Age	Commit message (Collapse)	Author	Lines
16 hours	Percent-encode a submodule path in its link	Bryce Kwon	-1/+4
	A module-link template takes the path as the tree names it, so a path holding a space or a byte past ASCII gave an href that broke there. The url is now encoded the way a clone url is.
16 hours	Add filename* for a download name past ASCII	Bryce Kwon	-1/+59
	The plain, snapshot and blob pages sent a name such as café.txt raw in filename, which a client may read as ISO-8859-1. RFC 6266 has such a name sent again as filename*, percent-encoded in UTF-8, for a client that knows the form, with filename first for one that does not. The extended form carries UTF-8 alone, so a name that is not valid UTF-8 keeps the plain filename by itself.
16 hours	Percent-encode a clone url in its href	Bryce Kwon	-2/+26
	An href carried git://example.org/dépôt.git as it was, and the space of "with space" too, neither of which a url may hold. html_url_path would also encode the plus of git+ssh, which breaks the scheme, and the percent of an escape the url was written with, so a whole url gets a helper that encodes only the bytes RFC 3986 allows nowhere.
16 hours	Keep a space in a repository url in one clone url	Bryce Kwon	-3/+13
	The shared repository "with space" listed git://example.org/with and space.git as two clone urls, and the head linked the first alone. The manual calls clone-url a list of templates, so the list is now split before each template is expanded. A clone-prefix was right already, as the url is joined to it after the split.
16 hours	Show a full index page for a negative offset	Bryce Kwon	-0/+5
	The query parser takes an offset as low as -1, which the stats page reads as every author, so the index clamps it for itself, as the log and refs pages already do.
17 hours	Refuse a config directory before reading it	Bryce Kwon	-1/+7
	Git's build swaps in an fopen that refuses a directory, but only where it defines FREAD_READS_DIRECTORIES, and not every system's build does. There a directory opened, and only a failed read(2) reported it. POSIX leaves that failure to the system, and one that read the raw entries would have parsed them as settings.
18 hours	Report a max-stats value that names no period	Bryce Kwon	-2/+27
	A value cgit does not know as a period turns the statistics page off, so a typo such as "decade" read the same as leaving stats off on purpose. The empty value stays quiet because a cached repolist writes it for a repository that turned stats off.
19 hours	Hold the mailmap to max-blob-size	Bryce Kwon	-1/+11
	git's read_mailmap_blob inflates the whole blob, so a pusher could make every page of a repository read as much as they liked. git reads a note and the .gitmodules file whole as well, and capping those would mean bounding the lookups where the log, commit and tree pages make them, so the manual names them instead. Nothing tested that mailmap.file and mailmap.blob are ignored either.
22 hours	Keep a followed diff to the path it names	Bryce Kwon	-2/+13
	A followed diff kept every path that started with the requested one, so following small.c took small.cpp in as well. A path now has to be the requested one or lie below it, as the pathspec of an unfollowed diff matches.
22 hours	Name the followed path in the diffstat header	Bryce Kwon	-4/+8
	With follow set the diff runs over the whole commit and show_filepair applies the path, so the header read (limited to '') whether or not a path was given.
23 hours	Pin the listing's blame link for a late NUL	Bryce Kwon	-0/+4
	The listing judges binary by the first 8000 bytes, as git does, while the blame page and the file view look for a NUL anywhere, since one past that window would end a Lua filter's output early. The listing keeps the cheap sniff, because the whole-blob check would inflate every blob it names, so a file whose NUL lies past the window gets a blame link that ends on "Blob is binary". The test pins that trade, so a change to either side has to face it.
3 days	Let a keyboard scroll the stats table	Bryce Kwon	-1/+4
	It scrolls sideways like the hex dump, with nothing inside a keyboard can reach, so it takes focus and the shared focus ring the same way.
3 days	Name a pinned commit with the page in the title	Bryce Kwon	-3/+28
3 days	Let a keyboard scroll the hex dump	Bryce Kwon	-1/+3
	Chrome and Firefox already let an overflowing box take focus, with their own ring, but WebKit does not. The table takes focus on a mouse click as well, where focus-visible keeps the ring off.
