summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2013-07-12t0008: avoid SIGPIPE race condition on fifoLibravatar Jeff King1-2/+10
To test check-ignore's --stdin feature, we use two fifos to send and receive data. We carefully keep a descriptor to its input open so that it does not receive EOF between input lines. However, we do not do the same for its output. That means there is a potential race condition in which check-ignore has opened the output pipe once (when we read the first line), and then writes the second line before we have re-opened the pipe. In that case, check-ignore gets a SIGPIPE and dies. The outer shell then tries to open the output fifo but blocks indefinitely, because there is no writer. We can fix it by keeping a descriptor open through the whole procedure. This should also help if check-ignore dies for any other reason (we would already have opened the fifo and would therefore not block, but just get EOF on read). However, we are technically still susceptible to check-ignore dying early, before we have opened the fifo. This is an unlikely race and shouldn't generally happen in practice, though, so we can hopefully ignore it. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-04-29t0008: use named pipe (FIFO) to test check-ignore streamingLibravatar Adam Spiers1-21/+17
sleeps in the check-ignore test suite are not ideal since they can fail when the system is under load, or when a tool like valgrind is used which drastically alters the timing. Therefore we replace them with a more robust solution using a named pipe (FIFO). Thanks to Jeff King for coming up with the redirection wizardry required to make this work. http://article.gmane.org/gmane.comp.version-control.git/220916 Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-04-11Documentation: add caveats about I/O buffering for check-{attr,ignore}Libravatar Adam Spiers3-2/+13
check-attr and check-ignore have the potential to deadlock callers which do not read back the output in real-time. For example, if a caller writes N paths out and then reads N lines back in, it risks becoming blocked on write() to check-*, and check-* is blocked on write back to the caller. Somebody has to buffer; the pipe buffers provide some leeway, but they are limited. Thanks to Peff for pointing this out: http://article.gmane.org/gmane.comp.version-control.git/220534 Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-04-11check-ignore: allow incremental streaming of queries via --stdinLibravatar Adam Spiers2-15/+28
Some callers, such as the git-annex web assistant, find it useful to invoke git check-ignore as a persistent background process, which can then have queries fed to its STDIN at any point, and the corresponding response consumed from its STDOUT. For this we need to invoke check_ignore() once per line of standard input, and flush standard output after each result. The above use case suggests that empty STDIN is actually a reasonable scenario (e.g. when the caller doesn't know in advance whether any queries need to be fed to the background process until after it's already started), so we make the minor behavioural change that "no pathspec given." is no longer emitted in when STDIN is empty. Even though check_ignore() could now be changed to operate on a single pathspec, we keep it operating on an array of pathspecs since that is a more convenient way of consuming the existing pathspec API. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-04-11check-ignore: move setup into cmd_check_ignore()Libravatar Adam Spiers1-20/+21
Initialisation of the dir_struct and path_exclude_check structs was previously done within check_ignore(). This was acceptable since check_ignore() was only called once per check-ignore invocation; however the next commit will convert it into an inner loop which is called once per line of STDIN when --stdin is given. Therefore moving the initialisation code out into cmd_check_ignore() ensures that initialisation is still only performed once per check-ignore invocation, and consequently that the output is identical whether pathspecs are provided as CLI arguments or via STDIN. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-04-11check-ignore: add -n / --non-matching optionLibravatar Adam Spiers3-49/+134
If `-n` or `--non-matching` are specified, non-matching pathnames will also be output, in which case all fields in each output record except for <pathname> will be empty. This can be useful when running check-ignore as a background process, so that files can be incrementally streamed to STDIN, and for each of these files, STDOUT will indicate whether that file matched a pattern or not. (Without this option, it would be impossible to tell whether the absence of output for a given file meant that it didn't match any pattern, or that the result simply hadn't been flushed to STDOUT yet.) Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-04-11t0008: remove duplicated test fixture dataLibravatar Adam Spiers1-15/+1
The expected contents of STDOUT for the final --stdin tests can be derived from the expected contents of STDOUT for the same tests when --verbose is given, in the same way that test_expect_success_multi derives this for earlier tests. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-02-19name-hash: allow hashing an empty stringLibravatar Junio C Hamano3-3/+8
