John Keeping [Sat, 6 Apr 2013 09:28:57 +0000 (10:28 +0100)]
use struct strbuf instead of static buffers
Use "struct strbuf" from Git to remove the limit on file path length.
Notes on scan-tree:
This is slightly involved since I decided to pass the strbuf into
add_repo() and modify if whenever a new file name is required, which
should avoid any extra allocations within that function. The pattern
there is to append the filename, use it and then reset the buffer to its
original length (retaining a trailing '/').
Notes on ui-snapshot:
Since write_archive modifies the argv array passed to it we
copy the argv_array values into a new array of char* and then free the
original argv_array structure and the new array without worrying about
what the values now look like.
John Keeping [Sat, 6 Apr 2013 09:49:22 +0000 (10:49 +0100)]
Remove redundant calls to fmt("%s", ...)
After this change there is one remaining call 'fmt("%s", delim)' in
ui-shared.c but is needed as delim is stack allocated and so cannot be
returned from the function.
John Keeping [Sat, 6 Apr 2013 10:23:52 +0000 (11:23 +0100)]
Convert cgit_print_error to a variadic function
This removes many uses of "fmt" which uses a fixed size static pool of
fixed size buffers. Instead of relying on these, we now pass around
argument lists for as long as possible before using a strbuf to render
content of an arbitrary size.
John Keeping [Sun, 7 Apr 2013 13:40:50 +0000 (14:40 +0100)]
html.c: add various strbuf and varadic helpers
This adds the fmtalloc helper, html_txtf, html_vtxtf, and html_attrf.
These takes a printf style format string like htmlf but escapes the
resulting string. The html_vtxtf variant takes a va_list whereas
html_txtf is variadic.
John Keeping [Mon, 1 Apr 2013 18:03:34 +0000 (19:03 +0100)]
Fix out-of-bounds memory accesses with virtual_root=""
The CGit configuration variable virtual_root is normalized so that it
does not have a trailing '/' character, but it is allowed to be empty
(the empty string and NULL have different meanings here) and there is
code that is insufficiently cautious when checking if it ends in a '/':
if (virtual_root[strlen(virtual_root) - 1] != '/')
Clearly this check is redundant, but rather than simply removing it we
get a slight efficiency improvement by switching the normalization so
that the virtual_root variable always ends in '/'. Do this with a new
"ensure_end" helper.
The code snippets for OBJ_TAG and other object types are almost
equivalent. Merge them and use a couple of inline if conditions to
select proper fields.
print_tag_header() is only called from cgit_print_tags() -- the
conditional invocation in print_tag() is never executed since
print_tag() is only called by cgit_print_tags() which already executes
print_tag_header() before (resulting in the global variable being always
set in when the condition is evaluated).
Remove the global variable and the conditional invocation.
John Keeping [Thu, 4 Apr 2013 17:32:24 +0000 (18:32 +0100)]
cgit.mk: fix dependency handling
Git calculates the dependency files to be included using a simply
expanded Makefile variable, so it does not include the CGit objects that
are added after that Makefile has been processed.
We therefore need to include the dependency files ourselves in order to
get the dependency calculations right. Do this.
Do not unnecessarily strdup() environment variables
This reverts the memory duplication introduced in commit 60a2627, while
keeping everything else that has been cleaned up. The environment
variables are never modified, so we do not need to call xstrdupn() here.
Also, remove xstrdupn() which is no longer needed.
These reflect the values of environment variables and should never be
changed. Add another xstrdup() when we assign environment variables to
strings that are potentially non-constant.
Return const char * in cgit_{httpscheme, hosturl, rooturl}()
The return values of these functions are essentially constant and should
never be modified.
Note that this will introduce a compiler warning when we try to free the
return value of any of these functions. However, given that all of these
currently return statically allocated strings in some cases, they need
to be refactored before this can be done anyway.
John Keeping [Mon, 8 Apr 2013 08:00:22 +0000 (09:00 +0100)]
Update git to v1.8.2.1
This requires a small change to how we handle notes, but otherwise just
works.
Note that we can't use anything from v1.8.0 until v1.8.2.1 because some
of the symbols that we need for graph drawing were made private in
v1.8.0 and this was not reverted until v1.8.2.1.
