We learned that our source files with ".pl" and ".py" extensions
are Perl and Python files respectively and changes to them are
better viewed as such with appropriate diff drivers.
* ab/perl-python-attrs:
.gitattributes: add a diff driver for Python
.gitattributes: use the "perl" differ for Perl
.gitattributes: add *.pl extension for Perl
The codepath around object-info API has been taught to take the
repository object (which in turn tells the API which object store
the objects are to be located).
* sb/oid-object-info:
cache.h: allow oid_object_info to handle arbitrary repositories
packfile: add repository argument to cache_or_unpack_entry
packfile: add repository argument to unpack_entry
packfile: add repository argument to read_object
packfile: add repository argument to packed_object_info
packfile: add repository argument to packed_to_object_type
packfile: add repository argument to retry_bad_packed_offset
cache.h: add repository argument to oid_object_info
cache.h: add repository argument to oid_object_info_extended
The transport protocol v2 is getting updated further.
* bw/server-options:
fetch: send server options when using protocol v2
ls-remote: send server options when using protocol v2
serve: introduce the server-option capability
"git gc" in a large repository takes a lot of time as it considers
to repack all objects into one pack by default. The command has
been taught to pretend as if the largest existing packfile is
marked with ".keep" so that it is left untouched while objects in
other packs and loose ones are repacked.
* nd/repack-keep-pack:
pack-objects: show some progress when counting kept objects
gc --auto: exclude base pack if not enough mem to "repack -ad"
gc: handle a corner case in gc.bigPackThreshold
gc: add gc.bigPackThreshold config
gc: add --keep-largest-pack option
repack: add --keep-pack option
t7700: have closing quote of a test at the beginning of line
The code has been taught to use the duplicated information stored
in the commit-graph file to learn the tree object name for a commit
to avoid opening and parsing the commit object when it makes sense
to do so.
* ds/lazy-load-trees:
coccinelle: avoid wrong transformation suggestions from commit.cocci
commit-graph: lazy-load trees for commits
treewide: replace maybe_tree with accessor methods
commit: create get_commit_tree() method
treewide: rename tree to maybe_tree
The code did not propagate the terminal width to subprocesses via
COLUMNS environment variable, which it now does. This caused
trouble to "git column" helper subprocess when "git tag --column=row"
tried to list the existing tags on a display with non-default width.
* nd/term-columns:
column: fix off-by-one default width
pager: set COLUMNS to term_columns()
* sb/object-store-replace:
get_main_ref_store: BUG() when outside a repository
object.c: clear replace map before freeing it
replace-object.c: remove the_repository from prepare_replace_object
object.c: free replace map in raw_object_store_clear
remote-curl: accept compressed responses with protocol v2
Configure curl to accept compressed responses when using protocol v2 by
setting `CURLOPT_ENCODING` to "", which indicates that curl should send
an "Accept-Encoding" header with all supported compression encodings.
Signed-off-by: Brandon Williams <bmwill@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
remote-curl: accept all encodings supported by curl
Configure curl to accept all encodings which curl supports instead of
only accepting gzip responses.
This fixes an issue when using an installation of curl which is built
without the "zlib" feature. Since aa90b9697 (Enable info/refs gzip
decompression in HTTP client, 2012-09-19) we end up requesting "gzip"
encoding anyway despite libcurl not being able to decode it. Worse,
instead of getting a clear error message indicating so, we end up
falling back to "dumb" http, producing a confusing and difficult to
debug result.
Since curl doesn't do any checking to verify that it supports the a
requested encoding, instead set the curl option `CURLOPT_ENCODING` with
an empty string indicating that curl should send an "Accept-Encoding"
header containing only the encodings supported by curl.
