index : cgit

#!/bin/sh

# Checks the ceilings a config can put on a page, the caps on refs listed,
# on the lines and files a diff renders inline, on the size of a blob any
# view will inflate, and the clamp on index query length. Crossing one has
# to trim the page or point somewhere better suited, never drop the request.

test_description='Check the ref listing and diff size limits'
. ./setup.sh

# The limits only show themselves against content that crosses them, so the
# fixture carries more branches and tags than max-ref-count allows and a
# commit whose oversized file diff sits beside a small one, which is what
# tells a per file limit apart from a whole page one.
test_expect_success 'set up limit fixtures' '
	mkrepo repos/limits 1 &&
	(
		cd repos/limits &&
		test_seq 1 60 >big.c &&
		printf "int x;\n" >small.c &&
		git add -A &&
		git commit -m sources &&
		test_seq 301 360 >big.c &&
		printf "int y;\n" >small.c &&
		git commit -am change &&
		for i in 1 2 3
		do
			git branch branch-$i &&
			git tag tag-$i || exit 1
		done
	) &&
	cgitrc_repo limitrc limits max-ref-count=2 max-diff-lines=20 max-diff-files=0 &&
	cgitrc_repo limitfilesrc limits max-diff-files=1
'

# An index query is compared against every repository the listing holds, so
# an enormous one multiplies out across the whole index. Clamping it rather
# than refusing it keeps the cost bounded without making an ordinary search
# behave any differently.
test_expect_success 'an enormous query is clamped, not rejected' '
	long=$(awk "BEGIN{s=\"\";for(i=0;i<4000;i++)s=s \"a\"; print s}") &&
	cgit_query "q=$long" >tmp &&
	grep "</html>" tmp &&
	longest=$(awk "{
		while (match(\$0, /aa*/)) {
			if (RLENGTH > n)
				n = RLENGTH
			\$0 = substr(\$0, RSTART + RLENGTH)
		}
	} END { print n }" tmp) &&
	test "$longest" -eq 512
'

test_expect_success 'an ordinary query still filters the index' '
	cgit_query "q=foo" >tmp &&
	grep ">foo</a>" tmp &&
	! grep ">bar<" tmp
'

# The combined refs page caps branches and tags separately, and the pages it
# hands the overflow to page on their own, so a limit that leaked between the
# two sections would show up as the wrong link or the wrong count here.
test_expect_success 'refs page lists max-ref-count branches and tags' '
	cgit_rc limitrc "url=limits/refs/" >tmp &&
	test $(grep -c "log/?h=" tmp) -eq 2 &&
	test $(grep -c "/tag/?h=tag-" tmp) -eq 2
'

test_expect_success 'refs page links each overflow to its own category' '
	cgit_ok limitrc "url=limits/refs/" >tmp &&
	test_grep "href=./limits/refs/heads.>\[\.\.\.\]</a>" tmp &&
	test_grep "href=./limits/refs/tags.>\[\.\.\.\]</a>" tmp
'

# The branches sort by name, so the second page holds branch-3 and master.
test_expect_success 'branch page paginates independently' '
	cgit_ok limitrc "url=limits/refs/heads/" >tmp &&
	test $(grep -c "log/?h=" tmp) -eq 2 &&
	test_grep "log/?h=branch-1." tmp &&
	test_grep "\[next\]" tmp &&
	cgit_ok limitrc "url=limits/refs/heads/&ofs=2" >tmp &&
	test_grep "\[prev\]</a> 3 - 4 of 4" tmp &&
	test_grep "log/?h=branch-3." tmp &&
	test_grep "href=./limits/log/.>master</a>" tmp &&
	test_grep ! "log/?h=branch-1." tmp &&
	test_grep ! "/tag/?h=tag-" tmp
'

test_expect_success 'tag page paginates independently' '
	cgit_ok limitrc "url=limits/refs/tags/" >tmp &&
	test_grep "/tag/?h=tag-1." tmp &&
	test_grep "1 - 2 of 3" tmp &&
	cgit_ok limitrc "url=limits/refs/tags/&ofs=2" >tmp &&
	test_grep "\[prev\]</a> 3 - 3 of 3" tmp &&
	test_grep "/tag/?h=tag-3." tmp &&
	test_grep ! "/tag/?h=tag-1." tmp &&
	test_grep ! "log/?h=" tmp
'

