Includes minor additions to test only.
Trunk tracking updates to sync to PSOl r3677.
Trunk tracking update to sync to PSOL r3677
Includes minor additions to test only.
It doesn't make sense to fetch_until the uncombined URL and then
grep for combined one: you might just get entirely unoptimized
output and fail the test. Instead wait for the combining to happen,
and make sure it combined the right # of things.
This makes nginx run in the background under valgrind,
with both a master and a child process.
Valgrind errors will be redirected to `valgrind.log`.
When `USE_VALGRIND` is set, all system tests will be run under valgrind,
and at the end a new test is appended which ensures no valgrind errors
where encountered.
It is also worth noting that:
- There is a new file, `valgrind.sup`, which contains a few suppressions.
- Some tests behave flakey under valgrind. For now these are appended
to the expected failures (when under valgrind only).
- 'Possibly lost' errors are all suppressed to get the amount of false
positives manageable.
The shared circular buffer wasn't hooked up fully, which meant loading
/ngx_pagespeed_messages didn't work. This fixes that and adds a test.
I also noticed while adding this that the 'Handling of large files' test
wasn't set up properly, so I converted that to use start_test.
Fixing that exposed another bug where the 'Handling of large files' test
was actually failing but being marked as an expected failure by being
grouped in with the test above. Adding `pagespeed MaxHtmlParseBytes 5000`
to the appropriate location made it test what it was supposed to be testing
again, and the underlying feature wasn't broken.
Changes:
* r3585: With downstream caching, don't touch `Cache-Control` headers.
* No longer require `Modify Caching Headers off`.
* Change from `modify_caching_headers` as a boolean to a three valued enum
`PreserveCachingHeaders`.
* r3596: Make tests less flaky.
* Changes WGET_DUMP to write to WGET_DIR instead of OUTDIR.
* r3597: Remove now-redundant system tests.
* Some system tests were moved into `automatic/system_test.sh` which means we
can remove our forks.
* r3598: Enable the shared memory metadata cache by default.
* The code change just worked and needed no changes, but it also added
substantial tests, which needed some porting.
* Deflake the `scrape stats` test by moving it before config reloading.
The keepalive tests were occasionally bumping up against
the 1s timeout we had set for curl, but a 1s 99.9% tail
latency is a little agressive. Drop it to 2s to make the
tests less flaky.
What we really need is a proper load testing setup, so
the keepalive test isn't doing double-duty.
This fixes a few valgrind errors on shutdown, plus a potential
issue when url_util.cc would lazily initialize in a thread-unsafe
manner. Also adds --leak-check=full to the suggested valgrind
command for the system tests.
* Adds IPRO tests that we should have copied from apache_system_test.sh back
when we first added IPRO.
* The "IPRO-optimized resources should have fixed size, not chunked" test
fails. The problem is as it says: IPRO optimized resources are being sent
out with `Transfer-Encoding: Chunked` instead of `Content-Length: ...`. How
do we fix this?
* install/system_test.sh moved to net/instaweb/automatic/system_test.sh
* minor method signature changes
* url.is_valid() is gone, replaced by .IsWebValid()
* I pushed Otto's #464 upstream, and now we need to make some IPRO changes.
* beacons have nonces now, so we need to update tests.
As of r3447 ServerContext::ChildInit is now in system/, which means we can
delete a lot of code.
DangerFetchFromUnknownHosts and FetchWithGzip are now handled by RewriteOptions.
ServerContext now needs a hostname and port, but we don't have one so we
hackishly provide a different unique id.
Squash-merge of chaizhenhua's work over many pull requests, especially #450.
The copy of in_place_resource_recorder.cc is temporary and can be removed after
the next PSOL release.
This makes messages written via the NgxMessageHandler from the
NgxServerContext be written with the nginx error_log configuration
specified at the server{} level, if specified.
Fixes https://github.com/pagespeed/ngx_pagespeed/issues/447
This fixes sending out the weak ETag for ngx_pagespeed_static/js_defer.hash.js,
and simplifies things somewhat for this code path.
It also adds a test for it.
Fixes https://github.com/pagespeed/ngx_pagespeed/issues/434
* ConsoleHandler now takes a RewriteOptions
* Console needs ports to override the statistics handler path. Before the
console just assumed statistics were available at /mod_pagespeed_statistics.
* The "nostore" test fails. I've marked this as an expected failure and will
fix it in a followup.
* downstream_cache_purges stat was renamed to downstream_cache_purge_attempts
* Proxying tests with gstatic no longer use 1.gif but are switched to
pss-architecture.png.
* DownstreamCacheLifetimeMs option was removed.
This matches two changes:
- Anupama made the caching tests use downstream_caching.html
(Thanks to her for help in updating this)
- Jan updated some configuration defaults, which made some images
compress more.
Adding a nginx conf snippet for constructing a proxy_cache_key based on the UserAgent and expected optimizations for that UserAgent class, bypassing the cache for all other cases. Sample config for downstream-cache proxy_cache_key construction now takes care of 2 classes of UserAgents (expected to get most of the traffic for a site) and also bypasses cache for all other desktop UAs, all mobile UAs and all tablet UAs.
(Fixing the startup phase of the nginx server to put out the original error message in case of failure.)
Add support for proxy_cache integration with pagespeed with:
a) Handling cache-hit-header to bypass content handling by pagespeed.
b) pagespeed_test_conf.template with the proxy_cache configuration blocks.
c) System tests for checking hit and miss headers and purge requests.
d) Checking for proxy_cache_purge module during start-up for tests.
Squash-merge of Anupama's #404
Updating trunk-tracking to compile and work with mod_pagespeed svn revision 3061 with 2 small modifications to serf_url_async_fetcher.cc to disable SERF_HTTPS_FETCHING.
Please note that anyone building mod_pagespeed from source will need to
a) edit serf_url_async_fetcher.h to change the line "#define SERF_HTTPS_FETCHING 1" to have 0.
b) edit serf_url_async_fetcher.cc to move the definition of "apr_status_t status = APR_SUCCESS;" on line 265 to be inside the "#if SERF_HTTPS_FETCHING" block (in order to avoid a compilation error for psol/mod_pagespeed).
Highlights of this change:
1) Propagating renaming changes related to
a) Furious -> Experiment
b) ModPagespeed -> PageSpeed.
2) UrlPollableAsyncFetcher going away.
3) RewriteOptions now requiring a ThreadSystem parameter for construction.
4) Removing client_property_cache references and usages (since this was removed from mod_pagespeed).
5) Propagating beacon construction changes (where POST requests can now carry the url query param in the request).
6) General fixes for API/test breakages in statistics_logger, fallback property page cache,
Add a keepalive test which is repeated 100 times, with us both
behind the gzip filter and directly behind nginx's output filter.
The reason for that is that modules running after us can mask
our errors, this at least tests the most used ones.
This CL adds a failure to the expected failures, as currently,
this tests gives suspicious messages in nginx's error.log:
"client prematurely closed connection while sending response to client"
We should probably also test other code paths, like .pagespeed.
resources and the beacon with post requests as well. If this approach
is generally ok, I can proceed to add those as well.
Jeff Kaufman