John Keeping [Sun, 7 Apr 2013 14:06:23 +0000 (15:06 +0100)]
ui-blob: don't segfault when no path is given
It it possible to inspect blobs by specifying only the SHA-1, and CGit
provides links to do so, for example if a tag points directly at a blob.
In this case the path_items structure is never used, but creating it
still causes strlen to be run on a null pointer. Fix this.
This error was introduced by commit c1633c6 (Update git to v1.7.6.5 -
2013-03-02).
Lukas Fleischer [Tue, 5 Mar 2013 15:48:27 +0000 (16:48 +0100)]
ui-summary.c: Move urls variable into print_urls()
There's no need for this variable to be global. Printing the header in
print_urls() instead of print_url() allows for moving this variable into
print_urls() without having to pass any status to print_url().
Note that this only works as long as we don't call print_urls() more
than once.
This squelches a gcc warning. It's also correct that we check to see if
there are any partial or failed writes. For now, we just print a warning
to stderr. In the future, perhaps it will prove wise to exit(1) on
partial writes.
Since tail is initialized to 0, we will never get a warning on the last
if statement, but recent gcc complains anyway. So, we initialize len as
well. Future gcc versions should be able to optimize this out anyway.
John Keeping [Wed, 6 Mar 2013 21:22:09 +0000 (21:22 +0000)]
cgit.mk: Use SHELL_PATH_SQ to run gen-version.sh
On some platforms (notably Solaris) /bin/sh doesn't support enough of
POSIX for gen-version.sh to run. Git's Makefile provides SHELL_PATH_SQ
to address this issue so we just have to use it.
John Keeping [Wed, 6 Mar 2013 21:22:08 +0000 (21:22 +0000)]
cgit.mk: don't rebuild everything if CGIT_VERSION changes
If CGIT_VERSION is in CGIT_CFLAGS then a change in version (for example
because you have committed your changes) causes all of the CGit objects
to be rebuilt. Avoid this by using EXTRA_CPPFLAGS to add the version
for only those files that are affected and make them depend on VERSION.
John Keeping [Wed, 6 Mar 2013 21:22:07 +0000 (21:22 +0000)]
ui-patch: use cgit_version not CGIT_VERSION
We already have a global cgit_version which is set from the #define'd
CGIT_VERSION in cgit.c. Change ui-patch.c to use this so that we only
need to rebuild cgit.o when the version changes.
John Keeping [Wed, 6 Mar 2013 21:22:06 +0000 (21:22 +0000)]
Makefile: re-use Git's Makefile where possible
Git does quite a lot of platform-specific detection in its Makefile,
which can result in it defining preprocessor variables that are used in
its header files. If CGit does not define the same variables it can
result in different sizes of some structures in different places in the
same application.
For example, on Solaris Git uses it's "compat" regex library which has a
different sized regex_t structure than that available in the platform
regex.h. This has a knock-on effect on the size of "struct rev_info"
and leads to hard to diagnose runtime issues.
In order to avoid all of this, introduce a "cgit.mk" file that includes
Git's Makefile and make all of the existing logic apply to CGit's
objects as well. This is slightly complicated because Git's Makefile
must run in Git's directory, so all references to CGit files need to be
prefixed with "../".
In addition, OBJECTS is a simply expanded variable in Git's Makefile so
we cannot just add our objects to it. Instead we must copy the two
applicable rules into "cgit.mk". This has the advantage that we can
split CGit-specific CFLAGS from Git's CFLAGS and hence avoid rebuilding
all of Git whenever a CGit-specific value changes.
Signed-off-by: John Keeping <john@keeping.me.uk> Acked-by: Jamie Couture <jamie.couture@gmail.com>
John Keeping [Wed, 6 Mar 2013 20:51:54 +0000 (20:51 +0000)]
ui-shared: fix return type of cgit_self_link
cgit_self_link() is a void function but implements each case it handles
by doing "return <another_void_function>" which is not valid C; section
6.8.6.4 of C11 says:
A return statement with an expression shall not appear in a
function whose return type is void.
Fix this by removing the return keywords and converting the final code
block into an "else".
