Changes - google-caja - Compiler for making third-party HTML, CSS and JavaScript safe for embedding

4891 - 4867 of 4891 Older ›

Committed Changes

Rev	Commit log message	Date	Author
r4891	run junit in parallel http://codereview.appspot.com/6252050 after 'ant clean AllTests', running 'ant runtests' takes about 6 minutes on my laptop. junit tests are independent, they can be run in parallel. with this change, 'ant runtests' takes about 4.5 minutes R=metaweta@gmail.com	Yesterday (21 hours ago)	felix8a
r4890	shorten the top-level var warning http://codereview.appspot.com/6244054 When building caja, one of the warning messages starts [transform] LINT : css-defs.js:46+1 - 777+7: Top level variable declarations incompatiable with Caja. Please rewrite as window[ 'cssSchema' ] = (function () { and continues for 730 lines. There are several messages like this. The one for Domado is 4328 lines. This change omits the rhs of these warning messages. R=metaweta@gmail.com	Yesterday (32 hours ago)	felix8a
r4889	speed up junit http://codereview.appspot.com/6250057 We're running junit with fork="yes" because we want the junit vm to have different properties than the ant vm. However, we're spawning a new vm for each junit test, which is unnecessary. This adds forkmode="once" to the junit tests. This reduces 'ant runtests' on my laptop from about 11 minutes to about 7 minutes. R=metaweta@gmail.com	Yesterday (36 hours ago)	felix8a
r4888	replace some minified js with closured js http://codereview.appspot.com/6257051 This replaces caja-minified.js and html-sanitizer-minified.js with versions compiled using Closure's advanced optimizations option. This is non-scary because all the relevant tests run fine with the closured versions. R=metaweta@gmail.com	Yesterday (37 hours ago)	felix8a
r4887	Fix handling of 'target' attributes http://codereview.appspot.com/6240043 Fixes http://code.google.com/p/google-caja/issues/detail?id=1471 This change required passing an out-of-band "this attribute was unspecified by the guest programmer" signal from TemplateCompiler to HtmlAttributeRewriter. See changes for details of how this is done. R=felix8a@gmail.com	May 23 (2 days ago)	ihab.awad
r4886	Adds special cases to the taming membrane for builtins. http://codereview.appspot.com/6241043 Adds special cases for Boolean, Date, Number, RegExp, and String to the taming membrane. Also whitelists toString and valueOf on Boolean.prototype. (No one ever noticed they were missing.) R=felix8a, ihab.awad	May 23 (2 days ago)	metaweta
r4885	publish r4884 for maven http://codereview.appspot.com/6213063 r4884 fixes a bug with tamePreviouslyConstructedObject, which affects how trusted container code passes objects to untrusted guest code.. R=metaweta@gmail.com	May 22 (3 days ago)	felix8a
r4884	Fix typo in taming membrane http://codereview.appspot.com/6218061 Fix typo in taming membrane as found by Felix. Add one test. R=felix8a@gmail.com	May 22 (3 days ago)	ihab.awad
r4883	publish r4882 for maven http://codereview.appspot.com/6208092 The previously published version of caja was either 4527 (2011-06-27) or 4504 (2001-06-14) depending on how you ask. So, this is about a year of changes. Highlights: - Improved ES5/SES mode - Replaced domita with domado - Support client-side sanitization of stylesheets - Better caja.js api - Stricter caja version checking - Better support for jquery and jquery-ui - Better support for Flash embeds - Better compatibility with Goog	May 21 (4 days ago)	felix8a
r4882	expose tamesTo http://codereview.appspot.com/6227049 shindig needs tamesTo, which isn't conveniently exposed in the same way tame and untame are. This patch exposes tamesTo. (shindig uses tamesTo to redirect gadgets.json functions to caja's special JSON functions.) R=ihab.awad@gmail.com	May 21 (4 days ago)	felix8a
r4881	Repair error message reporting in ES5 http://codereview.appspot.com/6214062 * Separate "compile" errors from "runtime" errors thrown my scripts * Corrected value of onerror that was being looked up to report the error R=metaweta@gmail.com	May 20 (5 days ago)	jasvir
r4880	Run jquery tests under ES5 http://codereview.appspot.com/6208082 Sets up the running of jQuery tests under both es5 and es53 Moves the directory in which jQuery tests get hosted so relative paths work correctly. Recognize link types without case sensitivity. R=metaweta@gmail.com	May 18, 2012	jasvir
