Original fixes in c169dcb5221cf3dd452c291bf97374bb459cc5b9; didn't get merged in 39cad55711 because the code had been broken out to another file, but manual merge went smooth.
These affect twitterstatusfetcher.php on all 32-bit installs and some 64-bit installs (depending on whether the version of the JSON library reads the large numbers as long or double internally). 64-bit bug is harder to see as it tends to manifest as off-by-one due to losing a bit of precision off the end.
Added StartRegistrationTry/EndRegistrationTry calls into those three, and moved the actual recording hook to EndUserRegister which is guaranteed to be called from User::register (so we don't need to worry about other auth methods forgetting to call the other UI-code hooks).
Fixes for Twitter bridge breakage on 32-bit servers. New "Snowflake" 64-bit IDs have become too big to fit in the integer portion of double-precision floats, so to reliably use these IDs we need to pull the new string form now.
Machines with 64-bit PHP installation should have had no problems (except on Windows, where integers are still 32 bits)
Conflicts:
plugins/TwitterBridge/twitterimport.php <- as this hasn't been broken out, the import code is NOT FULLY UPDATED HERE.
Piwik's current default recommended JS for loading creates a <script> tag via document.write(). In addition to being generally evil, this means the browser doesn't know it's going to need piwik.js until that chunk of script gets executed... which can't happen until all scripts referenced *before* it have been loaded and executed.
The only reason for that bit of script though seems to be to pick 'http' or 'https' depending on the current page's scheme. This can be done more simply by using a protocol-relative link (eg "//piwik.status.net/piwik.js"), which the browser will resolve as appropriate. Since it's now sitting in the <script> tag, the browser's lookahead code will now see it and be able to start loading it while earlier things are parsing/executing.
May be better still to move to an asynchronous load after DOM-ready, but I'm not sure if that'll screw with the analytics code (eg, not being able to start things on the DOM-ready events since they're past).
The default full build of OpenLayers.js is 943kb as of 2.10; this gzips down to a couple hundred kb
but is still rather nasty, plus loading it off a remote host could slow things down.
Using a local copy let us cut down the size significantly by discarding unused features, and further
minification with yui-compressor shaves a bit more off. Cuts down to about 1/5 the size of the
original.
Also threw in a bundled & minified copy of the Mapstraction classes plus our usermap.js,
which covers the common case of using the default OpenLayers provider. This cuts out three
additional script loads, two of which weren't getting launched until after the mxn.js main
file got loaded.
Included Makefile will recreate the OpenLayers.js using the statusnet.cfg strip configuration file
and yui-compressor to do some extra minification at the end. Requires fetching the OpenLayers
source download and dropping it in:
http://openlayers.org/download/OpenLayers-2.10.tar.gz
$config['twitter']['ignore_errors'] = true;
A longer-term solution is to patch up the indirect retry handling to count retries better, or delay for later retry sensibly.
Most annoying error case being where the notice was already faved or deleted on Twitter! :)
Such errors will now just fail out and log a note to the syslog -- the rest of what we were doing will continue on unhindered, so you can still delete, favorite, etc and it just won't sync the info over in that case.
When the retweet failed with a 403 error (say due to it being a private tweet, which can't be retweeted) we would end up mishandling the return value from our internal error handling.
Instead of correctly discarding the message and closing out the queue item, we ended up trying to save a bogus twitter<->local ID mapping, which threw another exception and lead the queue system to re-run it.
- Fixed the logic check and return values for the retweet case in broadcast_twitter().
- Added doc comments explaining the return values on some functions in twitter.php
- Added check on Notice_to_status::saveNew() for empty input -- throw an exception before we try to actually insert into db. :)
Data are sent to the 'info' level of logging, like so:
[lazarus.local:4812.86b23603 GET /mublog/api/statuses/friends_timeline.atom?since_id=1353]
STATLOG action:apitimelinefriends method:GET ssl:no query:since_id cookie:no auth:yes
ifmatch:no ifmod:no agent:Appcelerator Titanium/1.4.1 (iPhone/4.1; iPhone OS; en_US;)
Fields:
* action: case-normalized name of the action class we're acting on
* method: GET, POST, HEAD, etc
* ssl: Are we on HTTPS? 'yes' or 'no'
* query: Were we sent a query string? 'yes', 'no', or 'since_id' if the only parameter is a since_id
* cookie: Were we sent any cookies? 'yes' or 'no'
* auth: Were we sent an HTTP Authorization header? 'yes' or 'no'
* ifmatch: Were we sent an HTTP If-Match header for an ETag? 'yes' or 'no'
* ifmod: Were we sent an HTTP If-Modified-Since header? 'yes' or 'no'
* agent: User-agent string, to aid in figuring out what these things are
The most shared-cache-friendly requests will be non-SSL GET requests with no or very predictable
query parameters, no cookies, and no authorization headers. Private caching (eg within a supporting
user-agent) could still be friendly to SSL and auth'd GET requests.
We kind of expect that the most frequent hits from clients will be GETs for a few common timelines,
with auth headers, a since_id-only query, and no cookies. These should at least be amenable to
returning 304 matches for etags or last-modified headers with private caching, but it's very
possible that most clients won't actually think to save and send them. That would leave us expecting
to handle a lot of timeline since_id hits that return a valid API response with no notices.
At this point we don't expect to actually see if-match or if-modified-since a lot since most of our
API responses are marked as uncacheable; so even if we output them they're not getting sent back to
us.
Random subsampling can be enabled by setting the 'frequency' parameter smaller than 1.0:
addPlugin('ApiLogger', array(
'frequency' => 0.5 // Record 50% of API hits
));
If someone tries to register from an IP address that a silenced user
has registered from, prevent it.
When silencing someone, silence everyone else who registered from the
same IP address.
