines. To achieve that, the dirty_submodule bitfield is expanded to two bits. The output of "git status" inside the submodule is parsed to set the according bits. Signed-off-by: Jens Lehmann <> Signed-off-by: Junio C Hamano <>
2010-01-27Merge branch 'jl/diff-submodule-ignore'Junio C Hamano
* jl/diff-submodule-ignore: Teach diff --submodule that modified submodule directory is dirty git diff: Don't test submodule dirtiness with --ignore-submodules Make ce_uptodate() trustworthy again
2010-01-25git diff: Don't test submodule dirtiness with --ignore-submodulesJens Lehmann
The diff family suppresses the output of submodule changes when requested but checks them nonetheless. But since recently submodules get examined for their dirtiness, which is rather expensive. There is no need to do that when the --ignore-submodules option is used, as the gathered information is never used anyway. Signed-off-by: Jens Lehmann <> Signed-off-by: Junio C Hamano <>
2010-01-25Merge branch 'jc/fix-tree-walk'Junio C Hamano
* jc/fix-tree-walk: read-tree --debug-unpack unpack-trees.c: look ahead in the index unpack-trees.c: prepare for looking ahead in the index Aggressive three-way merge: fix D/F case traverse_trees(): handle D/F conflict case sanely more D/F conflict tests tests: move convenience regexp to match object names to Conflicts: builtin-read-tree.c unpack-trees.c unpack-trees.h
2010-01-24Make ce_uptodate() trustworthy againJunio C Hamano
The rule has always been that a cache entry that is ce_uptodate(ce) means that we already have checked the work tree entity and we know there is no change in the work tree compared to the index, and nobody should have to double check. Note that false ce_uptodate(ce) does not mean it is known to be dirty---it only means we don't know if it is clean. There are a few codepaths (refresh-index and preload-index are among them) that mark a cache entry as up-to-date based solely on the return value from ie_match_stat(); this function uses lstat() to see if the work tree entity has been touched, and for a submodule entry, if its HEAD points at the same commit as the commit recorded in the index of the superproject (a submodule that is not even cloned is considered clean). A submodule is no longer considered unmodified merely because its HEAD matches the index of the superproject these days, in order to prevent people from forgetting to commit in the submodule and updating the superproject index with the new submodule commit, before commiting the state in the superproject. However, the patch to do so didn't update the codepath that marks cache entries up-to-date based on the updated definition and instead worked it around by saying "we don't trust the return value of ce_uptodate() for submodules." This makes ce_uptodate() trustworthy again by not marking submodule entries up-to-date. The next step _could_ be to introduce a few "in-core" flag bits to cache_entry structure to record "this entry is _known_ to be dirty", call is_submodule_modified() from ie_match_stat(), and use these new bits to avoid running this rather expensive check more than once, but that can be a separate patch. Signed-off-by: Junio C Hamano <>
2010-01-19Performance optimization for detection of modified submodulesJens Lehmann
In the worst case is_submodule_modified() got called three times for each submodule. The information we got from scanning the whole submodule tree the first time can be reused instead. New parameters have been added to diff_change() and diff_addremove(), the information is stored in a new member of struct diff_filespec. Its value is then reused instead of calling is_submodule_modified() again. When no explicit "-dirty" is needed in the output the call to is_submodule_modified() is not necessary when the submodules HEAD already disagrees with the ref of the superproject, as this alone marks it as modified. To achieve that, get_stat_data() got an extra argument. Signed-off-by: Jens Lehmann <> Signed-off-by: Junio C Hamano <>
2010-01-17Show submodules as modified when they contain a dirty work treeJens Lehmann
Until now a submodule only then showed up as modified in the supermodule when the last commit in the submodule differed from the one in the index or the diffed against commit of the superproject. A dirty work tree containing new untracked or modified files in a submodule was undetectable when looking at it from the superproject. Now git status and git diff (against the work tree) in the superproject will also display submodules as modified when they contain untracked or modified files, even if the compared ref matches the HEAD of the submodule. Signed-off-by: Jens Lehmann <> Signed-off-by: Nanako Shiraishi <> Signed-off-by: Junio C Hamano <>