Reported-by: Anton Golubev <anton.golubev@gmail.com> Signed-off-by: Brandon Williams <bmwill@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* maint-2.16:
Git 2.16.4
Git 2.15.2
Git 2.14.4
Git 2.13.7
verify_path: disallow symlinks in .gitmodules
update-index: stat updated files earlier
verify_dotfile: mention case-insensitivity in comment
verify_path: drop clever fallthrough
skip_prefix: add case-insensitive variant
is_{hfs,ntfs}_dotgitmodules: add tests
is_ntfs_dotgit: match other .git files
is_hfs_dotgit: match other .git files
is_ntfs_dotgit: use a size_t for traversing string
submodule-config: verify submodule names as paths
* maint-2.15:
Git 2.15.2
Git 2.14.4
Git 2.13.7
verify_path: disallow symlinks in .gitmodules
update-index: stat updated files earlier
verify_dotfile: mention case-insensitivity in comment
verify_path: drop clever fallthrough
skip_prefix: add case-insensitive variant
is_{hfs,ntfs}_dotgitmodules: add tests
is_ntfs_dotgit: match other .git files
is_hfs_dotgit: match other .git files
is_ntfs_dotgit: use a size_t for traversing string
submodule-config: verify submodule names as paths
* maint-2.14:
Git 2.14.4
Git 2.13.7
verify_path: disallow symlinks in .gitmodules
update-index: stat updated files earlier
verify_dotfile: mention case-insensitivity in comment
verify_path: drop clever fallthrough
skip_prefix: add case-insensitive variant
is_{hfs,ntfs}_dotgitmodules: add tests
is_ntfs_dotgit: match other .git files
is_hfs_dotgit: match other .git files
is_ntfs_dotgit: use a size_t for traversing string
submodule-config: verify submodule names as paths
* maint-2.13:
Git 2.13.7
verify_path: disallow symlinks in .gitmodules
update-index: stat updated files earlier
verify_dotfile: mention case-insensitivity in comment
verify_path: drop clever fallthrough
skip_prefix: add case-insensitive variant
is_{hfs,ntfs}_dotgitmodules: add tests
is_ntfs_dotgit: match other .git files
is_hfs_dotgit: match other .git files
is_ntfs_dotgit: use a size_t for traversing string
submodule-config: verify submodule names as paths
Merge branch 'jk/submodule-fix-loose' into maint-2.13
* jk/submodule-fix-loose:
verify_path: disallow symlinks in .gitmodules
update-index: stat updated files earlier
verify_dotfile: mention case-insensitivity in comment
verify_path: drop clever fallthrough
skip_prefix: add case-insensitive variant
is_{hfs,ntfs}_dotgitmodules: add tests
is_ntfs_dotgit: match other .git files
is_hfs_dotgit: match other .git files
is_ntfs_dotgit: use a size_t for traversing string
submodule-config: verify submodule names as paths
We've recently forbidden .gitmodules to be a symlink in
verify_path(). And it's an easy way to circumvent our fsck
checks for .gitmodules content. So let's complain when we
see it.
Now that the internal fsck code has all of the plumbing we
need, we can start checking incoming .gitmodules files.
Naively, it seems like we would just need to add a call to
fsck_finish() after we've processed all of the objects. And
that would be enough to cover the initial test included
here. But there are two extra bits:
1. We currently don't bother calling fsck_object() at all
for blobs, since it has traditionally been a noop. We'd
actually catch these blobs in fsck_finish() at the end,
but it's more efficient to check them when we already
have the object loaded in memory.
2. The second pass done by fsck_finish() needs to access
the objects, but we're actually indexing the pack in
this process. In theory we could give the fsck code a
special callback for accessing the in-pack data, but
it's actually quite tricky:
a. We don't have an internal efficient index mapping
oids to packfile offsets. We only generate it on
the fly as part of writing out the .idx file.
b. We'd still have to reconstruct deltas, which means
we'd basically have to replicate all of the
reading logic in packfile.c.
Instead, let's avoid running fsck_finish() until after
we've written out the .idx file, and then just add it
to our internal packed_git list.
This does mean that the objects are "in the repository"
before we finish our fsck checks. But unpack-objects
already exhibits this same behavior, and it's an
acceptable tradeoff here for the same reason: the
quarantine mechanism means that pushes will be
fully protected.
In addition to a basic push test in t7415, we add a sneaky
pack that reverses the usual object order in the pack,
requiring that index-pack access the tree and blob during
the "finish" step.
This already works for unpack-objects (since it will have
written out loose objects), but we'll check it with this
sneaky pack for good measure.
unpack-objects: call fsck_finish() after fscking objects
As with the previous commit, we must call fsck's "finish"
function in order to catch any queued objects for
.gitmodules checks.
This second pass will be able to access any incoming
objects, because we will have exploded them to loose objects
by now.
This isn't quite ideal, because it means that bad objects
may have been written to the object database (and a
subsequent operation could then reference them, even if the
other side doesn't send the objects again). However, this is
sufficient when used with receive.fsckObjects, since those
loose objects will all be placed in a temporary quarantine
area that will get wiped if we find any problems.
Now that the internal fsck code is capable of checking
.gitmodules files, we just need to teach its callers to use
the "finish" function to check any queued objects.
With this, we can now catch the malicious case in t7415 with
git-fsck.
This patch detects and blocks submodule names which do not
match the policy set forth in submodule-config. These should
already be caught by the submodule code itself, but putting
the check here means that newer versions of Git can protect
older ones from malicious entries (e.g., a server with
receive.fsckObjects will block the objects, protecting
clients which fetch from it).