Lukas Fleischer [Mon, 4 Mar 2013 12:25:38 +0000 (13:25 +0100)]
cgit_print_snapshot_links(): Free prefix variable
Fixes following memory leak seen with "PATH_INFO=/cgit/commit/":
==16894== 12 bytes in 1 blocks are definitely lost in loss record 9 of 92
==16894== at 0x4C2C04B: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==16894== by 0x56F2DF1: strdup (in /usr/lib/libc-2.17.so)
==16894== by 0x46CAC8: xstrdup (wrapper.c:35)
==16894== by 0x414E34: cgit_print_snapshot_links (ui-shared.c:926)
==16894== by 0x40CFA1: cgit_print_commit (ui-commit.c:102)
==16894== by 0x407B06: commit_fn (cmd.c:54)
==16894== by 0x405E16: process_request (cgit.c:574)
==16894== by 0x4074C8: cache_process (cache.c:322)
==16894== by 0x406C4F: main (cgit.c:872)
Lukas Fleischer [Mon, 4 Mar 2013 12:25:37 +0000 (13:25 +0100)]
cgit_print_commit(): Free tmp variable
Fixes following memory leak seen with "PATH_INFO=/cgit/commit/":
==16894== 7 bytes in 1 blocks are definitely lost in loss record 4 of 92
==16894== at 0x4C2C04B: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==16894== by 0x56F2DF1: strdup (in /usr/lib/libc-2.17.so)
==16894== by 0x46CAC8: xstrdup (wrapper.c:35)
==16894== by 0x40CD6F: cgit_print_commit (ui-commit.c:70)
==16894== by 0x407B06: commit_fn (cmd.c:54)
==16894== by 0x405E16: process_request (cgit.c:574)
==16894== by 0x4074C8: cache_process (cache.c:322)
==16894== by 0x406C4F: main (cgit.c:872)
Lukas Fleischer [Mon, 4 Mar 2013 12:25:36 +0000 (13:25 +0100)]
cgit_print_tree(): Free curr_rev after usage
Fixes following memory leak seen with "PATH_INFO=/cgit/tree/":
==15715== 7 bytes in 1 blocks are definitely lost in loss record 4 of 51
==15715== at 0x4C2C04B: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==15715== by 0x56F2DF1: strdup (in /usr/lib/libc-2.17.so)
==15715== by 0x46CAA8: xstrdup (wrapper.c:35)
==15715== by 0x418A4C: cgit_print_tree (ui-tree.c:274)
==15715== by 0x407D91: tree_fn (cmd.c:131)
==15715== by 0x405E16: process_request (cgit.c:574)
==15715== by 0x4074C8: cache_process (cache.c:322)
==15715== by 0x406C4F: main (cgit.c:872)
Lukas Fleischer [Mon, 4 Mar 2013 12:25:35 +0000 (13:25 +0100)]
find_default_branch(): Free refmatch after usage
Fixes following memory leak seen with "PATH_INFO=/cgit/refs/":
==13408== 7 bytes in 1 blocks are definitely lost in loss record 4 of 52
==13408== at 0x4C2C04B: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==13408== by 0x56F2DF1: strdup (in /usr/lib/libc-2.17.so)
==13408== by 0x46CA78: xstrdup (wrapper.c:35)
==13408== by 0x405840: find_current_ref (cgit.c:426)
==13408== by 0x44BE5A: do_one_ref (refs.c:527)
==13408== by 0x44D3E0: do_for_each_ref_in_dir (refs.c:553)
==13408== by 0x44D85A: do_for_each_ref (refs.c:1298)
==13408== by 0x405889: find_default_branch (cgit.c:438)
==13408== by 0x405AC4: prepare_repo_cmd (cgit.c:490)
==13408== by 0x405D97: process_request (cgit.c:557)
==13408== by 0x407490: cache_process (cache.c:322)
==13408== by 0x406C18: main (cgit.c:864)
Lukas Fleischer [Mon, 4 Mar 2013 12:25:34 +0000 (13:25 +0100)]
print_tag_downloads(): Free ref variable
Make sure the ref variable is freed if we build a
"$basename-$version"-style ref.