# A count of 0 lists every ref, and a negative one has to mean the same
# rather than reach the sort as a size.
test_expect_success 'a negative ref count lists every ref' '
	for n in 0 -1
	do
		{
			echo "summary-branches=$n" &&
			echo "summary-tags=$n" &&
			sed "s/^max-ref-count=2$/max-ref-count=$n/" limitrc
		} >count$n.rc &&
		cgit_rc count$n.rc "url=limits/refs/" >refs$n &&
		cgit_rc count$n.rc "url=limits/" >summary$n ||
		return 1
	done &&
	for page in refs0 summary0
	do
		test $(grep -c "<tr class=.branch.>" $page) = 4 &&
		test_grep "href=./limits/log/.>master</a>" $page &&
		test $(grep -c "<tr class=.tag.>" $page) = 3 &&
		for n in 1 2 3
		do
			test_grep "href=./limits/log/?h=branch-$n.>" $page &&
			test_grep "href=./limits/tag/?h=tag-$n.>" $page || return 1
		done || return 1
	done &&
	test_cmp refs0 refs-1 &&
	test_cmp summary0 summary-1
'

# A file that busts max-diff-lines is replaced by a link to its own page,
# where the reader asked for that one file and the limit no longer applies.
# The rest of the commit still renders, so one huge file cannot hide the
# small change beside it.
test_expect_success 'oversized file diff is replaced by a link' '
	cgit_rc limitrc "url=limits/commit/" >tmp &&
	grep "too large to be rendered inline" tmp &&
	grep "View it on its own page" tmp
'

test_expect_success 'small file in the same commit still renders inline' '
	grep "class=.add.>+int y;" tmp
'

test_expect_success 'the single-file diff page always renders in full' '
	cgit_ok limitrc "url=limits/diff/big.c" >tmp &&
	test_grep "class=.hunk." tmp &&
	test_grep ! "too large to be rendered inline" tmp
'

# Following renames empties the prefix so the diff machinery sees the whole
# commit, yet the page still shows the one file it names, so the cap has to
# come from the path asked for. The links keeping follow=1 show that the page
# took that path.
test_expect_success 'a followed single-file diff page renders in full' '
	{
		echo "enable-follow-links=1" &&
		cat limitrc
	} >followrc &&
	cgit_ok followrc "url=limits/diff/big.c&follow=1" >tmp &&
	test_grep "href=./limits/diff/big.c?follow=1." tmp &&
	test_grep "class=.hunk." tmp &&
	test_grep ! "too large to be rendered inline" tmp
'

# max-diff-files counts the files in a commit rather than the lines in one,
# so it drops the whole body back to the diffstat while a request naming a
# single file stays unaffected.
test_expect_success 'a commit changing too many files shows stat only' '
	cgit_rc limitfilesrc "url=limits/commit/" >tmp &&
	grep "too large to be rendered inline" tmp &&
	! grep "class=.hunk." tmp
'

test_expect_success 'the single-file page is not limited by max-diff-files' '
	cgit_ok limitfilesrc "url=limits/diff/small.c" >tmp &&
	test_grep "class=.hunk." tmp
'