Usually we do not pass an empty string to the function hash_name() because we almost always ask for hash values for a path that is a candidate to be added to the index. However, check-ignore (and most likely check-attr, but I didn't check) apparently has a callchain to ask the hash value for an empty path when it was given a "." from the top-level directory to ask "Is the path . excluded by default?" Make sure that hash_name() does not overrun the end of the given pathname even when it is empty. Remove a sweep-the-issue-under-the-rug conditional in check-ignore that avoided to pass an empty string to the callchain while at it. It is a valid question to ask for check-ignore if the top-level is set to be ignored by default, even though the answer is most likely no, if only because there is currently no way to specify such an entry in the .gitignore file. But it is an unusual thing to ask and it is not worth optimizing for it by special casing at the top level of the call chain. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-02-19t0008: document test_expect_success_multiLibravatar Adam Spiers1-0/+10
test_expect_success_multi() helper function warrants some explanation, since at first sight it may seem like generic test framework plumbing, but is in fact specific to testing check-ignore, and allows more thorough testing of the various output formats without significantly increase the size of t0008. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-16clean.c, ls-files.c: respect encapsulation of exclude_list_groupsLibravatar Adam Spiers2-8/+13
Consumers of the dir.c traversal API should avoid assuming knowledge of the internal implementation of exclude_list_groups. Therefore when adding items to an exclude list, it should be accessed via the pointer returned from add_exclude_list(), rather than by referencing a location within dir.exclude_list_groups[EXC_CMDL]. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-10t0008: avoid brace expansionLibravatar René Scharfe1-2/+7
Brace expansion is a shell feature that's not required by POSIX and not supported by dash nor NetBSD's sh. Explicitly list all combinations instead. Also avoid calling touch by creating the test files with a redirection instead, as suggested by Junio. Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06add git-check-ignore sub-commandLibravatar Adam Spiers11-3/+905
This works in a similar manner to git-check-attr. Thanks to Jeff King and Junio C Hamano for the idea: http://thread.gmane.org/gmane.comp.version-control.git/108671/focus=108815 Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06setup.c: document get_pathspec()Libravatar Adam Spiers1-0/+19
Since we have just created a new pathspec-handling library, now is a good time to add some comments explaining get_pathspec(). Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06add.c: extract new die_if_path_beyond_symlink() for reuseLibravatar Adam Spiers3-4/+19
This will be reused by a new git check-ignore command. Also document validate_pathspec(). Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06add.c: extract check_path_for_gitlink() from treat_gitlinks() for reuseLibravatar Adam Spiers3-18/+38
Extract the body of the for loop in treat_gitlinks() into a separate check_path_for_gitlink() function so that it can be reused elsewhere. This paves the way for a new check-ignore sub-command. Also document treat_gitlinks(). Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06pathspec.c: rename newly public functions for clarityLibravatar Adam Spiers3-12/+13
Perform the following function renames to make it explicit that these pathspec handling functions are for matching against the index, rather than against a tree or the working directory. - fill_pathspec_matches() -> add_pathspec_matches_against_index() - find_used_pathspec() -> find_pathspecs_matching_against_index() Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06add.c: move pathspec matchers into new pathspec.c for reuseLibravatar Adam Spiers4-33/+67
Extract the following functions from builtin/add.c to pathspec.c, in preparation for reuse by a new git check-ignore command: - fill_pathspec_matches() - find_used_pathspec() The functions being extracted are not changed in any way, except removal of the 'static' qualifier. Also add comments documenting these newly public functions, including clarifications that they operate on the index. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06add.c: remove unused argument from validate_pathspec()Libravatar Adam Spiers1-3/+3
The 'argc' argument passed to validate_pathspec() was never used. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06dir.c: improve docs for match_pathspec() and match_pathspec_depth()Libravatar Adam Spiers2-12/+32
Fix a grammatical issue in the description of these functions, and make it more obvious how and why seen[] can be reused across multiple invocations. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06dir.c: provide clear_directory() for reclaiming dir_struct memoryLibravatar Adam Spiers3-0/+33
By the end of a directory traversal, a dir_struct instance will typically contains pointers to various data structures on the heap. clear_directory() provides a convenient way to reclaim that memory. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06dir.c: keep track of where patterns came fromLibravatar Adam Spiers4-12/+44
For exclude patterns read in from files, the filename is stored in the exclude list, and the originating line number is stored in the individual exclude (counting starting at 1). For exclude patterns provided on the command line, a string describing the source of the patterns is stored in the exclude list, and the sequence number assigned to each exclude pattern is negative, with counting starting at -1. So for example the 2nd pattern provided via --exclude would be numbered -2. This allows any future consumers of that data to easily distinguish between exclude patterns from files vs. from the CLI. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-01-06dir.c: use a single struct exclude_list per source of excludesLibravatar Adam Spiers6-39/+86