This fixes following memory leak seen with "PATH_INFO=/cgit/refs/":
==8784== 323 bytes in 29 blocks are definitely lost in loss record 41 of 53
==8784== at 0x4C2C04B: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==8784== by 0x56F2DF1: strdup (in /usr/lib/libc-2.17.so)
==8784== by 0x46CA28: xstrdup (wrapper.c:35)
==8784== by 0x410DA6: print_tag_downloads (ui-refs.c:115)
==8784== by 0x410F02: print_tag (ui-refs.c:141)
==8784== by 0x41128B: cgit_print_tags (ui-refs.c:230)
==8784== by 0x41134D: cgit_print_refs (ui-refs.c:250)
==8784== by 0x407C85: refs_fn (cmd.c:105)
==8784== by 0x405DDF: process_request (cgit.c:566)
==8784== by 0x407490: cache_process (cache.c:322)
==8784== by 0x406C18: main (cgit.c:864)
Lukas Fleischer [Mon, 4 Mar 2013 12:25:33 +0000 (13:25 +0100)]
Free reflists after usage
Free reflists in cgit_print_branches() and in cgit_print_tags() before
returning reflist structures to the stack.
This fixes following memory leaks seen with "PATH_INFO=/cgit/refs/":
==5710== 1,312 (32 direct, 1,280 indirect) bytes in 1 blocks are definitely lost in loss record 63 of 71
==5710== at 0x4C2C04B: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==5710== by 0x4C2C2FF: realloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==5710== by 0x46CA9B: xrealloc (wrapper.c:100)
==5710== by 0x40AAA6: cgit_add_ref (shared.c:156)
==5710== by 0x40ABC4: cgit_refs_cb (shared.c:186)
==5710== by 0x44BCBA: do_one_ref (refs.c:527)
==5710== by 0x44D240: do_for_each_ref_in_dir (refs.c:553)
==5710== by 0x44D6BA: do_for_each_ref (refs.c:1298)
==5710== by 0x410FE2: cgit_print_branches (ui-refs.c:191)
==5710== by 0x4111E9: cgit_print_refs (ui-refs.c:244)
==5710== by 0x407C85: refs_fn (cmd.c:105)
==5710== by 0x405DDF: process_request (cgit.c:566)
==5710==
==5710== 6,846 (256 direct, 6,590 indirect) bytes in 1 blocks are definitely lost in loss record 68 of 71
==5710== at 0x4C2C25E: realloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==5710== by 0x46CA9B: xrealloc (wrapper.c:100)
==5710== by 0x40AAA6: cgit_add_ref (shared.c:156)
==5710== by 0x40ABC4: cgit_refs_cb (shared.c:186)
==5710== by 0x44BCBA: do_one_ref (refs.c:527)
==5710== by 0x44D240: do_for_each_ref_in_dir (refs.c:553)
==5710== by 0x44D6EC: do_for_each_ref (refs.c:1288)
==5710== by 0x4110D5: cgit_print_tags (ui-refs.c:218)
==5710== by 0x4111FD: cgit_print_refs (ui-refs.c:246)
==5710== by 0x407C85: refs_fn (cmd.c:105)
==5710== by 0x405DDF: process_request (cgit.c:566)
==5710== by 0x407490: cache_process (cache.c:322)
John Keeping [Mon, 4 Mar 2013 21:58:49 +0000 (21:58 +0000)]
tests: "grep -e" is not portable to all platforms
The "-e" option to grep is not needed unless specifying more than one
pattern, which we don't do. Remove it to avoid restricting the tests on
platforms that do not have a grep that recognises "-e".
Jamie Couture [Sat, 17 Nov 2012 18:02:17 +0000 (13:02 -0500)]
Makefile: improve dependency generation
Makefile target generation would always be included for any makefile
target that was not clean. Only care to include the '.deps' directory
when building cgit, rather than generating and including dependencies
when calling other makefile targets.
Heavily borrowed from git's Makefile, but without definitions to test
for the compiler's header dependency feature. Previous Makefile
implementation never checked for this compiler feature anyway.
- Removed makecmdgoal 'clean' check
- Grouped like .PHONY target definitions
- Place build dependency targets under .SUFFIXES
- Re-arranged location of library inclusion definitions
- Use google code mirror instead of github
Signed-off-by: Jamie Couture <jamie.couture@gmail.com> Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Lukas Fleischer [Sun, 3 Mar 2013 15:49:47 +0000 (16:49 +0100)]
ui-tree.c: Declare the state variable globally
This allows for removing the header variable in a following patch. We
can use the state variable to check whether the tail needs to be printed
instead.