As a side effect, this means fsck will also complain about
.gitmodules files that cannot be parsed (or were larger than
core.bigFileThreshold).
fsck: handle promisor objects in .gitmodules check
If we have a tree that points to a .gitmodules blob but
don't have that blob, we can't check its contents. This
produces an fsck error when we encounter it.
But in the case of a promisor object, this absence is
expected, and we must not complain. Note that this can
technically circumvent our transfer.fsckObjects check.
Imagine a client fetches a tree, but not the matching
.gitmodules blob. An fsck of the incoming objects will show
that we don't have enough information. Later, we do fetch
the actual blob. But we have no idea that it's a .gitmodules
file.
The only ways to get around this would be to re-scan all of
the existing trees whenever new ones enter (which is
expensive), or to somehow persist the gitmodules_found set
between fsck runs (which is complicated).
In practice, it's probably OK to ignore the problem. Any
repository which has all of the objects (including the one
serving the promisor packs) can perform the checks. Since
promisor packs are inherently about a hierarchical topology
in which clients rely on upstream repositories, those
upstream repositories can protect all of their downstream
clients from broken objects.
In preparation for performing fsck checks on .gitmodules
files, this commit plumbs in the actual detection of the
files. Note that unlike most other fsck checks, this cannot
be a property of a single object: we must know that the
object is found at a ".gitmodules" path at the root tree of
a commit.
Since the fsck code only sees one object at a time, we have
to mark the related objects to fit the puzzle together. When
we see a commit we mark its tree as a root tree, and when
we see a root tree with a .gitmodules file, we mark the
corresponding blob to be checked.
In an ideal world, we'd check the objects in topological
order: commits followed by trees followed by blobs. In that
case we can avoid ever loading an object twice, since all
markings would be complete by the time we get to the marked
objects. And indeed, if we are checking a single packfile,
this is the order in which Git will generally write the
objects. But we can't count on that:
1. git-fsck may show us the objects in arbitrary order
(loose objects are fed in sha1 order, but we may also
have multiple packs, and we process each pack fully in
sequence).
2. The type ordering is just what git-pack-objects happens
to write now. The pack format does not require a
specific order, and it's possible that future versions
of Git (or a custom version trying to fool official
Git's fsck checks!) may order it differently.
3. We may not even be fscking all of the relevant objects
at once. Consider pushing with transfer.fsckObjects,
where one push adds a blob at path "foo", and then a
second push adds the same blob at path ".gitmodules".
The blob is not part of the second push at all, but we
need to mark and check it.
So in the general case, we need to make up to three passes
over the objects: once to make sure we've seen all commits,
then once to cover any trees we might have missed, and then
a final pass to cover any .gitmodules blobs we found in the
second pass.
We can simplify things a bit by loosening the requirement
that we find .gitmodules only at root trees. Technically
a file like "subdir/.gitmodules" is not parsed by Git, but
it's not unreasonable for us to declare that Git is aware of
all ".gitmodules" files and make them eligible for checking.
That lets us drop the root-tree requirement, which
eliminates one pass entirely. And it makes our worst case
much better: instead of potentially queueing every root tree
to be re-examined, the worst case is that we queue each
unique .gitmodules blob for a second look.
This patch just adds the boilerplate to find .gitmodules
files. The actual content checks will come in a subsequent
commit.
Because fscking a blob has always been a noop, we didn't
bother passing around the blob data. In preparation for
content-level checks, let's fix up a few things:
1. The fsck_object() function just returns success for any
blob. Let's a noop fsck_blob(), which we can fill in
with actual logic later.
2. The fsck_loose() function in builtin/fsck.c
just threw away blob content after loading it. Let's
hold onto it until after we've called fsck_object().
The easiest way to do this is to just drop the
parse_loose_object() helper entirely. Incidentally,
this also fixes a memory leak: if we successfully
loaded the object data but did not parse it, we would
have left the function without freeing it.
3. When fsck_loose() loads the object data, it
does so with a custom read_loose_object() helper. This
function streams any blobs, regardless of size, under
the assumption that we're only checking the sha1.
Instead, let's actually load blobs smaller than
big_file_threshold, as the normal object-reading
code-paths would do. This lets us fsck small files, and
a NULL return is an indication that the blob was so big
that it needed to be streamed, and we can pass that
information along to fsck_blob().
There's no need for us to manually check for ".git"; it's a
subset of the other filesystem-specific tests. Dropping it
makes our code slightly shorter. More importantly, the
existing code may make a reader wonder why ".GIT" is not
covered here, and whether that is a bug (it isn't, as it's
also covered in the filesystem-specific tests).
