| Age | Commit message (Collapse) | Author |
|---|
|
Add a copyright for changes done by me, also note in README the nature
of this repo
|
|
Check that a repo.path which does not exist or is not a repository is
reported, for the summary page, for an unknown page name and for the ci
page.
cgit used to hang on these requests rather than fail, so the helper puts a
deadline on each one: a regression should fail the test instead of wedging
the suite.
Signed-off-by: Saya Andy <saya.andy@posteo.com>
|
|
cgit cannot know whether the ref being viewed actually has a pipeline, so
the "ci" tab is offered for every ref and only reveals a missing one once
followed. Probing the ci system from cgit is not an option: the tab is
part of the page header, so it would mean a blocking request for every
page of every repository, cgit links no http client, and job pages are
usually not readable anonymously.
Add a ci filter instead, which receives the ref, whether it is a branch or
a tag, and the expanded url, and answers with its exit status. This keeps
credentials, timeouts and caching in a script, where they belong;
filters/ci-jenkins.sh demonstrates all three against Jenkins' REST API.
The verdict is memoized, so the filter runs once per request rather than
once for the tab and again for the redirect, and it governs the page as
well as the tab, so a hidden tab cannot be reached by typing the url.
Signed-off-by: Saya Andy <saya.andy@posteo.com>
|
|
Add a "ci" page which redirects to an external ci system for the ref
being viewed, along with a tab for it in the repository header.
The target is a url template, configurable globally or per repository,
with separate settings for branches and tags because pipelines for the
two commonly live at different locations: on a Jenkins multibranch
pipeline a branch is at job/<name>/job/<ref> while a tag is at
job/<name>/view/tags/job/<ref>. $ref, $repo and $slug are substituted,
$slug being the repository url with any ".git" suffix removed and slashes
replaced by dashes, which is the shape job names usually take.
An explicit repo.ci-url drops the branch and tag urls the repository
would otherwise inherit from the global settings, as it could never take
effect otherwise.
The expansion is written to the Location header verbatim rather than
through cgit_redirect(), which percent-encodes '?', '=' and '%' and would
corrupt any url carrying a query string. An expansion containing CR or LF
is refused so that it cannot smuggle in further headers.
Signed-off-by: Saya Andy <saya.andy@posteo.com>
|
|
print_header() fills the branch switcher from the main ref store whenever
ctx.repo is set. For a request which prints a page before
prepare_repo_cmd() has cleared ctx.repo -- an unknown page name for a
repository whose repo.path is broken, for instance -- the ref store has
never been initialized, and get_main_ref_store() aborts after the
response has already begun:
BUG: refs.c:2301: reference backend is unknown
get_main_ref_store
print_header at ui-shared.c:1057
cgit_print_error_page(fmt="Invalid request")
Add cgit_have_repository() to say whether the repository for this request
was opened, and skip the switcher when it was not.
Signed-off-by: Saya Andy <saya.andy@posteo.com>
|
|
prepare_repo_env() calls load_display_notes() unconditionally, even when
setup_git_directory_gently() has just reported that repo.path is not a
repository. init_notes() then resolves the default notes ref and
dereferences the repository's hash algorithm, which is still NULL:
EXC_BAD_ACCESS (code=1, address=0x18)
get_oid_basic(str="refs/notes/commits", len=18) at object-name.c:688
-> if (len == r->hash_algo->hexsz && !get_oid_hex(str, oid)) {
A typo in repo.path therefore took out the request. On Linux this
segfaults; on macOS nothing consumes the Mach exception, so the faulting
instruction is retried forever and the worker process spins at 100% CPU
instead of dying.
Only load the notes once a repository has been opened. prepare_repo_cmd()
reports the failure immediately afterwards, so such a request now renders
"Failed to open <repo>: No such file or directory".
Signed-off-by: Saya Andy <saya.andy@posteo.com>
|
|
`git describe` produces "v1.3.1-2-gfbb7", which is neither a Semantic
Version nor especially readable. Derive the version from the most recent
release tag and the number of commits made since it instead, giving
"v1.3.1+7", and plain "v1.3.1" when building the tag itself.