Note that the state variable will be moved into a context structure
later.
Lukas Fleischer [Sun, 3 Mar 2013 16:27:54 +0000 (17:27 +0100)]
ui-plain.c: Use a context structure in walk_tree()
Do not misuse global variables to save the context. Instead, use the
context pointer which was designed to share information between a
read_tree_fn and the caller.
Lukas Fleischer [Sun, 3 Mar 2013 16:27:46 +0000 (17:27 +0100)]
ui-blob.c: Use a context structure in walk_tree()
Do not misuse global variables to save the context. Instead, use the
context pointer which was designed to share information between a
read_tree_fn and the caller.
This also prevents from potential misuse of the global pointers
match_path and matched_sha1 after the referenced values have been
overwritten on the stack.
parse_archive_args does not pass PARSE_OPT_KEEP_ARGV0 to parse_args,
which means the first argument will be discarded, as though it were a
function being called from the command-line. Thus, we fill argv[0] with
a dummy argument to prevent this from happening.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Lukas Fleischer [Sun, 3 Mar 2013 15:04:29 +0000 (16:04 +0100)]
Fix several whitespace errors
* Remove whitespace at the end of lines.
* Replace space indentation by tabs.
* Add whitespace before/after several operators ("+", "-", "*", ...)
* Add whitespace to assignments ("foo = bar;").
* Fix whitespace in parameter lists ("foobar(foo, bar, 42)").
Matthew McClintock reported that older unannotated tags were not
correctly being cloned and did not appear in info/refs. Further
investigation revealed some dubious prefix comparison code for
determining whether or not to write refs in info/refs. After comparing
it with git's own http-backend.c, it appears upstream does not use this
prefix logic.
OTOH, I don't know what the reasoning was when the prefix logic was
introduced. It appears to me to just be buggy, though it's possible
there are other reasons, and we'll have to revisit this commit.
But for now, Works For Me.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> Signed-off-by: Matthew McClintock <msm@freescale.com> Reported-by: Matthew McClintock <msm@freescale.com>
Florian Pritz [Fri, 1 Feb 2013 09:59:13 +0000 (10:59 +0100)]
Make "owner" column on index page configurable
This is not really needed for personal sites where all repos belong to
the same person. Since it is pretty useful for shared sites however, it
should be configurable.
Signed-off-by: Florian Pritz <bluewind@xinu.at> Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Enhancements:
- path-selected submodule links
- intelligent default branch guessing
- /etc/mime.types lookup
- gitweb.* and cgit.* git-config support
- case insensitive sorting and age sorting
- commit, repository, and section sorting
- bold currently viewed page in pagination
- support BSDs in makefile
Bug Fixes:
- transition maintainer to Jason Donenfeld (zx2c4)
- download git snapshot from github instead of Lars' old server
- css fixes
- stablization of tests
- more compatible default highlight script
- suppress gzip timestamp so that tarballs only use tar timestamps
- treat ctags as target in makefile
- do not let global variables override certain local repo settings
- print ampersand as proper html entity
- use placeholder for empty commit subject
- format diff view for addition and removal of files
- point links at correct blob from ssdiff
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Julius Plenz [Tue, 30 Oct 2012 12:56:01 +0000 (13:56 +0100)]
ssdiff: point line links at exact blob by prepending "id="
Previously, the id_str (i.e. the current or diffed-against commit's
SHA1 ID) was simply concatenated to the URL. Now, prepend an "id="
string so that the links actually point to the right blobs and thus
the exact lines.
Signed-off-by: Julius Plenz <plenz@cis.fu-berlin.de> Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Michael Halstead [Wed, 14 Nov 2012 20:41:01 +0000 (12:41 -0800)]
Format git diff headers correctly when adding or removing files.
Copying the output of cgit and using it in patches now works when adding
files to or removing files from the repository. This is helpful for people
who use cgit in their patch workflow.
By not quoting the argument, an attacker with the ability to add files
to the repository could pass arbitrary arguments to the highlight
command, in particular, the --plug-in argument which can lead to
arbitrary command execution.
Tobias Bieniek [Tue, 9 Oct 2012 18:53:29 +0000 (20:53 +0200)]
ui-repolist: Add "section-sort" flag to control section sorting.
