Age	Commit message (Collapse)	Author	Lines
15 hours	Sort with git's QSORT	Bryce Kwon	-21/+9
	QSORT skips a list of fewer than two entries, so an empty list's null base never reaches qsort, which is what cgit_qsort was written for.
15 hours	Allocate formatted strings with git's xstrfmt	Bryce Kwon	-30/+15
15 hours	Spell non-zero the way the tree does	Bryce Kwon	-3/+3
15 hours	Say the highlighter needs Scintillua's detect()	Bryce Kwon	-2/+3
15 hours	Reuse the ids t0109 has already looked up	Bryce Kwon	-2/+0
15 hours	Drop the patch greps the comparisons already make	Bryce Kwon	-20/+0
	The From, Date, Subject and signature lines are part of what the next test compares in full against git format-patch, for the single commit and for the range alike.
15 hours	Put the valgrind comment above the loop it explains	Bryce Kwon	-3/+3
15 hours	Drop two setup.sh lines that change nothing	Bryce Kwon	-5/+3
	The library has read TEST_RESULTS_SAN_DIR for the last time by the time it returns, and the trash directory it makes is empty, so there is no cache to remove.
15 hours	Tighten the checks that match stray text	Bryce Kwon	-8/+8
	Every page holds foo in its footer element, the foo+bar row holds bar, the branch switcher holds master, and commit 1 and file-1 also match commits 10 to 19 and file-10 to file-19. A search page repeats its term in the search box. So these checks passed on pages that lacked what they look for, and each now matches the link itself.
15 hours	Require a whole page when a filter exits early	Bryce Kwon	-2/+3
	When a filter cannot run, cgit dies mid-page with its Status line already out, so the check passed with no filter at all. That is the case wherever true lives only in /bin, as on busybox. true now comes from the PATH, and the page has to reach its end.
15 hours	Make the main cgitrc filter check run the filter	Bryce Kwon	-3/+6
	The check asked for the summary page, which never runs an about filter, so it passed with exec:/bin/true on macOS, which has no such file, and with a filter that does not exist at all. Its first line read the stderr the previous test left behind. execvp searches the PATH, so a plain true runs the same on macOS, Debian and busybox.
15 hours	Let go of a slot lock that cannot take its key	Bryce Kwon	-0/+18
	A failed truncate or key write returned with the descriptor still open, so the lock was held while the page rendered uncached and the file was left behind. It is removed before its last close, as every holder does.
15 hours	Keep lock contention out of the error log	Bryce Kwon	-13/+45
	process_slot logged every failure to lock, so each request that found its slot being filled left a line, while refresh_slot logged none, so an expired slot in a directory the server cannot write was served stale for good without a word. lock_slot now logs the step that failed, as the scan does, and leaves out only another request at work.
15 hours	Lock the cached repolist with flock	Bryce Kwon	-9/+42
	The slot lock has to be an flock, since an exec filter drops a record lock when it closes the stdout a fill writes through. The scan only got by with fcntl because nothing reopens its file while it runs. One kind of lock lets the two share their code, and on Linux, where the kinds never see each other, an flock on the scan's file now holds a scan off like any other.
15 hours	Open the lock files close-on-exec	Bryce Kwon	-2/+31
	Each exec filter and snapshot compressor a fill runs inherited the slot's lock descriptor beside the stdout it writes into. An flock lasts until its last copy closes and nothing waits for a process a filter leaves behind, so a fill killed before its rename could leave the slot locked, and served stale or uncached, for as long as such a process lived. The scan holds its lock across no exec and opens the same way so that the two stay alike.
15 hours	Tidy the cache test's helpers and outputs	Bryce Kwon	-24/+18
	The header's last sentence named two of the many later subjects, so it goes too. The second request's stderr is still kept, for the lock tests that come next.
15 hours	Use test_dir_is_empty for the empty cache checks	Bryce Kwon	-16/+8
	The library's check also lists whatever was left behind when it fails.
16 hours	Catch a lexer's failure in one place	Bryce Kwon	-12/+9
	filter_close already runs highlight under pcall and serves the escaped text when it fails, so the pcalls around detect, load and lex inside it only turned a raise into the same fallback a step earlier.
16 hours	Write a session's expiry in whole seconds	Bryce Kwon	-0/+21
	On Lua 5.3 and later a length worked out with a fraction, such as 1.5 * 24 * 60 * 60, is a float, so the signed expiry came out as 1793609024.0, which validate_value refuses, and every login failed. Max-Age took the same .0, which makes a browser ignore it.
16 hours	Drop an auth prerequisite no test asks for	Bryce Kwon	-4/+2
16 hours	Set the Lua prerequisite once for every test	Bryce Kwon	-87/+27
	setup.sh already asks the binary, so the seven scripts that each turned the answer into a prerequisite now share one. test-lib names the missing prerequisite beside every test it skips, so the notes that cgit lacks Lua go, while the ones announcing a fallback stay. The list of standalone interpreters moves into lib.sh and matches the versions the extensions claim directly.