Previously each exclude_list could potentially contain patterns from multiple sources. For example dir->exclude_list[EXC_FILE] would typically contain patterns from .git/info/exclude and core.excludesfile, and dir->exclude_list[EXC_DIRS] could contain patterns from multiple per-directory .gitignore files during directory traversal (i.e. when dir->exclude_stack was more than one item deep). We split these composite exclude_lists up into three groups of exclude_lists (EXC_CMDL / EXC_DIRS / EXC_FILE as before), so that each exclude_list now contains patterns from a single source. This will allow us to cleanly track the origin of each pattern simply by adding a src field to struct exclude_list, rather than to struct exclude, which would make memory management of the source string tricky in the EXC_DIRS case where its contents are dynamically generated. Similarly, by moving the filebuf member from struct exclude_stack to struct exclude_list, it allows us to track and subsequently free memory buffers allocated during the parsing of all exclude files, rather than only tracking buffers allocated for files in the EXC_DIRS group. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-28dir.c: rename free_excludes() to clear_exclude_list()Libravatar Adam Spiers3-3/+7
It is clearer to use a 'clear_' prefix for functions which empty and deallocate the contents of a data structure without freeing the structure itself, and a 'free_' prefix for functions which also free the structure itself. http://article.gmane.org/gmane.comp.version-control.git/206128 Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-28dir.c: refactor is_path_excluded()Libravatar Adam Spiers2-9/+41
In a similar way to the previous commit, this extracts a new helper function last_exclude_matching_path() which return the last exclude_list element which matched, or NULL if no match was found. is_path_excluded() becomes a wrapper around this, and just returns 0 or 1 depending on whether any matching exclude_list element was found. This allows callers to find out _why_ a given path was excluded, rather than just whether it was or not, paving the way for a new git sub-command which allows users to test their exclude lists from the command line. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-28dir.c: refactor is_excluded()Libravatar Adam Spiers1-9/+29
In a similar way to the previous commit, this extracts a new helper function last_exclude_matching() which returns the last exclude_list element which matched, or NULL if no match was found. is_excluded() becomes a wrapper around this, and just returns 0 or 1 depending on whether any matching exclude_list element was found. This allows callers to find out _why_ a given path was excluded, rather than just whether it was or not, paving the way for a new git sub-command which allows users to test their exclude lists from the command line. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-28dir.c: refactor is_excluded_from_list()Libravatar Adam Spiers1-9/+28
The excluded function uses a new helper function called last_exclude_matching_from_list() to perform the inner loop over all of the exclude patterns. The helper just tells us whether the path is included, excluded, or undecided. However, it may be useful to know _which_ pattern was triggered. So let's pass out the entire exclude match, which contains the status information we were already passing out. Further patches can make use of this. This is a modified forward port of a patch from 2009 by Jeff King: http://article.gmane.org/gmane.comp.version-control.git/108815 Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-28dir.c: rename excluded() to is_excluded()Libravatar Adam Spiers3-8/+8
Continue adopting clearer names for exclude functions. This is_* naming pattern for functions returning booleans was discussed here: http://thread.gmane.org/gmane.comp.version-control.git/204661/focus=204924 Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-28dir.c: rename excluded_from_list() to is_excluded_from_list()Libravatar Adam Spiers3-10/+13
Continue adopting clearer names for exclude functions. This 'is_*' naming pattern for functions returning booleans was discussed here: http://thread.gmane.org/gmane.comp.version-control.git/204661/focus=204924 Also adjust their callers as necessary. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-28dir.c: rename path_excluded() to is_path_excluded()Libravatar Adam Spiers5-6/+6
Start adopting clearer names for exclude functions. This 'is_*' naming pattern for functions returning booleans was agreed here: http://thread.gmane.org/gmane.comp.version-control.git/204661/focus=204924 Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-28dir.c: rename cryptic 'which' variable to more consistent nameLibravatar Adam Spiers2-7/+7
'el' is only *slightly* less cryptic, but is already used as the variable name for a struct exclude_list pointer in numerous other places, so this reduces the number of cryptic variable names in use by one :-) Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-28Improve documentation and comments regarding directory traversal APILibravatar Adam Spiers3-6/+37
traversal API has a few potentially confusing properties. These comments clarify a few key aspects and will hopefully make it easier to understand for other newcomers in the future. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-28api-directory-listing.txt: update to match codeLibravatar Adam Spiers1-6/+6