That is a valid Semantic Version 2.0.0 with the commit distance as build
metadata. Note that build metadata is excluded from precedence, so
"v1.3.1+7" and "v1.3.1" compare equal; the distance identifies a build
rather than ordering it.
Tarball builds keep falling back to the version in the Makefile, and
release tags are expected to be named vMAJOR.MINOR.PATCH for the result
to stay well-formed.
Signed-off-by: Saya Andy <saya.andy@posteo.com>
|
|
cgit.mk generates dependency files into .depend, but clean removed .deps,
so the stale .o.d files survived. They embed absolute compiler and SDK
paths, which makes a tree that has been moved, shared between machines,
or built inside a container fail after "make clean" with missing targets
instead of rebuilding.
Signed-off-by: Saya Andy <saya.andy@posteo.com>
|
|
memrchr() is a GNU extension which glibc and musl provide but macOS does
not, so cgit_set_title_from_path() failed to compile there and the tree
could not be built on macOS at all.
Replace it with a small local helper. The search covers one path
component at a time, so there is nothing to gain from a libc version.
Signed-off-by: Saya Andy <saya.andy@posteo.com>
|
|
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
Update to git version v2.54.0, this requires changes for these
upstream commits:
* d9ecf268ef3f69130fa269012318470d908978f6
odb: embed base source in the "files" backend
* cb506a8a69c953f7b87bb3ae099e0bed8218d3ab
odb: introduce "files" source
... and probably more related.
Signed-off-by: Christian Hesse <mail@eworm.de>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
These would be largely invalid anyway (save, I suppose, for Linux file
paths that technically can contain new lines).
The actual problem is that these get printed back out into cached -- and
trusted -- cgitrc files, and if the fields have newlines, the git-config
way of less trusted users configuring repos on a shared system can be
abused to inject newlines, which then can be used to smuggle global
options (including filters, which execute code) into the cached cgitrc.
So now, only ever duplicate up to the newline, when dealing with these
inputs.
Reported-by: Adrian Denkiewicz <adrian@doyensec.com>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
There's no reason to pass around function pointers. It was never used
for anything beyond one function.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
lock_slot() opens the lock file with O_RDWR|O_CREAT but without
O_TRUNC. If a previous cgit process was killed between lock_slot() and
unlock_slot() (e.g. by a CGI timeout or OOM), the stale lock file
remains on disk with the old content, and the kernel releases the fcntl
lock.
The next process to claim the same cache slot then opens this stale lock
file, acquires the fcntl lock, writes its key and generated content on
top of the old bytes. If the new response is shorter than what was
previously in the file, trailing bytes from the old response survive
beyond the end of the new content. fstat() in fill_slot() reports the
total file size (including the stale tail), and print_slot() faithfully
sends all of it -- producing a response that is the correct page
followed by a fragment of whatever previously occupied that lock file.
Fix this by truncating the lock file after acquiring the lock and before
writing the new key.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
We don't get any return value from compile_grep_patterns calling
compile_regexp_failed, causing the default die routine to print to
stderr and then for cgit to exit ungracefully.
Instead override the default die routine to show a normal error page.
Perhaps compile_grep_patterns ought to change upstream to return an
error. But this commit here will handle future issues as well, so
perhaps not a bad idea to do anyway.
Link: https://lists.zx2c4.com/pipermail/cgit/2026-March/004982.html
Link: https://lists.zx2c4.com/pipermail/cgit/2026-March/004983.html
Reported-by: Adrian C. <anrxc@sysphere.org>
Reported-by: Aiden Woodruff <aiden@aidenw.net>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
This exists for other CPU heavy operations like blame, but doesn't for
the follow functionality. Add it for that.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
In git 2.36, the logic for diff_flush() changed, resulting in our object
getting freed, which led to following renames producing garbage commits
until it eventually crashed.
Fixes: bb02e24 ("git: update to v2.36.0")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
... specified in the URL, with a curved arrow ahead of the line number.
Signed-off-by: Loïc <lagiraudiere+cgit@free.fr>
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Signed-off-by: Christian Barcenas <christian@cbarcenas.com>
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Signed-off-by: Christian Barcenas <christian@cbarcenas.com>
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Since the favicon setting defaults to "/favicon.ico", the user can
only unset it with favicon=. However, that would write an empty string
as the config value. Previously, such empty string always satisfied
the if condition. In this case it is better to omit the <link> entirely.