Flag which, when set to "1", will sort the sections on the repository
listing by name. Set this flag to "0" if the order in the cgitrc file
should be preserved. Default value: "1".
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
scan-tree: Unify gitweb.* and cgit.* settings into one config option.
After some back and forth with Jamie and René, it looks like the git
config semantics are going to be like this:
- gitweb.category maps to the cgit repo config key "section"
- gitweb.description maps to the cgit repo config key "desc"
- gitweb.owner maps to the cgit repo config key "owner"
- cgit.* maps to all cgit repo config keys
This option can be enabled with "enable-git-config=1", and replaces
all previous "enable-gitweb-*" config keys.
The order of operations is as follows:
- git config settings are applied in the order that they exist in
the git config file
- if the owner is not set from git config, get the owner using the
usual getpwuid call
- if the description is not set from git config, look inside the
static $path/description file
- if section-from-path=1, override whatever previous settings were
inside of git config using the section-from-path logic
- parse $path/cgitrc for local repo.* settings, that override all
previous settings
repo_config: do not let globals override repo settings
This makes it possible to activate the enable_commit_graph,
enable_log_filecount, and enable_log_linecount for individual
repositories, even if the global setting is "0" (default).
The commit that introduced the broken behavior was e189344, and the
commit message of that makes it clear that this wasn't the intended
behavior.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Ferry Huberts [Tue, 9 Oct 2012 11:10:48 +0000 (13:10 +0200)]
syntax-highlight: when the file has no extension, assume text
There are 2 situations:
1- empty extension: assuming text is better than highlight
producing no output because of a missing argument.
2- no extension at all: assuming text is better than setting
the extension to the filename, which is what now happens.
Jim Meyering [Mon, 23 Apr 2012 20:06:35 +0000 (22:06 +0200)]
do not write outside heap buffer
* parsing.c (substr): Handle tail < head.
This started when I noticed some cgit segfaults on savannah.gnu.org.
Finding the offending URL/commit and then constructing a stand-alone
reproducer were far more time-consuming than writing the actual patch.
The problem arises with a commit like this, in which the user name
part of the "Author" field is empty:
(this is due to buf=malloc(0); strncpy (buf, head, -1);
where "head" may point to plenty of attacker-specified non-NUL bytes,
so we can overwrite a zero-length heap buffer with arbitrary data)
Invalid write of size 1
at 0x4A09361: strncpy (mc_replace_strmem.c:463)
by 0x408977: substr (parsing.c:61)
by 0x4089EF: parse_user (parsing.c:73)
by 0x408D10: cgit_parse_commit (parsing.c:153)
by 0x40A540: cgit_mk_refinfo (shared.c:171)
by 0x40A581: cgit_refs_cb (shared.c:181)
by 0x43DEB3: do_for_each_ref (refs.c:690)
by 0x41075E: cgit_print_branches (ui-refs.c:191)
by 0x416EF2: cgit_print_summary (ui-summary.c:56)
by 0x40780A: summary_fn (cmd.c:120)
by 0x40667A: process_request (cgit.c:544)
by 0x404078: cache_process (cache.c:322)
Address 0x4c718d0 is 0 bytes after a block of size 0 alloc'd
at 0x4A0884D: malloc (vg_replace_malloc.c:263)
by 0x455C85: xmalloc (wrapper.c:35)
by 0x40894C: substr (parsing.c:60)
by 0x4089EF: parse_user (parsing.c:73)
by 0x408D10: cgit_parse_commit (parsing.c:153)
by 0x40A540: cgit_mk_refinfo (shared.c:171)
by 0x40A581: cgit_refs_cb (shared.c:181)
by 0x43DEB3: do_for_each_ref (refs.c:690)
by 0x41075E: cgit_print_branches (ui-refs.c:191)
by 0x416EF2: cgit_print_summary (ui-summary.c:56)
by 0x40780A: summary_fn (cmd.c:120)
by 0x40667A: process_request (cgit.c:544)
