* respect purge requests when serving ipro requests from ngx_pagespeed
* Add TODO to share common base with ApacheFetch.
During system-tests, don't write into the source-controlled htdocs area. (#1197)
Instead, use a new (not-yet-committed) target in Makefile.tests to build
a mirror htdocs in test/tmp/root that sym-links the large readonly
section and deep copies the test areas for purge and flush tests.
Pass the install-path into Makefile (#1200)
* Pass the install-path into Makefile, rather than having the Makefile
depend on PWD.
* Move the double-quotes be before the value, to be consistent with other
Makefiles.
The report from some time ago mentioned three Vary: headers,
but I can now only reproduce two using trunk-tracking plus the
original repro-configuration.
This fix unflags r->gzip_vary as set by the gzip module when PSOL
hands us Vary: Accept-Encoding, to make sure that nginx's core
header filter doesn't append another one.
Fixes https://github.com/pagespeed/ngx_pagespeed/issues/1064
- Fix nginx-side flow so we handle .pagespeed. resources ok
when they will land on a customized 404 internal location.
- Additionally, check for a wiped request context and make sure
we do not dereference a null pointer, which is what hurt in
the flow we entered above as the IPRO lookup still was
generating events while the nginx side request context was
gone.
- Also, as a preliminary measure, do not check fail when we
receive a stale event originating from a NgxBaseFetch that
is no longer associated with our request context.
Do log a warning so we'll hear about this happening either
through system test failures or a bug report.
Fixes https://github.com/pagespeed/ngx_pagespeed/issues/1081
- Add handling for header only responses
- Fixup nginx i/o handling
- Fix issues in parser state
- Revert timeout for native fetcher dns resolve to 10 seconds
- Ignore warning that gets logged with this change in tests during
fetching from a host that refuses connections.
Should help stabilizing (valgrind) test runs some more.
Nginx sometimes receives an expected response from 8.8.8.8 when
resolving host names, and the subsequent retry doesn't finish up
in 10 seconds. Bumping the timeout helps, but does not fully
explain why nginx is sometimes receiving an unexpected response.
I'll create a separate issue for investigating that.
After sending a SIGHUP, wait for the new worker process with the
new configuration to handle further requests, or else the tests
running after that might get confused when the new worker becomes
active with fresh shared mem statistics and shared mem cache.
This change adds a flag to optionally position the module at HTTP_AUX_FILTER_MODULES
instead of just in front of the gzip module. Doing so allows header module(s) to
fully manipulate any response headers before they are send to the user-agent.
With https://github.com/pagespeed/ngx_pagespeed/pull/880, SSI no longer hangs,
and other modules using subrequests should not do so either. SSI subrequests are
not optimized when ngx_pagespeed runs at HTTP_AUX_FILTER_MODULES, so the test for
it is added to the expected failures when the tests are run for this module position.
To build with the module respositioned:
POSITION_AUX=true ./configure ......
The same variable can be set before running the system tests, in which case it will
perform additional header tests and add the SSI test to the expected failures.
- Fix valgrind errors
- Add a test to make sure all logged output looks sane by whitelisting
current errors/warnings.
- Stop our nginx test instances after we are done testing.
- Add tests for shutting down and reloading configuration under high
load (depends on ab).
- Reduce the number of keepalive requests in the keepalive tests to speed
up test runs.
- Fix exiting with open file descriptors, fix cleanup in nginx's cache
manager/loader processes
- Attempt to finish up queued up NgxBaseFetches/requests on shutdown/reload
- Under valgrind the blocking rewrite started failing after adding a test
for reloading configuration under high load.
I've added it to the expected failures for valgrind, looking into this
is up next.
- Decline in ps_resource_handler when nginx is quitting. This makes us
more reliable on continued stress during shutdown/reload.
This change passes on non succesful status codes for pagespeed
resources and other places where we act as a content handler to
nginx. This has two benefits:
- Instead of a blank page, the user agent receives a formatted
and (hopefully customized and informative) response.
- Header modules are able to operate on that response, which was
requested in https://github.com/pagespeed/ngx_pagespeed/issues/612#issuecomment-58855816
- Fixes an issue in the html flow that would prevent overriding the
value of the 'Server' response-header.
- Add tests that ensure we emit a single and correct server header
in all flows when not overriding it.
- Add tests that ensure overriding the 'Server' response header
works. The resource and IPRO flow are added to the expected failures
as those are not working yet (and will be adressed in a follow-up).
Fixed https://github.com/pagespeed/ngx_pagespeed/issues/864 (html flow)
Joshua Marantz