3 days	Read the file and diff tables as layout	Bryce Kwon	-2/+2
	The file view's table only sets the line numbers beside the code, and the unified diff's is a single cell holding every file.
3 days	Say which index column is sorted	Bryce Kwon	-1/+6
4 days	Put a Lua filter's print output where it runs	Bryce Kwon	-0/+31
	LuaJIT and Lua 5.1 leave print's line in stdio's buffer until cgit exits, after the end of the page.
4 days	Report a snapshot format name that matches nothing	Bryce Kwon	-1/+15
4 days	Let cgit.conf set Git's build options	Bryce Kwon	-3/+10
	cgit.conf was read only after Git's Makefile, which had already settled NO_REGEX and the other options it tests as it read them, so setting one there changed nothing in Git's part of the build. The second read stays for the defaults Git assigns outright, such as CC and CFLAGS. A "+=" there now appends twice to a variable nothing resets in between, such as CGIT_CFLAGS, which only repeats the flag.
4 days	Answer an empty repository's feed with a feed	Bryce Kwon	-1/+10
4 days	Keep a repository's own settings in a cached scan	Bryce Kwon	-12/+33
	A scanned repository is read back from the cached list under the global settings, so a setting that only differed from them by being off, or by naming the default value, came back as the global one. branch-sort was written only for age, per-path module links not at all, and a filter a repository turned off left no line behind. Writing an empty value for a filter turned off makes the list depend on whether each global filter is set, so the scan key names that too.
4 days	Keep the placeholder out of a cached scan	Bryce Kwon	-2/+9
	The cached list wrote "[no description]" like any other description, so once read back it no longer counted as the placeholder and became the meta description of every repository without one. Repositories read back from the list are also marked as having had their description file read, since the scan already did, so the index does not open those files again on every request.
4 days	Leave the scan uncached with a negative cache-size	Bryce Kwon	-1/+1
	The page cache counts a cache-size of 0 or less as off, but scan-path cached its list for any value other than 0, so a negative one still wrote a list into cache-root.
7 days	Keep the login page out of search indexes	Bryce Kwon	-3/+6
7 days	Keep a protected repository's header to itself	Bryce Kwon	-3/+7
	The login page stands in for any page of the repository, so it named the description and owner of one the visitor may not see, including one hidden from the index. The site's own description fills the line, as on the index, so the header keeps its height.
7 days	Name the logo after the site	Bryce Kwon	-1/+5
7 days	Head the tag table's column for what it holds	Bryce Kwon	-2/+4
7 days	Say when a log search matches nothing	Bryce Kwon	-0/+5
7 days	Answer a submodule path with the submodule	Bryce Kwon	-14/+43
	The tree page read the gitlink's commit as a blob, which outside the repository is a bad object id and inside it is the commit's raw text.
7 days	Say when nothing matches a mistyped url	Bryce Kwon	-1/+23
7 days	Lead an error page's title with the error	Bryce Kwon	-5/+8
	The config error page had a title of its own, in the reverse order, and now takes the same one as any other 404, since its message already says the repository failed to open.
7 days	Keep the range in a large diff's patch link	Bryce Kwon	-6/+14
	The link named only id, so for a diff between two commits it led to a patch of the newer one alone, and without an id it lost the branch as well and led to the default branch's tip.
7 days	Offer blame only where it works	Bryce Kwon	-4/+37
	The blame page answers 404 for a symlink, since git blames a regular file alone, and it refuses a binary file and one past max-blob-size, yet a tree listing offered blame for all of them, and so did a symlink's own page. Each row already knows its size, and git calls a file binary by its first 8000 bytes alone, which a stream reads without inflating the rest. With blame on, a listing of 2,000 delta-packed files went from about 11 to 15 ms.