2010-01-13Merge branch 'nd/sparse'Junio C Hamano
* nd/sparse: (25 commits) t7002: test for not using external grep on skip-worktree paths t7002: set test prerequisite "external-grep" if supported grep: do not do external grep on skip-worktree entries commit: correctly respect skip-worktree bit ie_match_stat(): do not ignore skip-worktree bit with CE_MATCH_IGNORE_VALID tests: rename duplicate t1009 sparse checkout: inhibit empty worktree Add tests for sparse checkout read-tree: add --no-sparse-checkout to disable sparse checkout support unpack-trees(): ignore worktree check outside checkout area unpack_trees(): apply $GIT_DIR/info/sparse-checkout to the final index unpack-trees(): "enable" sparse checkout and load $GIT_DIR/info/sparse-checkout unpack-trees.c: generalize verify_* functions unpack-trees(): add CE_WT_REMOVE to remove on worktree alone Introduce "sparse checkout" dir.c: export excluded_1() and add_excludes_from_file_1() excluded_1(): support exclude files in index unpack-trees(): carry skip-worktree bit over in merged_entry() Read .gitignore from index if it is skip-worktree Avoid writing to buffer in add_excludes_from_file_1() ... Conflicts: .gitignore Documentation/config.txt Documentation/git-update-index.txt Makefile entry.c t/
2010-01-07unpack-trees.c: look ahead in the indexJunio C Hamano
This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <>
2010-01-07unpack-trees.c: prepare for looking ahead in the indexJunio C Hamano
This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <>
2009-12-26Merge branch 'jc/1.7.0-diff-whitespace-only-status'Junio C Hamano
* jc/1.7.0-diff-whitespace-only-status: diff.c: fix typoes in comments Make test case number unique diff: Rename QUIET internal option to QUICK diff: change semantics of "ignore whitespace" options Conflicts: diff.h
2009-10-11diff-lib.c: fix misleading comments on oneway_diff()Junio C Hamano
20a16eb (unpack_trees(): fix diff-index regression., 2008-03-10) adjusted diff-index to the new world order since 34110cd (Make 'unpack_trees()' have a separate source and destination index, 2008-03-06). Callbacks are expected to return anything non-negative as "success", and instead of reporting how many index entries they have processed, they are expected to advance o->pos themselves. The code did so, but a stale comment was left behind. Signed-off-by: Junio C Hamano <>
2009-08-29Merge branch 'jc/shortstatus'Junio C Hamano
* jc/shortstatus: git commit --dry-run -v: show diff in color when asked Documentation/git-commit.txt: describe --dry-run wt-status: collect untracked files in a separate "collect" phase Make git_status_config() file scope static to builtin-commit.c wt-status: move wt_status_colors[] into wt_status structure wt-status: move many global settings to wt_status structure commit: --dry-run status: show worktree status of conflicted paths separately wt-status.c: rework the way changes to the index and work tree are summarized diff-index: keep the original index intact diff-index: report unmerged new entries
2009-08-24Teach Git to respect skip-worktree bit (reading part)Nguyễn Thái Ngọc Duy
grep: turn on --cached for files that is marked skip-worktree ls-files: do not check for deleted file that is marked skip-worktree update-index: ignore update request if it's skip-worktree, while still allows removing diff*: skip worktree version Signed-off-by: Nguyễn Thái Ngọc Duy <> Signed-off-by: Junio C Hamano <>
2009-08-12Prevent diff machinery from examining assume-unchanged entries on worktreeNguyễn Thái Ngọc Duy
Signed-off-by: Nguyễn Thái Ngọc Duy <> Signed-off-by: Junio C Hamano <>
2009-08-05diff-index: keep the original index intactJunio C Hamano
When comparing the index and a tree, we used to read the contents of the tree into stage #1 of the index and compared them with stage #0. In order not to lose sight of entries originally unmerged in the index, we hoisted them to stage #3 before reading the tree. Commit d1f2d7e (Make run_diff_index() use unpack_trees(), not read_tree(), 2008-01-19) changed all this. These days, we instead use unpack_trees() API to traverse the tree and compare the contents with the index, without modifying the index at all. There is no reason to hoist the unmerged entries to stage #3 anymore. Signed-off-by: Junio C Hamano <>