If fsck reports an error, we say only "Error in object".
This isn't quite as bad as it might seem, since the fsck
code would have dumped some errors to stderr already. But it
might help to give a little more context. The earlier output
would not have even mentioned "fsck", and that may be a clue
that the "fsck.*" or "*.fsckObjects" config may be relevant.
Merge branch 'jk/submodule-name-verify-fix' into jk/submodule-name-verify-fsck
* jk/submodule-name-verify-fix:
verify_path: disallow symlinks in .gitmodules
update-index: stat updated files earlier
verify_path: drop clever fallthrough
skip_prefix: add icase-insensitive variant
is_{hfs,ntfs}_dotgitmodules: add tests
path: match NTFS short names for more .git files
is_hfs_dotgit: match other .git files
is_ntfs_dotgit: use a size_t for traversing string
submodule-config: verify submodule names as paths
Note that this includes two bits of evil-merge:
- there's a new call to verify_path() that doesn't actually
have a mode available. It should be OK to pass "0" here,
since we're just manipulating the untracked cache, not an
actual index entry.
- the lstat() in builtin/update-index.c:update_one() needs
to be updated to handle the fsmonitor case (without this
it still behaves correctly, but does an unnecessary
lstat).
There are a few reasons it's not a good idea to make
.gitmodules a symlink, including:
1. It won't be portable to systems without symlinks.
2. It may behave inconsistently, since Git may look at
this file in the index or a tree without bothering to
resolve any symbolic links. We don't do this _yet_, but
the config infrastructure is there and it's planned for
the future.
With some clever code, we could make (2) work. And some
people may not care about (1) if they only work on one
platform. But there are a few security reasons to simply
disallow it:
a. A symlinked .gitmodules file may circumvent any fsck
checks of the content.
b. Git may read and write from the on-disk file without
sanity checking the symlink target. So for example, if
you link ".gitmodules" to "../oops" and run "git
submodule add", we'll write to the file "oops" outside
the repository.
Again, both of those are problems that _could_ be solved
with sufficient code, but given the complications in (1) and
(2), we're better off just outlawing it explicitly.
Note the slightly tricky call to verify_path() in
update-index's update_one(). There we may not have a mode if
we're not updating from the filesystem (e.g., we might just
be removing the file). Passing "0" as the mode there works
fine; since it's not a symlink, we'll just skip the extra
checks.
In the update_one(), we check verify_path() on the proposed
path before doing anything else. In preparation for having
verify_path() look at the file mode, let's stat the file
earlier, so we can check the mode accurately.
This is made a bit trickier by the fact that this function
only does an lstat in a few code paths (the ones that flow
down through process_path()). So we can speculatively do the
lstat() here and pass the results down, and just use a dummy
mode for cases where we won't actually be updating the index
from the filesystem.
We check ".git" and ".." in the same switch statement, and
fall through the cases to share the end-of-component check.
While this saves us a line or two, it makes modifying the
function much harder. Let's just write it out.
We have the convenient skip_prefix() helper, but if you want
to do case-insensitive matching, you're stuck doing it by
hand. We could add an extra parameter to the function to
let callers ask for this, but the function is small and
somewhat performance-critical. Let's just re-implement it
for the case-insensitive version.
When we started to catch NTFS short names that clash with .git, we only
looked for GIT~1. This is sufficient because we only ever clone into an
empty directory, so .git is guaranteed to be the first subdirectory or
file in that directory.
However, even with a fresh clone, .gitmodules is *not* necessarily the
first file to be written that would want the NTFS short name GITMOD~1: a
malicious repository can add .gitmodul0000 and friends, which sorts
before `.gitmodules` and is therefore checked out *first*. For that
reason, we have to test not only for ~1 short names, but for others,
too.
It's hard to just adapt the existing checks in is_ntfs_dotgit(): since
Windows 2000 (i.e., in all Windows versions still supported by Git),
NTFS short names are only generated in the <prefix>~<number> form up to
number 4. After that, a *different* prefix is used, calculated from the
long file name using an undocumented, but stable algorithm.
For example, the short name of .gitmodules would be GITMOD~1, but if it
is taken, and all of ~2, ~3 and ~4 are taken, too, the short name
GI7EBA~1 will be used. From there, collisions are handled by
incrementing the number, shortening the prefix as needed (until ~9999999
is reached, in which case NTFS will not allow the file to be created).
We'd also want to handle .gitignore and .gitattributes, which suffer
from a similar problem, using the fall-back short names GI250A~1 and
GI7D29~1, respectively.
