AgeCommit message (Collapse)AuthorLines
Pass over a notes ref that names no treeBryce Kwon-0/+9
A pusher can point refs/notes/commits or a signature notes ref at any object, and git dies loading one that is no tree, which ended every page of the repository, the dumb transport included. Such a ref now counts as absent.
Diff without asking xdiff for a minimal resultBryce Kwon-1/+3
XDF_NEED_MINIMAL turns off the cost limit in xdiff, so a file whose lines were reordered took time quadratic in its length: 2.1 s for one of 40000 lines, now 38 ms. Git leaves it off by default as well.
Hold object sizes in size_t as git 2.55 doesBryce Kwon-4/+4
Git 2.55 moved object sizes from unsigned long to size_t. The two are the same type on 64-bit hosts, but not on a 32-bit one, where GCC 14 rejects the mismatched pointers, so cgit stopped compiling there.
Share the max-blob-size check and its refusalBryce Kwon-3/+6
Five copies of the test and four of the error page had to agree by hand. The new check pins the four refusals, which no test covered.
Mark the parameters callbacks leave unusedBryce Kwon-1/+1
With these and the two changes before them, cgit builds under git's DEVELOPER=1 settings, which also annotate their own callbacks this way.
Pass callbacks to git by addressBryce Kwon-4/+4
ISO C has no conversion between a function pointer and void *, so the callbacks handed through git's data pointers and the write found by dlsym stopped the pedantic half of a DEVELOPER=1 build.
Keep string literals behind const pointersBryce Kwon-1/+1
Git's DEVELOPER=1 build turns on -Wwrite-strings as an error, and the literals stored through plain char pointers failed cgit's own objects there. The exec filter's argv takes the cast execvp needs once, the way git's run-command does.
Skip an indented comment in the mimetype fileBryce Kwon-3/+1
Only a hash in the first column marked a comment, so an indented one was read as the type "#" for every word after it. The manual already describes comments by their first non-blank character.
Keep date-format=local-local from failing pagesBryce Kwon-1/+8
The guard took local as a name that may carry the -local suffix, but git reads local only as a whole alias of default-local and dies on anything after it, so that one value turned every page into a 500.
Gate a readme ending in a colon like a disk pathBryce Kwon-0/+7
trust-scan-config passed any readme holding a colon as a tracked file, while the parser takes one only when a path follows the colon, so a scanned repository could still name a file on disk outside itself. The two now share one test.
Keep a repository with an empty url ignoredBryce Kwon-1/+0
The field resets at the end of cgit_add_repo cleared the ignore flag that the empty url branch had just set, so such a repository was listed on the index and selectable with ?r=, where every link to it read the byte before its url.
Replace the stray bytes of text that is not UTF-8Bryce Kwon-0/+49
A commit whose encoding header names something iconv cannot convert, or whose text is labelled UTF-8 without being it, reached a page declared UTF-8 with its bytes untouched. The atom feed already replaced them, so its UTF-8 reader moves to shared.c and every ident, subject, message and tag text passes through it.
Limit rawdiff to the requested pathBryce Kwon-11/+16
The diff and patch pages honour the path in the url, which the parameter table documents for every page, but rawdiff never handed it to git and answered with the whole diff.
Rename enable-trailers to enable-commit-trailersBryce Kwon-1/+1
Harden the request path, scan and error recoveryBryce Kwon-26/+79
Apply the mailmap at HEAD to every identBryce Kwon-0/+1
Split trailers out of the commit messageBryce Kwon-0/+2
Clean up the whole treeBryce Kwon-30/+17
Trim the comments and dead code across the treeBryce Kwon-15/+9
Rename the inconsistent config keysBryce Kwon-1/+1
noplainemail enable-plain-email noheader enable-header cache-root-ttl cache-index-ttl cache-repo-ttl cache-summary-ttl cache-scanrc-ttl cache-scan-ttl agefile age-file renamelimit rename-limit extra-head-content head-content
Fold `enable-stats` back into `max-stats`Bryce Kwon-1/+0
A `max-stats` period enables the page again, as it did before v2.2.0, so one key does both jobs.