2009-08-05diff-index: report unmerged new entriesJunio C Hamano
Since an earlier change to diff-index by d1f2d7e (Make run_diff_index() use unpack_trees(), not read_tree(), 2008-01-19), we stopped reporting an unmerged path that does not exist in the tree, but we should. Signed-off-by: Junio C Hamano <>
2009-07-29diff: Rename QUIET internal option to QUICKJunio C Hamano
The option "QUIET" primarily meant "find if we have _any_ difference as quick as possible and report", which means we often do not even have to look at blobs if we know the trees are different by looking at the higher level (e.g. "diff-tree A B"). As a side effect, because there is no point showing one change that we happened to have found first, it also enables NO_OUTPUT and EXIT_WITH_STATUS options, making the end result look quiet. Rename the internal option to QUICK to reflect this better; it also makes grepping the source tree much easier, as there are other kinds of QUIET option everywhere. Signed-off-by: Junio C Hamano <>
2009-06-21Merge branch 'jc/cache-tree'Junio C Hamano
* jc/cache-tree: Avoid "diff-index --cached" optimization under --find-copies-harder Optimize "diff-index --cached" using cache-tree t4007: modernize the style cache-tree.c::cache_tree_find(): simplify internal API write-tree --ignore-cache-tree
2009-05-25Avoid "diff-index --cached" optimization under --find-copies-harderJunio C Hamano
When find-copies-harder is in effect, the diff frontends are expected to feed all paths, not just changed paths, to the diffcore, so that copy sources can be picked up. In such a case, not descending into subtrees using the cache-tree information is simply wrong. Signed-off-by: Junio C Hamano <>
2009-05-25Optimize "diff-index --cached" using cache-treeJunio C Hamano
When running "diff-index --cached" after making a change to only a small portion of the index, there is no point unpacking unchanged subtrees into the index recursively, only to find that all entries match anyway. Tweak unpack_trees() logic that is used to read in the tree object to catch the case where the tree entry we are looking at matches the index as a whole by looking at the cache-tree. As an exercise, after modifying a few paths in the kernel tree, here are a few numbers on my Athlon 64X2 3800+: (without patch, hot cache) $ /usr/bin/time git diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.07user 0.02system 0:00.09elapsed 102%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+9407minor)pagefaults 0swaps (with patch, hot cache) $ /usr/bin/time ../git.git/git-diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.02user 0.00system 0:00.02elapsed 103%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+2446minor)pagefaults 0swaps Cold cache numbers are very impressive, but it does not matter very much in practice: (without patch, cold cache) $ su root sh -c 'echo 3 >/proc/sys/vm/drop_caches' $ /usr/bin/time git diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.06user 0.17system 0:10.26elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k 247032inputs+0outputs (1172major+8237minor)pagefaults 0swaps (with patch, cold cache) $ su root sh -c 'echo 3 >/proc/sys/vm/drop_caches' $ /usr/bin/time ../git.git/git-diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.02user 0.01system 0:01.01elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k 18440inputs+0outputs (79major+2369minor)pagefaults 0swaps This of course helps "git status" as well. (without patch, hot cache) $ /usr/bin/time ../git.git/git-status >/dev/null 0.17user 0.18system 0:00.35elapsed 100%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+5336outputs (0major+10970minor)pagefaults 0swaps (with patch, hot cache) $ /usr/bin/time ../git.git/git-status >/dev/null 0.10user 0.16system 0:00.27elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+5336outputs (0major+3921minor)pagefaults 0swaps Signed-off-by: Junio C Hamano <>
2009-05-23Merge branch 'lt/maint-diff-reduce-lstat'Junio C Hamano
* lt/maint-diff-reduce-lstat: Teach 'git checkout' to preload the index contents Avoid unnecessary 'lstat()' calls in 'get_stat_data()'
2009-05-10Avoid unnecessary 'lstat()' calls in 'get_stat_data()'Linus Torvalds