16 hours	Move the highlighter fallback check into t0502	Bryce Kwon	-24/+12
	t0502 owns the shipped highlighter and already writes a cgitrc naming it, so the check needs no config of its own there. t0204's header still called it the last case long after it stopped being one.
16 hours	Share the request helpers in the auth checks	Bryce Kwon	-77/+49
	The two identical posts for an unknown user become one, and no two checks share a name, so a failure report points at a single check.
16 hours	Let an unmirrored sink fail the way Lua would	Bryce Kwon	-12/+0
16 hours	Drop the auth stubs no check reaches	Bryce Kwon	-20/+6
	Every check replaces get_secret, the only caller of posix.sys.stat, link and unlink, and nothing outside the harness calls preload.
16 hours	Collect an email filter's writes like the others	Bryce Kwon	-10/+12
16 hours	Say what really differs between the auth filters	Bryce Kwon	-6/+8
	The lookups were never the only difference, since the configuration and trim's place differ as well. The configuration is now set off by blank lines, as in the other extensions, instead of a heading.
16 hours	Drop auth guards for input cgit never sends	Bryce Kwon	-21/+10
	cgit names no repository with an empty string, never nil, so the check that the index stays up proved nothing, and auth-file in fact refuses the index too while its repos file cannot be read. A session_seconds of 0 signs a session that has expired by the time it comes back, so it needs no branch of its own in set_cookie.
16 hours	Strip control characters from the location alone	Bryce Kwon	-20/+12
	set_cookie only ever writes an empty value or a token secure_value built from url-encoded text, digits and hex, so stripping it changed nothing.
16 hours	Drop the URL helpers' guards that cannot fire	Bryce Kwon	-18/+2
	Every caller hands the helpers a string, and a value with a line break is refused by validate_value whether or not CR LF is folded on the way.
16 hours	Keep only the request fields the auth filters read	Bryce Kwon	-42/+10
	Seven of the twelve fields filter_open stored were never read. Naming all twelve parameters keeps the order cgit passes them in on record.
16 hours	Take a signed cookie apart with one match	Bryce Kwon	-82/+11
	The loop let a valid cookie through with a field appended, a doubled separator or a zero in front of its expiry. The payload now carries the expiry as written, so only the spelling secure_value issues verifies.
16 hours	Drop a highlighter check that tested the harness	Bryce Kwon	-6/+0
	The filter hands a NUL through untouched, and only the harness's copy of cgit's sink cut the text there, so the check would have failed the day the filter replaced NULs the way about-render does.
16 hours	Drop the highlighter's extension fallback	Bryce Kwon	-61/+13
	The table of extensions and the second lookup served only a Scintillua with no detect(), which now leaves every file uncoloured but escaped. The fake detect() names a lexer after the file's suffix, so every test lexer is reached through it.
16 hours	Call string.match like the rest of the highlighter	Bryce Kwon	-1/+1
16 hours	Drop a stray blank line from the about-render tests	Bryce Kwon	-1/+0
16 hours	Carry a man page's text as plain strings	Bryce Kwon	-23/+16
	lpeg's constant, query and numbered captures already do what the token helpers did, and nothing ever read the tokens of an escape that draws nothing.
16 hours	Take a no fill line's plain text from its nodes	Bryce Kwon	-7/+1
16 hours	Wrap a man page's font runs with the macros' helper	Bryce Kwon	-27/+11
16 hours	Reuse the digit pattern where it is spelt out	Bryce Kwon	-2/+2
16 hours	Build the block start test from the line grammars	Bryce Kwon	-2/+1
16 hours	Count a rule among the blocks that end a paragraph	Bryce Kwon	-7/+4
16 hours	Inline the list check before indented code	Bryce Kwon	-6/+2
16 hours	Read a setext heading's level off its capture	Bryce Kwon	-10/+4
16 hours	Build every code block in one place	Bryce Kwon	-17/+11
16 hours	Replace a readme's NUL bytes in one substitution	Bryce Kwon	-19/+3
16 hours	Let the dispatch serve markdown without lpeg	Bryce Kwon	-5/+0
	Without lpeg both renderers stay nil, and the handler lookup already falls back to plain text for a format it has no renderer for.
16 hours	Render markdown and man pages through one function	Bryce Kwon	-28/+18
	Everything that can raise already runs inside the parse and emit pcalls, so the outer one in filter_close could only ever have sent the page a second time. The size check moves to the dispatch with it.
16 hours	Drop three needless forward declarations	Bryce Kwon	-9/+3
16 hours	Drop fallbacks for a match that cannot fail	Bryce Kwon	-3/+3
	A table capture of a pattern repeated zero or more times matches every string, the empty one included, so the inline grammars never return nil.