To accommodate for that, we could reimplement the hashing algorithm, but
it is just safer and simpler to provide the known prefixes. This
algorithm has been reverse-engineered and described at
https://usn.pw/blog/gen/2015/06/09/filenames/, which is defunct but
still available via https://web.archive.org/.
These can be recomputed by running the following Perl script:
Both verify_path() and fsck match ".git", ".GIT", and other
variants specific to HFS+. Let's allow matching other
special files like ".gitmodules", which we'll later use to
enforce extra restrictions via verify_path() and fsck.
is_ntfs_dotgit: use a size_t for traversing string
We walk through the "name" string using an int, which can
wrap to a negative value and cause us to read random memory
before our array (e.g., by creating a tree with a name >2GB,
since "int" is still 32 bits even on most 64-bit platforms).
Worse, this is easy to trigger during the fsck_tree() check,
which is supposed to be protecting us from malicious
garbage.
Note one bit of trickiness in the existing code: we
sometimes assign -1 to "len" at the end of the loop, and
then rely on the "len++" in the for-loop's increment to take
it back to 0. This is still legal with a size_t, since
assigning -1 will turn into SIZE_MAX, which then wraps
around to 0 on increment.
Submodule "names" come from the untrusted .gitmodules file,
but we blindly append them to $GIT_DIR/modules to create our
on-disk repo paths. This means you can do bad things by
putting "../" into the name (among other things).
Let's sanity-check these names to avoid building a path that
can be exploited. There are two main decisions:
1. What should the allowed syntax be?
It's tempting to reuse verify_path(), since submodule
names typically come from in-repo paths. But there are
two reasons not to:
a. It's technically more strict than what we need, as
we really care only about breaking out of the
$GIT_DIR/modules/ hierarchy. E.g., having a
submodule named "foo/.git" isn't actually
dangerous, and it's possible that somebody has
manually given such a funny name.
b. Since we'll eventually use this checking logic in
fsck to prevent downstream repositories, it should
be consistent across platforms. Because
verify_path() relies on is_dir_sep(), it wouldn't
block "foo\..\bar" on a non-Windows machine.
2. Where should we enforce it? These days most of the
.gitmodules reads go through submodule-config.c, so
I've put it there in the reading step. That should
cover all of the C code.
We also construct the name for "git submodule add"
inside the git-submodule.sh script. This is probably
not a big deal for security since the name is coming
from the user anyway, but it would be polite to remind
them if the name they pick is invalid (and we need to
expose the name-checker to the shell anyway for our
test scripts).
This patch issues a warning when reading .gitmodules
and just ignores the related config entry completely.
This will generally end up producing a sensible error,
as it works the same as a .gitmodules file which is
missing a submodule entry (so "submodule update" will
barf, but "git clone --recurse-submodules" will print
an error but not abort the clone.
There is one minor oddity, which is that we print the
warning once per malformed config key (since that's how
the config subsystem gives us the entries). So in the
new test, for example, the user would see three
warnings. That's OK, since the intent is that this case
should never come up outside of malicious repositories
(and then it might even benefit the user to see the
message multiple times).
Credit for finding this vulnerability and the proof of
concept from which the test script was adapted goes to
Etienne Stalmans.
The '--progress' was introduced in 72c5f88311d (clone: pass --progress
decision to recursive submodules, 2016-09-22) to fix the progress reporting
of the clone command. Also add the progress option to the 'submodule add'
command. The update command already supports the progress flag, but it
is not documented.
Signed-off-by: Casey Fitzpatrick <kcghost@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
'recommend_shallow' and 'jobs' variables do not need quotes. They only hold a
single token value, and even if they were multi-token it is likely we would want
them split at IFS rather than pass a single string.
'progress' is a boolean value. Treat it like the other boolean values in the
script by using a substitution.
Signed-off-by: Casey Fitzpatrick <kcghost@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The strings allocated in `setup_unpack_trees_porcelain()` are never
freed. Provide a function `clear_unpack_trees_porcelain()` to do so and
call it where we use `setup_unpack_trees_porcelain()`. The only
non-trivial user is `unpack_trees_start()`, where we should place the
new call in `unpack_trees_finish()`.
We keep the string pointers in an array, mixing pointers to static
memory and memory that we allocate on the heap. We also keep several
copies of the individual pointers. So we need to make sure that we do
not free what we must not free and that we do not double-free. Let a
separate argv_array take ownership of all the strings we create so that
we can easily free them.
Zero the whole array of string pointers to make sure that we do not
leave any dangling pointers.
Note that we only take responsibility for the memory allocated in
`setup_unpack_trees_porcelain()` and not any other members of the
`struct unpack_trees_options`.
Helped-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
if (some condition) {
msg = "constant string message";
... other logic ...
