Brion Vibber [Sat, 18 Dec 2010 00:22:26 +0000 (16:22 -0800)]
Sort indexing fix for profile sidebar: add group_member_profile_id_created_idx to group_member table, streamlines sorting of your group memberships in the sidebar
Brion Vibber [Fri, 17 Dec 2010 21:03:18 +0000 (13:03 -0800)]
Notice::whereSinceId() and Notice::whereMaxId() encapsulate logic for building where clauses for since_id/max_id parameters. Can override the field names from 'id' and 'created'.
Brion Vibber [Fri, 17 Dec 2010 20:09:02 +0000 (12:09 -0800)]
Notice::getAsTimestamp() static function to look up the timestamp for a given notice, even if it's been deleted. To be used for converting since_id/max_id processing to use timestamp sorting internally.
Brion Vibber [Fri, 17 Dec 2010 01:02:02 +0000 (17:02 -0800)]
Tickets #2112, 2333, 1677, 2362, 2831: fix AJAX form posting on SSL page views with ssl=sometimes
These have been failing for ages due to our outputting full URLs all the time, usually with the default protocol instead of the current one.
Forms would get output with an http: URL in their contents even when destined for an HTTPS page; while a regular form submission would just warn you about the secure->insecure transition, the AJAX code was failing outright and then not bothering to fall back to the regular submission.
I found it was easy to detect the mismatch -- just check the target URL and the current page's protocol before submitting.
Since failing over to non-AJAX submission to the HTTP URL throws up a warning, I figured it'd be easier (and much nicer for users) to just let it rewrite the target URL to use the secure protocol & hostname before doing the final submit.
This check is now automatically done for anything that calls SN.U.FormXHR() -- making most of our buttons on notices and profile/group headers work naturally.
The notice form setup code also runs the rewrite, which gets posting working without an error dialog.
I'd prefer in the long run to simply use relative URLs in most of our output; it avoids this problem completely and lets users simply stay in the current protocol mode instead of being constantly switched back to HTTP when clicking around.
(Note that folks using the SSLAlways extension to Firefox, for instance, will have their browsers constantly sending them back to HTTP pages, mimicking the desired user experience even though we haven't fully implemented it. These folks are likely going to be a lot happier with forms that submit correctly to go along with it!)
Brion Vibber [Fri, 17 Dec 2010 00:18:49 +0000 (16:18 -0800)]
Fix for ticket #2910: fix inconsistencies in notice posting response display that broke help command, could be generally wonky
Previous code was importing nodes from the XHR result into current document, then pulling text content of what might be the right element, then concat'ing that straight into HTML. Eww! Now pulling the text content straight from the XHR result -- same element that we check for existence of -- and using jQuery's own text() to do the getting and setting of text. Also note that some browsers might have been pulling HTML instead of text, or other funkiness.
Brion Vibber [Wed, 15 Dec 2010 22:57:09 +0000 (14:57 -0800)]
Fix for ticket #2942: character counter now updates on cut and paste operations made with mouse or menu
This uses the 'copy' and 'paste' DOM events to trigger a counter update. I haven't had a chance to 100% confirm that middle-button click on X11 triggers the event, but it ought to.
Cut and paste events from context menu and main edit menu known good in:
* Firefox 4.08b-pre
* IE 9 preview 7
* IE 8 current
* Chrome 8 beta current
* Safari 5.0.3
Opera is listed as not supporting these events, oh well.
Note that using a *delete* command from a menu doesn't trigger an event. Sigh, you can't win everything.
Brion Vibber [Wed, 15 Dec 2010 20:14:25 +0000 (12:14 -0800)]
Cleaner code to avoid a couple PHP notices from accessing uninitialized variables in ostatus profile discovery (these cases hit checking diaspora accounts)
Brion Vibber [Wed, 15 Dec 2010 00:14:15 +0000 (16:14 -0800)]
Mark OembedAction, XrdAction, and (plugin) AutocompleteAction as read-only. Tweaked ApiStatusesShow and ApiTimelineUser to still claim read-only when hit with a HEAD request (usually link checkers or a precursor to a GET, and should be semantically equivalent to a GET without actually transferring data)
Brion Vibber [Mon, 13 Dec 2010 20:12:22 +0000 (12:12 -0800)]
TwitterBridge: partial merge of id_str usage from 0.9.x for improved 32-bit and pre-5.2.10 compatibility. (on 64-bit in 5.2.6 we can pull the integer IDs, but silently lose some precision on the end.)
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.
Brion Vibber [Fri, 10 Dec 2010 22:08:36 +0000 (22:08 +0000)]
Workaround for locally-handled sessions breaking on PHP 5.3 with APC enabled.
Big thanks to the folks at http://pecl.php.net/bugs/bug.php?id=16745 for the secret juju!
Classes were being torn down before session save handlers got called at the end of the request, which exploded with complaints about being unable to find various classes.
Registering a shutdown function lets us explicitly close out the session before everything gets torn down.
Brion Vibber [Fri, 10 Dec 2010 22:08:36 +0000 (22:08 +0000)]
Workaround for locally-handled sessions breaking on PHP 5.3 with APC enabled.
Big thanks to the folks at http://pecl.php.net/bugs/bug.php?id=16745 for the secret juju!
Classes were being torn down before session save handlers got called at the end of the request, which exploded with complaints about being unable to find various classes.
Registering a shutdown function lets us explicitly close out the session before everything gets torn down.
Brion Vibber [Mon, 3 May 2010 23:49:59 +0000 (16:49 -0700)]
extlibs updates: PEAR::Mail to 1.2.0, PEAR::Net_SMTP to 1.4.2 (need to go together as a pair)
PEAR::Mail updated to 1.2.0 from 1.1.4, fixes deprecation warnings on PHP 5.3, as well as:
1.2.0:
• QA release - stable.