Resolve submodule links from `.gitmodules`Bryce Kwon-0/+1
Log scan lock failures that are not contentionBryce Kwon-2/+3
Remove the repository homepage featureBryce Kwon-1/+0
Restyle the sources and fix the audit's findingsBryce Kwon-317/+303
Settle the names, types and layout of the sourcesBryce Kwon-34/+36
Add `enable-relative-dates` and `date-format`Bryce Kwon-0/+33
Cap diff blobs by `max-blob-size`Bryce Kwon-3/+22
The diff family inflated a blob of any size just to render it. A file over the limit is reported as binary instead.
Skip a ref whose object cannot be parsedBryce Kwon-1/+7
A dangling or unreadable ref was described through a null object.
Gate the stats page and add a language breakdownBryce Kwon-0/+1
`max-stats` only bounds the selectable periods now and no longer doubles as the enable switch. The tree walk runs before the history walk on purpose. Releasing commit memory while walking history resets each commit slab index, and a commit graph lookup afterwards would read another commit slot and walk the wrong tree. The stats fixture writes a commit graph so the tests cover that path. The history walk bounds the window in process rather than passing a formatted since date to `setup_revisions`, and parses each commit once.
Remove the owner filter hookBryce Kwon-1/+0
Point the fork's links and credits at itselfBryce Kwon-1/+1
AUTHORS now records the upstream developers as the project's history rather than as maintainers of the fork.
Bail out on a read error in `read_first_line`Bryce Kwon-2/+11
`read_in_full` returns a signed -1 on error, which became SIZE_MAX once stored in the size_t length and then wrote a terminator far out of bounds.
Tolerate a missing or empty repository URLBryce Kwon-1/+2
`trim_end()` returns NULL when a repo url is empty or all slashes and the later newline trim dereferenced it. The legacy `r=` and `p=about` path also read the last byte of the url without checking it was set.
Reorganize into source, assets and librariesBryce Kwon-0/+0
The C sources move to source/, the served files to assets/, the bundled Git submodule to libraries/git and the filter scripts to extensions/. Everything the build generates now lands in build/, so a clean is a single remove and the tree stays clean.
global: fix libc constness warningsJason A. Donenfeld-1/+2
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
cgit: truncate all config values at the newlineJason A. Donenfeld-2/+12
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>
ui-log: allow link following to be disabled per-repoJason A. Donenfeld-0/+1
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>
git: update to v2.53.0Christian Hesse-3/+2
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>
git: update to v2.52.0Christian Hesse-1/+1
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>
git: update to v2.51.1Christian Hesse-1/+1
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>
git: update to v2.47.0Christian Hesse-2/+2
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>
git: update to v2.46.0Christian Hesse-0/+2
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>
git: update to v2.41.0Christian Hesse-6/+11
Update to git version v2.41.0, with lots of changes... This requires changes for these upstream commits: * 60ff56f50372c1498718938ef504e744fe011ffb banned.h: mark `strtok()` and `strtok_r()` as banned * 52acddf36c8cb3778ab2098a0d95cc2e375a4069 string-list: multi-delimiter `string_list_split_in_place()` * d850b7a545fcfbd97460a921c7f7c59d933eb0f7 cocci: apply the "cache.h" part of "the_repository.pending" * cb338c23d6d518947bf6f7240bf30e2ec232bd3b cocci: apply the "commit-reach.h" part of "the_repository.pending" * ecb5091fd4301ac647db0bd2504112b38f7ee06d cocci: apply the "commit.h" part of "the_repository.pending" * 085390328f5fe1dfba67039b1fd6cc51546a4e41 