Commits
-------
351910b added twig comments
2b7e8d9 added phpDocs for TokenParsers, filters and tests
Discussion
----------
PHPDocs updated
fixed copy/paste errors and added docs to Sandbox and Use tags
Commits
-------
e4b8371 Improve performance of is defined in non strict mode
Discussion
----------
Improve performance of is defined in non strict mode
Addresses issue #380: This gives a good performance improvement on defined tests in non-strict code. About 2x.
Reasoning: It doesn't make sense to propagate the `is_defined_test` flag into deeper levels in non-strict mode, because all getAttr functions and name accesses will just return `null` if it doesn't exist and `null` is just as good as `false`.
---------------------------------------------------------------------------
by nikic at 2011/07/02 03:22:53 -0700
Though, do not merge yet please. Changes behavior concerning methods.
---------------------------------------------------------------------------
by nikic at 2011/07/02 13:34:32 -0700
Hm, I really don't know what I was thinking when I wrote my last comment. There shouldn't be any behavior changes introduced by that change (and tests pass).
Commits
-------
5b91f12 Update tests
a5ef326 Use Template->getContext in non-strict mode too
Discussion
----------
Use Template->getContext() when in non-strict mode, too
As described in #380 using the current `isset($context['test']) ? $context['test'] : null` approach can cause problems in some cases: The ternary operator always returns by value and thus PHP's copy-on-write concept does not apply. So `$context['test']` needs to be copied in any case, even though a write never happens to it. Basically Twig is copying the values of all variables ever used inside Twig if it operates in non-strict mode. This is problematic when `$context['test']` contains big arrays as the whole array is copied.
In this patch I change Twig to use `Template->getContext` when in non-strict mode too and add an `isStrictVariables` check in there.
Commits
-------
75748a5 revert accidental mode change
681d2b8 fix nested buffers issue #374
Discussion
----------
fix nested buffers issue #374
see #374 for details
---------------------------------------------------------------------------
by fabpot at 2011/07/09 01:02:50 -0700
You have inadvertently changed the mode to 0755. Should stay at 0644.
---------------------------------------------------------------------------
by macolu at 2011/07/09 03:07:42 -0700
Mode change reverted.
About this Windows bug... Are you talking about the zlib issue? When zlib compression is activated, it adds a buffer level that ob_end_clean() can't close. So ob_get_level() never go below 1.
http://cksource.com/forums/viewtopic.php?t=21653http://www.php.net/manual/fr/function.ob-get-level.php#52945
In that case, the first ob_get_level() call will always return something greater than 0. So $level var will be > 0. This should avoid infinite loops.
Commits
-------
c178c28 Making docs more explict as to what kind of object can be used with Twig_Function_Method. Fixes#308
Discussion
----------
Fix for #308
Making docs more explict as to what kind of object can be used with Twig_Function_Method.
Fixes#308
Commits
-------
6593761 Allow setting of timezone on twig_date_format_filter method
Discussion
----------
Allow setting of timezone on twig_date_format_filter method
Commits
-------
f2a1c2b Refactor and add additional default filter tests
e760483 Make `a.b is defined` not throw an exception if a is not defined (in strict mode)
Discussion
----------
Make `a.b is defined` not throw an exception if a is not defined (in strict mode)
This commit does two things: Firstly (that's the main part) it makes `a.b is defined` not throw an error in strict mode if `a` is undefined (applies for `a.b.c` aso too) [fixes#337]. Additionally it allows `a.b|default` just like it is allowed to do `a|default`.
---------------------------------------------------------------------------
by fabpot at 2011/06/09 22:21:11 -0700
This is quite a big change. Can you add some tests?
---------------------------------------------------------------------------
by nikic at 2011/06/10 08:19:49 -0700
I added some more tests for the default filter in the fixtures. But I wasn't sure how I should test the default test's behavior in strict mode. If you could point me in the right direction here I would add those tests, too ;)
Jannis Grimm
Fabien Potencier
Robert Gruendler
Matthieu Robin
nikic
Mark Story
Josh Turmel