Signed-off-by: Christian Barcenas <christian@cbarcenas.com>
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
This fixes an error which was introduced by
2f50b47c72cbc4270bbd12ae7f520486d5f42736. Git 2.42.0 added a new argument
to config_fn_t, and it was added to gitconfig_config(), but not named.
This causes compile warnings/errors. This commit fixes that by naming the
new parameter, and marking it unused.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.53.0, this requires changes for these
upstream commits:
* bdbebe5714b25dc9d215b48efbb80f410925d7dd
refs: introduce wrapper struct for `each_ref_fn`
* 589127caa73090040200989ff4d24c3d54f473f2
packfile: move list of packs into the packfile store
* 5a5c7359f77ecd1bc4b0e172563161d602f131d3
refs: drop `current_ref_iter` hack
* b6e4cc8c32850315323961659e553d1d14591f7f
tag: support arbitrary repositories in parse_tag()
* 84f0e60b28de69d1ccb7a51b729af6202b6cf4c8
packfile: move packfile store into object source
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.52.0, this requires changes for these
upstream commits:
* 9f6dfe43c8a55b833ae16486bcafe29b543461f9
string-list: align string_list_split() with its _in_place() counterpart
* 78237ea53d6546aeab7adb2c7547a1177311ccde
packfile: split up responsibilities of `reprepare_packed_git()`
* 751808b2a18acba76b824aed4d8b7442bd7f5fca
packfile: refactor `get_packed_git()` to work on packfile store
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.51.2, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.51.1, this requires changes for these
upstream commits:
* e1d062e8ba0b72f49e9ef9713cc7011c330baab8
odb: drop deprecated wrapper functions
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.51.0, this requires changes for these
upstream commits:
* 8f49151763cb81adf4bcec53c1ae67057081b02d
object-store: rename files to "odb.{c,h}"
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.50.1, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.50.0, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.49.0, this requires changes for these
upstream commits:
* 88dd321cfedc6ee190dfafe4670a83ea33cdf4a3
path: drop `git_path()` in favor of `repo_git_path()`
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.48.0, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.48.0, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.47.1, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.47.0, this requires changes for these
upstream commits:
* e8207717f1623325fe1c95338fb03c1104ed5687
refs: add referent to each_ref_fn
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.46.2, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.46.1, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.46.0, this requires changes for these
upstream commits:
* e7da9385708accf518a80a1e17969020fb361048
global: introduce `USE_THE_REPOSITORY_VARIABLE` macro
* 9da95bda74cf10e1475384a71fd20914c3b99784
hash: require hash algorithm in `oidread()` and `oidclr()`
* 30aaff437fddd889ba429b50b96ea4c151c502c5
refs: pass repo when peeling objects
* c8f815c2083c4b340d4148a15d45c55f2fcc7d3f
refs: remove functions without ref store
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
fix error that is given because of macro overlapping cgit_filter member:
../filter.c:388:10: error: no member named '__fprintf_chk' in
'struct cgit_filter'
388 | filter->fprintf(filter, f, prefix);
| ~~~~~~ ^
/usr/include/bits/stdio2.h:92:3: note: expanded from macro 'fprintf'
92 | __fprintf_chk (stream, __USE_FORTIFY_LEVEL - 1, __VA_ARGS__)
| ^
1 error generated.
Signed-off-by: Denis Pronin <dannftk@yandex.ru>
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.45.2, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.45.1, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.45.0, this requires changes for these
upstream commits:
* 9720d23e8caf4adee44b3a32803a9bb0480118bd
date: make DATE_MODE thread-safe
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.44.0, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.43.2, no additional changes required.
(Git v2.43.1 fails to build, thus skipping.)
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.43.0, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
|
Update to git version v2.42.1, no additional changes required.
Signed-off-by: Christian Hesse <mail@eworm.de>
|
Saya Andy
Jason A. Donenfeld
Christian Hesse
Loïc
Christian Barcenas
Kian Kasad
Denis Pronin