• Updated minimum dependencies (Net_SMTP, PEAR, PHP)
• Doc Bug #15620 Licence change to BSD
• Bug #13659 Mail parse error in special condition
• Bug #16200 - Security hole allow to read/write Arbitrary File
_hasUnclosedQuotes() doesn't properly handle a double slash before an end quote (slusarz@curecanti.org, Bug #9137).
• Make sure Net_SMTP is defined when calling getSMTPObject() directly (slusarz@curecanti.org, Bug #13772).
• Add addServiceExtensionParameter() to the SMTP driver (slusarz@curecanti.org, Bug #13764).
• Add a method to obtain the Net_SMTP object from the SMTP driver (slusarz@curecanti.org, Bug #13766).
PEAR::Net_SMTP updated to 1.4.2 from 1.3.1, needed to support updated PEAR::Mail:
1.4.2:
• Fixing header string quoting in data(). (Bug #17199)
1.4.1:
• The auth() method now includes an optional $tls parameter that determines whether or not TLS should be attempted (if supported by the PHP runtime and the remote SMTP server). This parameter defaults to true. (Bug #16349)
• Header data can be specified separately from message body data by passing it as the optional second parameter to ``data()``. This is especially useful when an open file resource is being used to supply message data because it allows header fields (like *Subject:*) to be built dynamically at runtime. (Request #17012)
1.4.0:
• The data() method now accepts either a string or a file resource containing the message data. (Request #16962)
1.3.4:
• All Net_Socket write failures are now recognized. (Bug #16831)
1.3.3:
• Added getGreeting(), for retrieving the server's greeting string. (Request #16066) [needed for PEAR::Mail]
• We no longer attempt a TLS connection if we're already using a secure socket. (Bug #16254)
• You can now specify a debug output handler via setDebug(). (Request #16420)
1.3.2:
• TLS connection only gets started if no AUTH methods are sent. (Bug #14944)
Brion Vibber [Mon, 3 May 2010 23:49:59 +0000 (16:49 -0700)]
extlibs updates: PEAR::Mail to 1.2.0, PEAR::Net_SMTP to 1.4.2 (need to go together as a pair)
PEAR::Mail updated to 1.2.0 from 1.1.4, fixes deprecation warnings on PHP 5.3, as well as:
1.2.0:
• QA release - stable.
• Updated minimum dependencies (Net_SMTP, PEAR, PHP)
• Doc Bug #15620 Licence change to BSD
• Bug #13659 Mail parse error in special condition
• Bug #16200 - Security hole allow to read/write Arbitrary File
_hasUnclosedQuotes() doesn't properly handle a double slash before an end quote (slusarz@curecanti.org, Bug #9137).
• Make sure Net_SMTP is defined when calling getSMTPObject() directly (slusarz@curecanti.org, Bug #13772).
• Add addServiceExtensionParameter() to the SMTP driver (slusarz@curecanti.org, Bug #13764).
• Add a method to obtain the Net_SMTP object from the SMTP driver (slusarz@curecanti.org, Bug #13766).
PEAR::Net_SMTP updated to 1.4.2 from 1.3.1, needed to support updated PEAR::Mail:
1.4.2:
• Fixing header string quoting in data(). (Bug #17199)
1.4.1:
• The auth() method now includes an optional $tls parameter that determines whether or not TLS should be attempted (if supported by the PHP runtime and the remote SMTP server). This parameter defaults to true. (Bug #16349)
• Header data can be specified separately from message body data by passing it as the optional second parameter to ``data()``. This is especially useful when an open file resource is being used to supply message data because it allows header fields (like *Subject:*) to be built dynamically at runtime. (Request #17012)
1.4.0:
• The data() method now accepts either a string or a file resource containing the message data. (Request #16962)
1.3.4:
• All Net_Socket write failures are now recognized. (Bug #16831)
1.3.3:
• Added getGreeting(), for retrieving the server's greeting string. (Request #16066) [needed for PEAR::Mail]
• We no longer attempt a TLS connection if we're already using a secure socket. (Bug #16254)
• You can now specify a debug output handler via setDebug(). (Request #16420)
1.3.2:
• TLS connection only gets started if no AUTH methods are sent. (Bug #14944)
Evan Prodromou [Thu, 9 Dec 2010 15:45:36 +0000 (10:45 -0500)]
Additional in-process cache plugin InProcessCache
Some of our caching systems, like the disk cache or memcached, have
significant overhead (network connections or disk I/O).
This plugin adds an additional layer of in-process cache, so we don't
need to reconnect to external cache systems when we've already
received a data item from the cache. There are some concurrency issues
here, but typically they won't be important at the level of a single
web hit.
Brion Vibber [Thu, 9 Dec 2010 01:39:04 +0000 (17:39 -0800)]
Tweak to PiwikAnalytics plugin to help browsers to pre-load piwik.js, may shave a little off load time.
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).
Brion Vibber [Wed, 8 Dec 2010 22:54:02 +0000 (14:54 -0800)]
Mapstraction plugin: use minified sources for OpenLayers
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.
Brion Vibber [Wed, 8 Dec 2010 22:52:26 +0000 (14:52 -0800)]
Create a bundled & minified JS file for Mapstraction's common case (using OpenLayers); this'll avoid waiting on additional script loads for mxn.core.js and mxn.openlayers.core.js, and removes the need to load usermap.js separately as well.
Brion Vibber [Wed, 8 Dec 2010 21:52:35 +0000 (13:52 -0800)]
Add stripped and minified local copy of OpenLayers 2.10, about 1/5 the size of the full version.
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: