| Commit message (Collapse) | Author | Age | Lines |
|\ |
|
| |
| |
| |
| | |
large databases, the sorting of such large batches causes an extreme slowdown).
|
| |
| |
| |
| |
| |
| |
| | |
To be fair this change does not seem to make any tests pass that
didn’t pass before, but (based on my limited understanding) it
seems like a good idea. It also conforms to how FakeWeb is used
in application_controller_spec.rb.
|
|/
|
|
|
|
| |
There were some order-dependent test failures that turned out to
be caused by the fact that the RoutingFilters were cleared and
not subsequently restored, by some tests.
|
|\ |
|
| |\ |
|
| | |
| | |
| | |
| | | |
a race condition and fixes #304
|
| |/ |
|
|/
|
|
|
|
|
| |
(If DEBUG_RECORD_MEMORY is enabled)
This is useful for tracking down bugs that cause Rails to go into an infinite or very long
loop, as the recent Xapian wildcard bug does.
|
|\ |
|
| |
| |
| |
| | |
Work (with subtly reduced resilience) on earlier versions.
|
|\| |
|
| |\ |
|
| | |\
| | | |
| | | |
| | | |
| | | | |
Conflicts:
app/views/public_body/_search_ahead.rhtml
|
| | | | |
|
| | | | |
|
| | | | |
|
| | | |
| | | |
| | | |
| | | | |
fails.
|
| | | | |
|
| | | |\ |
|
| | | | |
| | | | |
| | | | |
| | | | | |
didn't call "integrate_views", so the rendering part was mocked by RSpec instead of executed.
|
| | |\ \ \
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Conflicts:
app/controllers/application_controller.rb
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Wildcard searches in Xapian can expand uncontrollably, consuming
all available RAM and crashing the server. This has been a real
problem on WhatDoTheyKnow.com. The underlying issue is tracked
in http://trac.xapian.org/ticket/350
This changeset imposes a limit on wildcard expansion. The type-ahead
search will first try a wildcard query, and if that fails because
of too much expansion will fall back to a plain non-wildcard search.
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
...by forcing the IncomingMessage to be re-parsed after we have
edited it.
|
| | | | | | |
|
| | |/ / /
| |/| | | |
|
| | | | |
| | | | |
| | | | |
| | | | | |
Another temporary patch to protect the live site.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Do not use FLAG_WILDCARD or FLAG_PARTIAL anywhere, because wildcard
searches can crash the whole server as per http://trac.xapian.org/ticket/350
(e.g. http://www.whatdotheyknow.com/body/search_ahead?query=app will
crash the server prior to this commit.)
This is a temporary fix while we investigate the possibilities of
the set_max_wildcard_expansion() method, which should make it possible
to alleviate this issue.
|
| |/ / / |
|
| | | | |
|
| | | | |
|
| | | | |
|
|/ / / |
|
| | |
| | |
| | |
| | | |
Same problem as previous commit to this (wdtk) branch.
|
| |/
|/|
| |
| |
| | |
Pagination for authority search results in the new request workflow
seems to be broken very badly, so disable it for now.
|
|\ \
| |/
|/| |
|
| |
| |
| |
| |
| | |
The behaviour of elinks is locale-dependent. This patch forces LANG=C
which makes the behaviour consistent across platforms.
|
| | |
|
| |
| |
| |
| | |
Because we want the new version of external_command.rb
|
| |
| |
| |
| |
| | |
because for some reason iconv is assuming big-endian on my MacBook Pro
(OS X 10.7.2) unless told explicitly, and this causes test failure.
|
| |
| |
| |
| |
| | |
Allow the WDG HTML validation utility "validate" to be anywhere
in UTILITY_SEARCH_PATH, rather than hard-coding /usr/bin.
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The trouble with `which command` is twofold:
- It spawns a whole shell just to find out the path to a binary, every time;
- The results are environment-dependent, since they depend on $PATH. It would
be better to specify the search path in the configuration file where everything
else is specified rather than in the environment.
This commit replaces it with the new mechanism from AlaveteliExternalCommand.
|
| | |
|
| |
| |
| |
| |
| |
| | |
So do not rely on a hard-coded path. Also change the order of arguments
and use /dev/stdout rather than -, so it works with the Mac/BSD version
of uudecode as well as the Linux version.
|
| | |
|
| |
| |
| |
| |
| |
| | |
Look for external commands in a config-defined path (defaulting
to /usr/bin:/usr/local/bin), rather than requiring the path to be
hard-coded or the caller to resolve it.
|
| |\ |
|
| |/
|/| |
|