} else {
msg = argv_array_pushf(&to_free, "format %s", var);
}
... use "msg" ...
... do other things ...
argv_array_clear(&to_free);
Note that argv_array_pushl() and argv_array_pushv() are used to push
one or more strings with a single call, so we do not return any one
of these strings from these two functions in order to reduce the
chance to misuse the API.
Signed-off-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Switch from gcc-4.8 to gcc-8. Newer compilers come with more warning
checks (usually in -Wextra). Since -Wextra is enabled in developer
mode (which is also enabled in travis), this lets travis report more
warnings before other people do it.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
regex: do not call `regfree()` if compilation fails
It is apparently undefined behavior to call `regfree()` on a regex where
`regcomp()` failed. The language in [1] is a bit muddy, at least to me,
but the clearest hint is this (`preg` is the `regex_t *`):
Upon successful completion, the regcomp() function shall return 0.
Otherwise, it shall return an integer value indicating an error as
described in <regex.h>, and the content of preg is undefined.
Funnily enough, there is also the `regerror()` function which should be
given a pointer to such a "failed" `regex_t` -- the content of which
would supposedly be undefined -- and which may investigate it to come up
with a detailed error message.
In any case, the example in that document shows how `regfree()` is not
called after `regcomp()` fails.
We have quite a few users of this API and most get this right. These
three users do not.
Several implementations can handle this just fine [2] and these code paths
supposedly have not wreaked havoc or we'd have heard about it. (These
are all in code paths where git got bad input and is just about to die
anyway.) But let's just avoid the issue altogether.
config: let `config_store_data_clear()` handle `key`
Instead of remembering to free `key` in each code path, let
`config_store_data_clear()` handle that.
We still need to free it before replacing it, though. Move that freeing
closer to the replacing to be safe. Note that in that same part of the
code, we can no longer set `key` to the original pointer, but need to
`xstrdup()` it.
Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
config: let `config_store_data_clear()` handle `value_regex`
Instead of duplicating the logic for clearing up `value_regex`, let
`config_store_data_clear()` handle that.
When `regcomp()` fails, the current code does not call `regfree()`. Make
sure we do the same by immediately invalidating `value_regex`. Some
implementations are able to handle such an extra `regfree()`-call [1],
but from the example in [2], we should not do so. (The language itself
in [2] is not super-clear on this.)
Researched-by: Eric Sunshine <sunshine@sunshineco.com> Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
config: free resources of `struct config_store_data`
Commit fee8572c6d (config: avoid using the global variable `store`,
2018-04-09) dropped the staticness of a certain struct, instead letting
the users create an instance on the stack and pass around a pointer.
We do not free all the memory that the struct tracks. When the struct
was static, the memory would always be reachable. Now that we keep the
struct on the stack, though, as soon as we return, it goes out of scope
and we leak the memory it points to. In particular, we leak the memory
pointed to by the `parsed` and `seen` fields.
Introduce and use a helper function `config_store_data_clear()` to plug
these leaks. The memory tracked here is config parser events. Once the
users (`git_config_set_multivar_in_file_gently()` and
`git_config_copy_or_rename_section_in_file()` at the moment) are done,
no-one should be holding on to a pointer into this memory.
There are two more members of the struct that are candidates for freeing
in this new function (`key` and `value_regex`). Those are actually
already being taken care of. The next couple of patches will move their
freeing into the function we are adding here.
Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
This test enumerates log entries and then sorts them. For SHA-1, this
produces results that happen to sort in the order specified in the test,
but for other hash algorithms they sort differently. Ensure we sort the
log entries in a hash-independent way by sorting on the ref name instead
of the object ID.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Adjust the test code so that it computes variables for blobs instead of
using hard-coded hashes. This makes t4033 and t4050 (the patience and
histogram tests) pass.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
We typically indent our tests with a single tab, partially so that we
can take advantage of indented heredocs. Make this change and move the
quote marks to be in the typical position for our tests.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Strip out the index lines in the diff before comparing them, as these
will differ between hash algorithms. This leads to a smaller, simpler
change than editing the index line.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
completion: allow to customize the completable command list
By default we show porcelain, external commands and a couple others
that are also popular. If you are not happy with this list, you can
now customize it a new config variable.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
By providing aliases via --list-cmds=, we could simplify command
collection code in the script. We only issue one git command. Before
this patch that is "git config", after it's "git --list-cmds=". In
"git help" completion case we actually reduce one "git" process (for
getting guides) but that call was added in this series so it does not
really count.
A couple of bash functions are removed because they are not needed
anymore. __git_compute_all_commands() and $__git_all_commands stay
because they are still needed for completing pager.* config and
without "alias" group, the result is still cacheable.