When we ask get_stat_data() to get the mode and size of an index entry, we can avoid the lstat() call if we have marked the index entry as being uptodate due to earlier lstat() calls. This avoids a lot of unnecessary lstat() calls in eg 'git checkout', where the last phase shows the differences to the working tree (requiring a diff), but earlier phases have already verified the index. On the kernel repo (with a fast machine and everything cached), this changes timings of a nul 'git checkout' from - Before (best of ten): 0.14user 0.05system 0:00.19elapsed 100%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+13237minor)pagefaults 0swaps - After 0.11user 0.03system 0:00.15elapsed 98%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+13235minor)pagefaults 0swaps so it can obviously be noticeable, although equally obviously it's not a show-stopper on this particular machine. The difference is likely larger on slower machines, or with operating systems that don't do as good a job of name caching. Signed-off-by: Linus Torvalds <> Signed-off-by: Junio C Hamano <>
2009-03-18Merge branch 'kb/checkout-optim'Junio C Hamano
* kb/checkout-optim: Revert "lstat_cache(): print a warning if doing ping-pong between cache types" checkout bugfix: use stat.mtime instead of stat.ctime in two places Makefile: Set compiler switch for USE_NSEC Create USE_ST_TIMESPEC and turn it on for Darwin Not all systems use st_[cm]tim field for ns resolution file timestamp Record ns-timestamps if possible, but do not use it without USE_NSEC write_index(): update index_state->timestamp after flushing to disk verify_uptodate(): add ce_uptodate(ce) test make USE_NSEC work as expected fix compile error when USE_NSEC is defined check_updates(): effective removal of cache entries marked CE_REMOVE lstat_cache(): print a warning if doing ping-pong between cache types show_patch_diff(): remove a call to fstat() write_entry(): use fstat() instead of lstat() when file is open write_entry(): cleanup of some duplicated code create_directories(): remove some memcpy() and strchr() calls unlink_entry(): introduce schedule_dir_for_removal() lstat_cache(): swap func(length, string) into func(string, length) lstat_cache(): generalise longest_match_lstat_cache() lstat_cache(): small cleanup and optimisation
2009-02-11Generalize and libify index_is_dirty() to index_differs_from(...)Stephan Beyer
index_is_dirty() in builtin-revert.c checks if the index is dirty. This patch generalizes this function to check if the index differs from a revision, i.e. the former index_is_dirty() behavior can now be achieved by index_differs_from("HEAD", 0). The second argument "diff_flags" allows to set further diff option flags like DIFF_OPT_IGNORE_SUBMODULES. See DIFF_OPT_* macros in diff.h for a list. index_differs_from() seems to be useful for more than builtin-revert.c, so it is moved into diff-lib.c and also used in builtin-commit.c. Yet to mention: - "rev.abbrev = 0;" can be safely removed. This has no impact on performance or functioning of neither setup_revisions() nor run_diff_index(). - rev.pending.objects is free()d because this fixes a leak. (Also see 295dd2ad "Fix memory leak in traverse_commit_list") Mentored-by: Daniel Barkalow <> Mentored-by: Christian Couder <> Signed-off-by: Stephan Beyer <> Signed-off-by: Junio C Hamano <>
2009-02-10lstat_cache(): swap func(length, string) into func(string, length)Kjetil Barvik
Swap function argument pair (length, string) into (string, length) to conform with the commonly used order inside the GIT source code. Also, add a note about this fact into the coding guidelines. Signed-off-by: Kjetil Barvik <> Signed-off-by: Junio C Hamano <>
2009-01-11Cleanup of unused symcache variable inside diff-lib.cKjetil Barvik
Commit c40641b77b0274186fd1b327d5dc3246f814aaaf, 'Optimize symlink/directory detection' by Linus Torvalds, removed the 'char *symcache' parameter to the has_symlink_leading_path() function. This made all variables currently named 'symcache' inside diff-lib.c unnecessary. This also let us throw away the 'struct oneway_unpack_data', and instead directly use the 'struct rev_info *revs' member, which was the only member left after removal of the 'symcache[] array' member. The 'struct oneway_unpack_data' was introduced by the following commit: 948dd346 "diff-files: careful when inspecting work tree items" Impact: cleanup PATH_MAX bytes less memory stack usage in some cases Signed-off-by: Kjetil Barvik <> Signed-off-by: Junio C Hamano <>