r4879	Support external scripts and stylesheets loading in ES5 http://codereview.appspot.com/6060046 * Adds JSONP support to cajoling service to proxy a url and return it as a json object * Modifies caja.js to provide a default jsonp based fetcher for external urls using doc.write * Modifies html-sanitizer to make the sax parser restartable in some places (eg. at the end of a tag) * Modifies html-emitter to stop processing tokens until content is returned by fetcher and sandboxing the fetched content	May 17, 2012	jasvir
r4878	small caja.js tweaks http://codereview.appspot.com/6214051 1. closurization inadvertently changed a test typeof window.console.log === 'function' into typeof window['console'] === 'function' which is not the same thing. this fixes that. 2. when debug==true, add ?debug=1 to the frame scripts, which will signal to shindig to not compile/minify the scripts. R=metaweta@gmail.com	May 16, 2012	felix8a
r4877	more precajoler stuff http://codereview.appspot.com/6201082 - Add a servlet /precajole/info that tells you what's precajoled. There's also a link to it in the upper right of the playground. - Add the latest version of jquery to the precajoler. - Add the latest version of jquery-ui to the precajoler. R=ihab.awad@gmail.com	May 15, 2012	felix8a
r4876	fix handling of naive uri policies in server-side rewriter http://codereview.appspot.com/6209054 a little while ago we fixed the client-side uri rewriter so that when a container supplies a naive URI policy, we still block unsafe URIs like javascript: this fixes it for the server-side uri rewriter too. R=metaweta@gmail.com	May 11, 2012	felix8a
r4875	Removes unnecessary uses of slice(arguments, 0) http://codereview.appspot.com/6203062 Removes some unnecessary uses of slice(arguments, 0) and adds notes when it's actually necessary. This is usually when the fake arguments object is given directly to the apply method, where only arrays and real arguments objects are allowed. R=erights@google.com, felix8a	May 9, 2012	metaweta
r4874	Add more hints to URI policy, and improve "target" attribute handling http://codereview.appspot.com/6182049 * Consistently send more "hints" to the container's URI policy upon invocation. * Modify "target" attribute handling so that the server side cajoler always produces markup with target="_blank" then always consults a default and whitelist supplied by the container on the client side for rewriting. R=felix8a@gmail.com	May 9, 2012	ihab.awad
r4873	Fixes 1362, the communications channel in Date.prototype. http://codereview.appspot.com/6188056 This change wraps all setter methods of Date so that they throw if invoked on the prototype object itself. R=felix8a, ihab.awad	May 8, 2012	metaweta
r4872	Warnings ES5/3 developers of code that will not be supported under ES5 http://codereview.appspot.com/6196049 ES5 does not handle communication between script blocks via top-level var and function declarations. These need to be rewritten as: var x = 1; => window['x'] = 1; function y() {} => window['y'] = function() {} This change warns ES5/3 developers of this need. R=metaweta@gmail.com	May 8, 2012	jasvir
r4871	allow caller to supply both cajoled and uncajoled code http://codereview.appspot.com/6035056 In some cases, a server like shindig has both the cajoled and uncajoled version of some guest code, and on the server side we don't know if the browser will support ses or not. So we want to send to the browser code that will create a guest frame and then do frame.code(...).cajoled(...).run() so when ses is active, the code(...) will be used, but when es53 is active, the cajoled(...) will be used.	May 8, 2012	felix8a
r4870	test closure-ized caja.js http://codereview.appspot.com/6131063 This runs closure-ized caja.js through our browser test machinery, to verify that it works (incidentally fixing some problems that prevented it from working.) R=metaweta@gmail.com	May 8, 2012	felix8a
r4869	misc changes to browser test framework http://codereview.appspot.com/6161046 - build.xml has logic to remove generated files that have version in them, so that incremental builds work correctly if your svn version changes. but it wasn't removing all generated files. this patch fixes that. - GeneralBrowserTest takes several minutes to run, and it isn't easy to see what parts are slow. this patch adds a print before each browser test page, to make it easier to track test speed. - in cases whe	May 8, 2012	felix8a
r4868	minor playground UI changes http://codereview.appspot.com/6180045 1. build.xml has a rule to check if the playground needs to be rebuilt. The rule doesn't look at all relevant files. 2. tweak button layout. R=jasvir@gmail.com	May 3, 2012	felix8a
r4867	Created wiki page through web user interface.	May 2, 2012	ihab.awad

4891 - 4867 of 4891 Older ›