There is a slight (good) change in _git_help() with this patch: before
"git help <tab>" shows external commands (as in _not_ part of git) as
well as part of $__git_all_commands. We have finer control over
command listing now and can exclude that because we can't provide a
man page for external commands anyway.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The following commands are removed from the complete list:
- annotate obsolete, discouraged to use
- filter-branch not often used
- get-tar-commit-id not often used
- imap-send not often used
- interpreter-trailers not for interactive use
- name-rev plumbing, just use git-describe
- p4 too short and probably not often used (*)
- svn same category as p4 (*)
- verify-commit not often used
(*) to be fair, send-email command which is in the same foreignscminterface
group as svn and p4 does get completion, just because it's used by git
and kernel development. So maybe we should include them.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
completion: let git provide the completable command list
Instead of maintaining a separate list of command classification,
which often could go out of date, let's centralize the information
back in git.
While the function in git-completion.bash implies "list porcelain
commands", that's not exactly what it does. It gets all commands (aka
--list-cmds=main,others) then exclude certain non-porcelain ones. We
could almost recreate this list two lists list-mainporcelain and
others. The non-porcelain-but-included-anyway is added by the third
category list-complete.
Note that the current completion script incorrectly classifies
filter-branch as porcelain and t9902 tests this behavior. We keep it
this way in t9902 because this test does not really care which
particular command is porcelain or plumbing, they're just names.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
help: use command-list.txt for the source of guides
The help command currently hard codes the list of guides and their
summary in C. Let's move this list to command-list.txt. This lets us
extract summary lines from Documentation/git*.txt. This also
potentially lets us list guides in git.txt, but I'll leave that for
now.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
help: add "-a --verbose" to list all commands with synopsis
This lists all recognized commands [1] by category. The group order
follows closely git.txt.
[1] We may actually show commands that are not built (e.g. if you set
NO_PERL you don't have git-instaweb but it's still listed here). I
ignore the problem because on Linux a git package could be split
anyway. The "git-core" package may not contain git-instaweb even if
it's built because it may end up in a separate package. We can't know
anyway.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
This allows us to select any group of commands by a category defined
in command-list.txt. This is an internal/hidden option so we don't
have to be picky about the category name or worried about exposing too
much.
This will be used later by git-completion.bash to retrieve certain
command groups.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
completion: implement and use --list-cmds=main,others
This is part of the effort to break down and provide commands by
category in machine-readable form. This could be helpful later on when
completion script switches to use --list-cmds for selecting
completable commands. It would be much easier for the user to choose
to complete _all_ commands instead of the default selection by passing
different values to --list-cmds in git-completino.bash.
While at there, replace "git help -a" in git-completion.bash with
--list-cmds since it's better suited for this task.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
git --list-cmds: collect command list in a string_list
Instead of printing the command directly one by one, keep them in a
list and print at the end. This allows more modification before we
print out (e.g. sorting, removing duplicates or even excluding some
items).
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Even if these are hidden options, let's make them a bit more generic
since we're introducing more listing types shortly. The code is
structured to allow combining multiple listing types together because
we will soon add more types the 'builtins'.
'parseopt' remains separate because it has separate (SPC) to match
git-completion.bash needs and will not combine with others.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Functions like die_errno() use fmt_with_err() to combine the
caller-provided format with the strerror() string. We use a
fixed stack buffer because we're already handling an error
and don't have any way to report another one. Our buffer
should generally be big enough to fit this, but if it's not,
truncation is our best option. Let's add a comment to that
effect, so that anybody auditing the code for truncation
bugs knows that this is fine.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
We convert the ref_rev_parse_rules array into scanf formats
on the fly, and use snprintf() to write into each string. We
should have enough memory to hold everything because of the
earlier total_len computation. Let's use xsnprintf() to
give runtime confirmation that this is the case, and to make
it easy for people auditing the code to know there's no
truncation bug.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
fsmonitor: use internal argv_array of struct child_process
Avoid magic array sizes and indexes by constructing the fsmonitor
command line using the embedded argv_array of the child_process. The
resulting code is shorter and easier to extend.
Getting rid of the snprintf() calls is a bonus -- even though the
buffers were big enough here to avoid truncation -- as it makes auditing
the remaining callers easier.
Inspired-by: Jeff King <peff@peff.net> Signed-off-by: Rene Scharfe <l.s.r@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we write a MIME attachment, we write the mime headers
into fixed-size buffers. These are likely to be big enough
in practice, but technically the input could be arbitrarily
large (e.g., if the caller provided a lot of content in the
extra_headers string), in which case we'd quietly truncate
it and generate bogus output. Let's convert these buffers to
strbufs.