7 days	Dump only the first 64 KB of a binary file	Bryce Kwon	-6/+16
7 days	Blame the previous revision under its own name	Bryce Kwon	-4/+6
7 days	Lead a blame page's folder crumbs to the tree	Bryce Kwon	-2/+17
7 days	Link snapshots of a slash ref by a dashed name	Bryce Kwon	-4/+27
	The snapshot page refuses to work a revision out of a name holding a slash, since that name becomes the top directory of the archive, so every link for a ref like release/2.0 answered 404. The ref now goes in the link as the id, and the archive unpacks into the repository's own prefix, as any snapshot asked for by id does.
8 days	Show a renamed file's diff as the rename	Bryce Kwon	-11/+116
	A diff limited to the new name cannot see the old one, so the diffstat's link for a renamed file, the commit page reached from that file's log, and the file's rawdiff and patch all showed every line as added. A path the old tree lacks is now diffed across the whole commit with rename detection aimed at it alone, the way git blame looks for a rename, so it is compared with the commit's deleted files and nothing else. Doing it in cgit_diff_tree keeps the log's counts for that commit in step with the page they link to. The rawdiff and patch pages leave the diff to git's own output, so for them that search finds the old name, and both names go into the limit with rename detection on.
8 days	Say why the build leaves out Lua and what to do	Bryce Kwon	-1/+15
	A LUA_PKGCONFIG that pkg-config could not resolve, with no LUA_LIBS or LUA_CFLAGS given, still claimed to link with it and then failed on lua.h, and a build that found no pkg-config dropped Lua with no hint of why. A warning rather than an error lets the build go on, since flags supplied another way, such as CGIT_LIBS in cgit.conf, may still make it link.
8 days	Log the path of a repository that will not open	Bryce Kwon	-0/+9
	The 404 page names only the repository and nothing reached the log, so a mistyped repo.path left no trace of the path cgit had tried.
8 days	Read the description file without repo.desc	Bryce Kwon	-14/+39
	Only a scanned repository fell back to its description file, so one listed in cgitrc without repo.desc showed "[no description]" however its file read. repo.desc may follow repo.path, so the file is read where a page shows the description, which also spares every other request the reads.
8 days	Report a word where cgitrc expects a number	Bryce Kwon	-70/+136
	atoi reads "yes" or "true" as 0, so repo.hide=yes, or cgit.hide true in a scanned repository's git config, silently did nothing. The values keep that meaning, since honouring them now would hide repositories on upgrade, but each one is logged with its file and line.
8 days	Report a cgitrc or include that cannot be read	Bryce Kwon	-6/+33
	A missing or unreadable cgitrc left only "No repositories found" and an empty log, the commonest first mistake with no trace of its cause. A scanned repository's own cgitrc stays quiet, since most have none.
8 days	Keep a scan with a negative ttl for good	Bryce Kwon	-1/+2
	"The cache" promises that a negative ttl never expires, but no age was ever below a negative cache-scan-ttl, so every request started a background rescan.
8 days	List scanned repositories in path order	Bryce Kwon	-4/+14
	readdir hands entries back in whatever order the filesystem keeps them, so proj111 could come before proj105 and a section could start twice. section-sort is off by default to keep the order written in cgitrc, and a scan has none to keep.
8 days	Key the cached scan on every setting it uses	Bryce Kwon	-5/+59
	Two cgitrc files sharing a cache-root read one list, so strict-export and trust-scan-config set in one applied to the other, and a changed setting stayed stale until the list expired. The first request after an upgrade scans again, and the files under the old names are never read.
8 days	Log an owner lookup only when it fails	Bryce Kwon	-2/+6
	getpwuid leaves errno alone for a uid with no passwd entry, so a scan logged "Unable to read the owner" with "Success" as the reason for each repository such a uid owns.
8 days	Reset the repository title before reusing it	Bryce Kwon	-1/+3
8 days	Release only finished walks, never cgit's mailmap	Bryce Kwon	-4/+4
	release_revisions frees revs->mailmap, and the log page points it at the static list cgit_mailmap returns, so every log of a repository with enable-mailmap aborted. A walk whose setup failed holds commits that are already freed, so releasing it could free them twice.
8 days	Find Lua under the BSDs' package names	Bryce Kwon	-1/+4