7c4c97c0ac turned the flags in struct dir_struct into a single bitfield variable, but forgot to update this document. Signed-off-by: Adam Spiers <git@adamspiers.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-10-15attr: more matching optimizations from .gitignoreLibravatar Nguyễn Thái Ngọc Duy5-32/+64
.gitattributes and .gitignore share the same pattern syntax but has separate matching implementation. Over the years, ignore's implementation accumulates more optimizations while attr's stays the same. This patch reuses the core matching functions that are also used by excluded_from_list. excluded_from_list and path_matches can't be merged due to differences in exclude and attr, for example: * "!pattern" syntax is forbidden in .gitattributes. As an attribute can be unset (i.e. set to a special value "false") or made back to unspecified (i.e. not even set to "false"), "!pattern attr" is unclear which one it means. * we support attaching attributes to directories, but git-core internally does not currently make use of attributes on directories. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-10-15gitignore: make pattern parsing code a separate functionLibravatar Nguyễn Thái Ngọc Duy2-23/+50
This function can later be reused by attr.c. Also turn to_exclude field into a flag. Signed-off-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-10-15exclude: split pathname matching code into a separate functionLibravatar Nguyễn Thái Ngọc Duy1-32/+53
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-10-15exclude: fix a bug in prefix compare optimizationLibravatar Nguyễn Thái Ngọc Duy2-1/+7
When "namelen" becomes zero at this stage, we have matched the fixed part, but whether it actually matches the pattern still depends on the pattern in "exclude". As demonstrated in t3001, path "three/a.3" exists and it matches the "three/a.3" part in pattern "three/a.3[abc]", but that does not mean a true match. Don't be too optimistic and let fnmatch() do the job. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-10-15exclude: split basename matching code into a separate functionLibravatar Nguyễn Thái Ngọc Duy1-13/+24
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-10-15exclude: stricten a length check in EXC_FLAG_ENDSWITH caseLibravatar Nguyễn Thái Ngọc Duy1-2/+3
This block of code deals with the "basename" part only, which has the length of "pathlen - (basename - pathname)". Stricten the length check and remove "pathname" from the main expression to avoid confusion. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-10-05Merge commit 'f9f6e2c' into nd/attr-match-optim-moreLibravatar Junio C Hamano2-67/+78
* commit 'f9f6e2c': exclude: do strcmp as much as possible before fnmatch dir.c: get rid of the wildcard symbol set in no_wildcard() Unindent excluded_from_list()
2012-10-05attr: avoid searching for basename on every matchLibravatar Nguyễn Thái Ngọc Duy1-6/+9
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-10-05attr: avoid strlen() on every matchLibravatar Nguyễn Thái Ngọc Duy1-2/+4
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-09-14Git 1.7.11.7Libravatar Junio C Hamano4-3/+50
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-09-14Merge branch 'jk/maint-quiet-is-synonym-to-s-in-log' into maint-1.7.11Libravatar Junio C Hamano2-1/+13
* jk/maint-quiet-is-synonym-to-s-in-log: log: fix --quiet synonym for -s
2012-09-14Merge branch 'jc/maint-ident-missing-human-name' into maint-1.7.11Libravatar Junio C Hamano2-3/+20
* jc/maint-ident-missing-human-name: split_ident_line(): make best effort when parsing author/committer line
2012-09-14Merge branch 'rj/test-regex' into maint-1.7.11Libravatar Junio C Hamano4-0/+27
* rj/test-regex: test-regex: Add a test to check for a bug in the regex routines
2012-09-14Merge branch 'da/gitk-reload-tag-contents' into maint-1.7.11Libravatar Junio C Hamano1-8/+9
* da/gitk-reload-tag-contents: gitk: Rename 'tagcontents' to 'cached_tagcontent' gitk: Teach "Reread references" to reload tags gitk: Avoid Meta1-F5
2012-09-14Merge branch 'jc/maint-checkout-fileglob-doc' into maint-1.7.11Libravatar Junio C Hamano2-1/+30
* jc/maint-checkout-fileglob-doc: gitcli: contrast wildcard given to shell and to git gitcli: formatting fix Document file-glob for "git checkout -- '*.c'"
2012-09-12Merge branch 'jc/apply-binary-p0' into maint-1.7.11Libravatar Junio C Hamano2-46/+76
"git apply -p0" did not parse pathnames on "diff --git" line correctly. This caused patches that had pathnames in no other places to be mistakenly rejected (most notably, binary patch that does not rename nor change mode). Textual patches, renames or mode changes have preimage and postimage pathnames in different places in a form that can be parsed unambiguously and did not suffer from this problem. * jc/apply-binary-p0: apply: compute patch->def_name correctly under -p0
2012-09-12Merge branch 'jc/dotdot-is-parent-directory' into maint-1.7.11Libravatar Junio C Hamano5-4/+56
"git log .." errored out saying it is both rev range and a path when there is no disambiguating "--" is on the command line. Update the command line parser to interpret ".." as a path in such a case. * jc/dotdot-is-parent-directory: specifying ranges: we did not mean to make ".." an empty set
2012-09-12Merge branch 'jc/maint-doc-checkout-b-always-takes-branch-name' into ↵Libravatar Junio C Hamano1-3/+3
maint-1.7.11 The synopsis said "checkout [-B branch]" to make it clear the branch name is a parameter to the option, but the heading for the option description was "-B::", not "-B branch::", making the documentation misleading. * jc/maint-doc-checkout-b-always-takes-branch-name: doc: "git checkout -b/-B/--orphan" always takes a branch name