cocci: apply the "diff.h" part of "the_repository.pending" * bc726bd075929aab6b3e09d4dd5c2b0726fd5350 cocci: apply the "object-store.h" part of "the_repository.pending" * bab821646a74c446370fa8d01ca851f247df5033 cocci: apply the "pretty.h" part of "the_repository.pending" * afe27c889429438829bc8818ed17e4960bd3ef02 cocci: apply the "packfile.h" part of "the_repository.pending" * 12cb1c10a64170a5d600dd1c6c8abfeec105fb6b cocci: apply the "refs.h" part of "the_repository.pending" * 035c7de9e9ea11d26df5f9e4bb117f91ed11a9fd cocci: apply the "revision.h" part of "the_repository.pending" ... and some more I missed to list 😜 - for example the move and cleanup of headers and includes (see changes in `cgit.h`) comes to mind... Signed-off-by: Christian Hesse <mail@eworm.de>
shared: fix bad free in cgit_diff_treeJune McEnroe-7/+5
Since git commit 244c27242f44e6b88e3a381c90bde08d134c274b, > diff.[ch]: have diff_free() call clear_pathspec(opts.pathspec) calling diff_flush calls free(3) on opts.pathspec.items, so it can't be a pointer to a stack variable. Signed-off-by: Christian Hesse <mail@eworm.de>
ui-tree: allow per repository override for enable-blameChristian Hesse-0/+1
The blame operation can cause high cost in terms of CPU load for huge repositories. Let's add a per repository override for enable-blame. Signed-off-by: Christian Hesse <mail@eworm.de>
git: update to v2.20.0Christian Hesse-1/+1
Update to git version v2.20.0. Required changes follow upstream commits: * 00436bf1b1c2a8fe6cf5d2c2457d419d683042f4 (archive: initialize archivers earlier) * 611e42a5980a3a9f8bb3b1b49c1abde63c7a191e (xdiff: provide a separate emit callback for hunks) Signed-off-by: Christian Hesse <mail@eworm.de>
git: update to v2.19.1Christian Hesse-1/+1
Update to git version v2.19.1. Required changes follow upstream commits: * commit: add repository argument to get_cached_commit_buffer (3ce85f7e5a41116145179f0fae2ce6d86558d099) * commit: add repository argument to lookup_commit_reference (2122f6754c93be8f02bfb5704ed96c88fc9837a8) * object: add repository argument to parse_object (109cd76dd3467bd05f8d2145b857006649741d5c) * tag: add repository argument to deref_tag (a74093da5ed601a09fa158e5ba6f6f14c1142a3e) * tag: add repository argument to lookup_tag (ce71efb713f97f476a2d2ab541a0c73f684a5db3) * tree: add repository argument to lookup_tree (f86bcc7b2ce6cad68ba1a48a528e380c6126705e) * archive.c: avoid access to the_index (b612ee202a48f129f81f8f6a5af6cf71d1a9caef) * for_each_*_object: move declarations to object-store.h (0889aae1cd18c1804ba01c1a4229e516dfb9fe9b) Signed-off-by: Christian Hesse <mail@eworm.de>
Fix gcc 8.1.1 compiler warningsJason A. Donenfeld-3/+4
CC ../shared.o ../shared.c: In function ‘expand_macro’: ../shared.c:487:3: warning: ‘strncpy’ specified bound depends on the length of the source argument [-Wstringop-overflow=] strncpy(name, value, len); ^~~~~~~~~~~~~~~~~~~~~~~~~ ../shared.c:484:9: note: length computed here len = strlen(value); ^~~~~~~~~~~~~ ../ui-shared.c: In function ‘cgit_repobasename’: ../ui-shared.c:136:2: warning: ‘strncpy’ specified bound 1024 equals destination size [-Wstringop-truncation] strncpy(rvbuf, reponame, sizeof(rvbuf)); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ CC ../ui-ssdiff.o ../ui-ssdiff.c: In function ‘replace_tabs’: ../ui-ssdiff.c:142:4: warning: ‘strncat’ output truncated copying between 1 and 8 bytes from a string of length 8 [-Wstringop-truncation] strncat(result, spaces, 8 - (strlen(result) % 8)); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
extra-head-content: introduce another option for meta tagsJason A. Donenfeld-0/+1
This is to support things like go-import meta tags, which are on a per-repo basis. Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Copyright © 2026 Bryce Kwon. All Rights Deprecated.