Invalid write of size 1
at 0x4A09400: strncpy (mc_replace_strmem.c:463)
by 0x408977: substr (parsing.c:61)
by 0x4089EF: parse_user (parsing.c:73)
by 0x408D10: cgit_parse_commit (parsing.c:153)
by 0x40A540: cgit_mk_refinfo (shared.c:171)
by 0x40A581: cgit_refs_cb (shared.c:181)
by 0x43DEB3: do_for_each_ref (refs.c:690)
by 0x41075E: cgit_print_branches (ui-refs.c:191)
by 0x416EF2: cgit_print_summary (ui-summary.c:56)
by 0x40780A: summary_fn (cmd.c:120)
by 0x40667A: process_request (cgit.c:544)
by 0x404078: cache_process (cache.c:322)
Address 0x4c7192b is not stack'd, malloc'd or (recently) free'd
Invalid write of size 1
at 0x4A0940E: strncpy (mc_replace_strmem.c:463)
by 0x408977: substr (parsing.c:61)
by 0x4089EF: parse_user (parsing.c:73)
by 0x408D10: cgit_parse_commit (parsing.c:153)
by 0x40A540: cgit_mk_refinfo (shared.c:171)
by 0x40A581: cgit_refs_cb (shared.c:181)
by 0x43DEB3: do_for_each_ref (refs.c:690)
by 0x41075E: cgit_print_branches (ui-refs.c:191)
by 0x416EF2: cgit_print_summary (ui-summary.c:56)
by 0x40780A: summary_fn (cmd.c:120)
by 0x40667A: process_request (cgit.c:544)
by 0x404078: cache_process (cache.c:322)
Address 0x4c7192d is not stack'd, malloc'd or (recently) free'd
Process terminating with default action of signal 11 (SIGSEGV)
Access not within mapped region at address 0x502F000
at 0x4A09400: strncpy (mc_replace_strmem.c:463)
by 0x408977: substr (parsing.c:61)
by 0x4089EF: parse_user (parsing.c:73)
by 0x408D10: cgit_parse_commit (parsing.c:153)
by 0x40A540: cgit_mk_refinfo (shared.c:171)
by 0x40A581: cgit_refs_cb (shared.c:181)
by 0x43DEB3: do_for_each_ref (refs.c:690)
by 0x41075E: cgit_print_branches (ui-refs.c:191)
by 0x416EF2: cgit_print_summary (ui-summary.c:56)
by 0x40780A: summary_fn (cmd.c:120)
by 0x40667A: process_request (cgit.c:544)
by 0x404078: cache_process (cache.c:322)
This happens when tail - head == -1 here:
(parsing.c)
char *parse_user(char *t, char **name, char **email, unsigned long *date)
{
char *p = t;
int mode = 1;
while (p && *p) {
if (mode == 1 && *p == '<') {
*name = substr(t, p - 1);
t = p;
mode++;
} else if (mode == 1 && *p == '\n') {
The fix is to handle the case of (tail < head) before calling xmalloc,
thus avoiding passing an invalid value to xmalloc.
And here's the reproducer:
It was tricky to reproduce, because git prohibits use of an empty "name"
in a commit ID. To construct the offending commit, I had to resort to
using "git hash-object".
git init -q foo &&
( cd foo &&
echo a > j && git add . && git ci -q --author='au <T at h.or>' -m. . &&
h=$(git cat-file commit HEAD|sed 's/au //' \
|git hash-object -t commit -w --stdin) &&
git co -q -b test $h &&
git br -q -D master &&
git br -q -m test master)
git clone -q --bare foo foo.git
AFAICS, this is not exploitable thanks (ironically) to the use of strncpy.
Since that -1 translates to SIZE_MAX and this is strncpy, not only does it
copy whatever is in "head" (up to first NUL), but it also writes
SIZE_MAX - strlen(head) NUL bytes into the destination buffer, and that
latter is guaranteed to evoke a segfault. Since cgit is single-threaded,
AFAICS, there is no way that the buffer clobbering can be turned into
an exploit.
ui-snapshot: pass -n to gzip, to suppress timestamp
Since cgit snapshots of tags are often used for releases, we don't
want the rarely used feature of the gzip compressor that includes
an embedded timestamp into the archive, since this makes each tarball
of the same (potentially signed) tag different.
This commit refactors the archive handling code a bit so that each
different format is able to run with an arbitrary argv for the filter.
ui-repolist: Case insensitive sorting and age sort
Add two options, one for doing the ordinary name sorts in a
case-insensitive manner, and another for choosing to sort repos in each
section by age instead of by name.