The memory ownership here is a bit funny. The original fixed
buffers were static, and we merely pass out pointers to them
to be used by the caller (and in one case, we even just
stuff our value into the opt->diffopt.stat_sep value).
Ideally we'd actually pass back heap buffers, and the caller
would be responsible for freeing them.
This patch punts on that cleanup for now, and instead just
marks the strbufs as static. That means we keep ownership in
this function, making it not a complete leak. This also
takes us one step closer to fixing it in the long term
(since we can eventually use strbuf_detach() to hand
ownership to the caller, once it's ready).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
We keep the names of incoming packs and objects in fixed
PATH_MAX-size buffers, and snprintf() into them. This is
unlikely to end up with truncated filenames, but it is
possible (especially on systems where PATH_MAX is shorter
than actual paths can be). Let's switch to using strbufs,
which makes the question go away entirely.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
config: die when --blob is used outside a repository
If you run "config --blob" outside of a repository, then we
eventually try to resolve the blob name and hit a BUG().
Let's catch this earlier and provide a useful message.
Note that we could also catch this much lower in the stack,
in git_config_from_blob_ref(). That might cover other
callsites, too, but it's unclear whether those ones would
actually be bugs or not. So let's leave the low-level
functions to assume the caller knows what it's doing (and
BUG() if it turns out it doesn't).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
get_main_ref_store: BUG() when outside a repository
If we don't have a repository, then we can't initialize the
ref store. Prior to 64a741619d (refs: store the main ref
store inside the repository struct, 2018-04-11), we'd try to
access get_git_dir(), and outside a repository that would
trigger a BUG(). After that commit, though, we directly use
the_repository->git_dir; if it's NULL we'll just segfault.
Let's catch this case and restore the BUG() behavior.
Obviously we don't ever want to hit this code, but a BUG()
is a lot more helpful than a segfault if we do.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The tests added in 2f271cd9cf (t9902-completion: add tests
demonstrating issues with quoted pathnames, 2018-05-08) and in 2ab6eab4fe (completion: improve handling quoted paths in 'git
ls-files's output, 2018-03-28) have a few shortcomings:
- All these tests use the 'test_completion' helper function, thus
they are exercising the whole completion machinery, although they
are only interested in how git-aware path completion, specifically
the __git_complete_index_file() function deals with unusual
characters in pathnames and on the command line.
- These tests can't satisfactorily test the case of pathnames
containing spaces, because 'test_completion' gets the words on the
command line as a single argument and it uses space as word
separator.
- Some of the tests are protected by different FUNNYNAMES_* prereqs
depending on whether they put backslashes and double quotes or
separator characters (FS, GS, RS, US) in pathnames, although a
filesystem not allowing one likely doesn't allow the others
either.
- One of the tests operates on paths containing '|' and '&'
characters without being protected by a FUNNYNAMES prereq, but
some filesystems (notably on Windows) don't allow these characters
in pathnames, either.
Replace these tests with basically equivalent, more focused tests that
call __git_complete_index_file() directly. Since this function only
looks at the current word to be completed, i.e. the $cur variable, we
can easily include pathnames containing spaces in the tests, so use
such pathnames instead of pathnames containing '|' and '&'. Finally,
use only a single FUNNYNAMES prereq for all kinds of special
characters.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
completion: don't return with error from __gitcomp_file_direct()
In __gitcomp_file_direct() we tell Bash that it should handle our
possible completion words as filenames with the following piece of
cleverness:
# use a hack to enable file mode in bash < 4
compopt -o filenames +o nospace 2>/dev/null ||
compgen -f /non-existing-dir/ > /dev/null
Unfortunately, this makes this function always return with error
when it is not invoked in real completion, but e.g. in tests of
't9902-completion.sh':
- First the 'compopt' line errors out
- either because in Bash v3.x there is no such command,
- or because in Bash v4.x it complains about "not currently
executing completion function",
- then 'compgen' just silently returns with error because of the
non-existing directory.
Since __gitcomp_file_direct() is now the last command executed in
__git_complete_index_file(), that function returns with error as well,
which prevents it from being invoked in tests directly as is, and
would require extra steps in test to hide its error code.
So let's make sure that __gitcomp_file_direct() doesn't return with
error, because in the tests coming in the following patch we do want
to exercise __git_complete_index_file() directly,
__gitcomp_file() contains the same construct, and thus it, too, always
returns with error. Update that function accordingly as well.
While at it, also remove the space from between the redirection
operator and the filename in both functions.
Signed-off-by: SZEDER Gábor <szeder.dev@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>