Redesign the caching layer

The original caching layer in cgit has no upper bound on the number of
concurrent cache entries, so when cgit is traversed by a spider (like the
googlebot), the cache might end up filling your disk. Also, if any error
occurs in the cache layer, no content is returned to the client.

This patch redesigns the caching layer to avoid these flaws by
* giving the cache a bound number of slots
* disabling the cache for the current request when errors occur

The cache size limit is implemented by hashing the querystring (the cache
lookup key) and generating a cache filename based on this hash modulo the
cache size. In order to detect hash collisions, the full lookup key (i.e.
the querystring) is stored in the cache file (separated from its associated
content by ascii 0).

The cache filename is the reversed 8-digit hexadecimal representation of

  hash(key) % cache_size

which should make the filesystem lookup pretty fast (if directory content
is indexed/sorted); reversing the representation avoids the problem where
all keys have equal prefix.

There is a new config option, cache-size, which sets the upper bound for
the cache. Default value for this option is 0, which has the same effect
as setting nocache=1 (hence nocache is now deprecated).

Included in this patch is also a new testfile which verifies that the
new option works as intended.

Signed-off-by: Lars Hjemli <hjemli@gmail.com>

2 files changed, 68 insertions, 1 deletions

diff --git a/tests/setup.sh b/tests/setup.sh
index 66bf406..e37306e 100755
--- a/tests/setup.sh
+++ b/tests/setup.sh
@@ -35,25 +35,25 @@ mkrepo() {
 }
 
 setup_repos()
 {
 	rm -rf trash/cache
 	mkdir -p trash/cache
 	mkrepo trash/repos/foo 5 >/dev/null
 	mkrepo trash/repos/bar 50 >/dev/null
 	cat >trash/cgitrc <<EOF
 virtual-root=/
 cache-root=$PWD/trash/cache
 
-nocache=0
+cache-size=1021
 snapshots=tar.gz tar.bz zip
 enable-log-filecount=1
 enable-log-linecount=1
 summary-log=5
 summary-branches=5
 summary-tags=5
 
 repo.url=foo
 repo.path=$PWD/trash/repos/foo/.git
 # Do not specify a description for this repo, as it then will be assigned
 # the constant value "[no description]" (which actually used to cause a
 # segfault).
diff --git a/tests/t0020-validate-cache.sh b/tests/t0020-validate-cache.sh
new file mode 100755
index 0000000..53ec2eb
--- a/dev/null
+++ b/tests/t0020-validate-cache.sh
@@ -0,0 +1,67 @@
+#!/bin/sh
+
+. ./setup.sh
+
+prepare_tests 'Validate cache'
+
+run_test 'verify cache-size=0' '
+
+	rm -f trash/cache/* &&
+	sed -i -e "s/cache-size=1021$/cache-size=0/" trash/cgitrc &&
+	cgit_url "" &&
+	cgit_url "foo" &&
+	cgit_url "foo/refs" &&
+	cgit_url "foo/tree" &&
+	cgit_url "foo/log" &&
+	cgit_url "foo/diff" &&
+	cgit_url "foo/patch" &&
+	cgit_url "bar" &&
+	cgit_url "bar/refs" &&
+	cgit_url "bar/tree" &&
+	cgit_url "bar/log" &&
+	cgit_url "bar/diff" &&
+	cgit_url "bar/patch" &&
+	test 0 -eq $(ls trash/cache | wc -l)
+'
+
+run_test 'verify cache-size=1' '
+
+	rm -f trash/cache/* &&
+	sed -i -e "s/cache-size=0$/cache-size=1/" trash/cgitrc &&
+	cgit_url "" &&
+	cgit_url "foo" &&
+	cgit_url "foo/refs" &&
+	cgit_url "foo/tree" &&
+	cgit_url "foo/log" &&
+	cgit_url "foo/diff" &&
+	cgit_url "foo/patch" &&
+	cgit_url "bar" &&
+	cgit_url "bar/refs" &&
+	cgit_url "bar/tree" &&
+	cgit_url "bar/log" &&
+	cgit_url "bar/diff" &&
+	cgit_url "bar/patch" &&
+	test 1 -eq $(ls trash/cache | wc -l)
+'
+
+run_test 'verify cache-size=1021' '
+
+	rm -f trash/cache/* &&
+	sed -i -e "s/cache-size=1$/cache-size=1021/" trash/cgitrc &&
+	cgit_url "" &&
+	cgit_url "foo" &&
+	cgit_url "foo/refs" &&
+	cgit_url "foo/tree" &&
+	cgit_url "foo/log" &&
+	cgit_url "foo/diff" &&
+	cgit_url "foo/patch" &&
+	cgit_url "bar" &&
+	cgit_url "bar/refs" &&
+	cgit_url "bar/tree" &&
+	cgit_url "bar/log" &&
+	cgit_url "bar/diff" &&
+	cgit_url "bar/patch" &&
+	test 13 -eq $(ls trash/cache | wc -l)
+'
+
+tests_done