# A path naming a directory still covers every file below it, so the caps
# go on applying there.
test_expect_success 'the diff caps hold when the path names a directory' '
	test_create_repo repos/capdir &&
	mkdir repos/capdir/dir &&
	for n in 1 2 3
	do
		echo $n >repos/capdir/dir/f$n || return 1
	done &&
	git -C repos/capdir add dir &&
	git -C repos/capdir commit -q -m dir &&
	cgitrc_repo capdirrc capdir max-diff-files=2 &&
	cgit_ok capdirrc "url=capdir/diff/dir" >tmp &&
	test_grep "too large to be rendered inline" tmp &&
	cgit_ok capdirrc "url=capdir/diff/dir/f1" >tmp &&
	test_grep "class=.hunk." tmp &&
	test_grep ! "too large to be rendered inline" tmp
'

# The patch the notice links to has to hold the change the page compared,
# both ends of a range and a branch other than the default.
test_expect_success 'the stat only notice links the patch of the same change' '
	new=$(git -C repos/limits rev-parse HEAD) &&
	old=$(git -C repos/limits rev-parse HEAD~2) &&
	cgit_rc limitfilesrc "url=limits/diff/&id=$new&id2=$old" >tmp &&
	grep "href=./limits/patch/?id=$new&amp;id2=$old.>patch</a> link" tmp &&
	cgit_rc limitfilesrc "url=limits/diff/&h=branch-1" >tmp &&
	grep "href=./limits/patch/?h=branch-1.>patch</a> link" tmp
'

# The diff views must not inflate a blob past max-blob-size merely to render
# it, which is the same defence the tree and plain views make and is easy to
# miss on this path. Such a file is reported as binary instead.
test_expect_success 'a diff of an oversized blob is not inlined' '
	mkrepo repos/blobdiff 1 &&
	(
		cd repos/blobdiff &&
		awk "BEGIN{for(i=0;i<400;i++)print \"aaaaaaaa\"}" >big.txt &&
		git add -A &&
		git commit -m add &&
		awk "BEGIN{for(i=0;i<401;i++)print \"aaaaaaaa\"}" >big.txt &&
		git commit -am change
	) &&
	cgitrc_repo blobdiffrc blobdiff max-blob-size=1 &&
	sha=$(git -C repos/blobdiff rev-parse HEAD) &&
	cgit_rc blobdiffrc "url=blobdiff/commit/&id=$sha" >tmp &&
	grep "Binary files differ" tmp &&
	! grep "aaaaaaaa" tmp
'

test_expect_success 'every page showing a blob refuses an oversized one alike' '
	{
		echo "enable-blame=1" &&
		cat blobdiffrc
	} >blobpagesrc &&
	for page in tree plain blob blame
	do
		cgit_rc blobpagesrc "url=blobdiff/$page/big.txt" >tmp &&
		grep "^Status: 413 Content Too Large" tmp &&
		grep "Blob size (3 KB) exceeds the limit (1 KB)" tmp || return 1
	done
'

# The limit is in whole kilobytes, so a blob of exactly that size is still
# within it.
test_expect_success 'a blob of exactly max-blob-size is shown' '
	printf "%01023d\n" 0 >repos/blobdiff/exact.txt &&
	test $(wc -c <repos/blobdiff/exact.txt) = 1024 &&
	git -C repos/blobdiff add exact.txt &&
	git -C repos/blobdiff commit -q -m exact &&
	for page in tree plain blob blame
	do
		cgit_ok blobpagesrc "url=blobdiff/$page/exact.txt" >tmp &&
		test_grep "00000000" tmp || return 1
	done
'

# The patch and rawdiff pages hand the work to git, whose own limit on an
# inlined blob is far past max-blob-size, so the cap is passed through to it.
test_expect_success 'patch and rawdiff show an oversized blob as binary' '
	test_seq 1 2000 >repos/limits/big.txt &&
	git -C repos/limits add big.txt &&
	git -C repos/limits commit -q -m "big blob" &&
	{
		echo "max-blob-size=1" &&
		cat limitrc
	} >blobrc &&
	for p in patch rawdiff
	do
		cgit_rc blobrc "url=limits/$p/" >tmp &&
		grep "^Binary files" tmp &&
		! grep "^+1999$" tmp || return 1
	done
'

test_done