6 (codename to be selected, release date to be announced)
8 - internal attributes now raise an internal attribute error now instead
9 of returning an undefined. This fixes problems when passing undefined
10 objects to Python semantics expecting APIs.
11 - traceback support now works properly for PyPy. (Tested with 1.4)
12 - implemented operator intercepting for sandboxed environments. This
13 allows application developers to disable builtin operators for better
14 security. (For instance limit the mathematical operators to actual
15 integers instead of longs)
16 - groupby filter now supports dotted notation for grouping by attributes
18 - scoped blocks not properly treat toplevel assignments and imports.
19 Previously an import suddenly "disappeared" in a scoped block.
20 - automatically detect newer Python interpreter versions before loading code
21 from bytecode caches to prevent segfaults on invalid opcodes. The segfault
22 in earlier Jinja2 versions here was not a Jinja2 bug but a limitation in
23 the underlying Python interpreter. If you notice Jinja2 segfaulting in
24 earlier versions after an upgrade of the Python interpreter you don't have
25 to upgrade, it's enough to flush the bytecode cache. This just no longer
26 makes this necessary, Jinja2 will automatically detect these cases now.
27 - the sum filter can now sum up values by attribute. This is a backwards
28 incompatible change. The argument to the filter previously was the
29 optional starting index which defaultes to zero. This now became the
30 second argument to the function because it's rarely used.
31 - like sum, sort now also makes it possible to order items by attribute.
35 (re-release of 2.5.4 with built documentation removed for filesize.
36 Released on October 18th 2010)
38 - built documentation is no longer part of release.
42 (bugfix release, released on October 17th 2010)
44 - Fixed extensions not loading properly with overlays.
45 - Work around a bug in cpython for the debugger that causes segfaults
46 on 64bit big-endian architectures.
50 (bugfix release, released on October 17th 2010)
52 - fixed an operator precedence error introduced in 2.5.2. Statements
53 like "-foo.bar" had their implicit parentheses applied around the
54 first part of the expression ("(-foo).bar") instead of the more
59 (bugfix release, released on August 18th 2010)
61 - improved setup.py script to better work with assumptions people
62 might still have from it (``--with-speedups``).
63 - fixed a packaging error that excluded the new debug support.
67 (bugfix release, released on August 17th 2010)
69 - StopIteration exceptions raised by functions called from templates
70 are now intercepted and converted to undefineds. This solves a
71 lot of debugging grief. (StopIteration is used internally to
72 abort template execution)
73 - improved performance of macro calls slightly.
74 - babel extraction can now properly extract newstyle gettext calls.
75 - using the variable `num` in newstyle gettext for something else
76 than the pluralize count will no longer raise a :exc:`KeyError`.
77 - removed builtin markup class and switched to markupsafe. For backwards
78 compatibility the pure Python implementation still exists but is
79 pulled from markupsafe by the Jinja2 developers. The debug support
80 went into a separate feature called "debugsupport" and is disabled
81 by default because it is only relevant for Python 2.4
82 - fixed an issue with unary operators having the wrong precendence.
86 (codename Incoherence, relased on May 29th 2010)
88 - improved the sort filter (should have worked like this for a
89 long time) by adding support for case insensitive searches.
90 - fixed a bug for getattribute constant folding.
91 - support for newstyle gettext translations which result in a
92 nicer in-template user interface and more consistent
93 catalogs. (:ref:`newstyle-gettext`)
94 - it's now possible to register extensions after an environment
99 (bugfix release, released on April 20th 2010)
101 - fixed an error reporting bug for undefineds.
105 (codename Correlation, released on April 13th 2010)
107 - the environment template loading functions now transparently
108 pass through a template object if it was passed to it. This
109 makes it possible to import or extend from a template object
110 that was passed to the template.
111 - added a :class:`ModuleLoader` that can load templates from
112 precompiled sources. The environment now features a method
113 to compile the templates from a configured loader into a zip
115 - the _speedups C extension now supports Python 3.
116 - added support for autoescaping toggling sections and support
117 for evaluation contexts (:ref:`eval-context`).
118 - extensions have a priority now.
122 (bugfix release, released on February 19th 2010)
124 - fixed an error reporting bug on all python versions
125 - fixed an error reporting bug on Python 2.4
129 (3000 Pythons, released on February 10th 2010)
131 - fixes issue with code generator that causes unbound variables
132 to be generated if set was used in if-blocks and other small
134 - include tags are now able to select between multiple templates
135 and take the first that exists, if a list of templates is
137 - fixed a problem with having call blocks in outer scopes that
138 have an argument that is also used as local variable in an
140 - greatly improved error message reporting (#339)
141 - implicit tuple expressions can no longer be totally empty.
142 This change makes ``{% if %}...{% endif %}`` a syntax error
144 - added support for translator comments if extracted via babel.
145 - added with-statement extension.
146 - experimental Python 3 support.
150 (bugfix release, released on September 14th 2009)
152 - fixes some smaller problems for Jinja2 on Jython.
156 (codename Kong, released on September 13th 2009)
158 - Include statements can now be marked with ``ignore missing`` to skip
159 non existing templates.
160 - Priority of `not` raised. It's now possible to write `not foo in bar`
161 as an alias to `foo not in bar` like in python. Previously the grammar
162 required parentheses (`not (foo in bar)`) which was odd.
163 - Fixed a bug that caused syntax errors when defining macros or using the
164 `{% call %}` tag inside loops.
165 - Fixed a bug in the parser that made ``{{ foo[1, 2] }}`` impossible.
166 - Made it possible to refer to names from outer scopes in included templates
167 that were unused in the callers frame (#327)
168 - Fixed a bug that caused internal errors if names where used as iteration
169 variable and regular variable *after* the loop if that variable was unused
170 *before* the loop. (#331)
171 - Added support for optional `scoped` modifier to blocks.
172 - Added support for line-comments.
173 - Added the `meta` module.
174 - Renamed (undocumented) attribute "overlay" to "overlayed" on the
175 environment because it was clashing with a method of the same name.
176 - speedup extension is now disabled by default.
182 - Fixed a translation error caused by looping over empty recursive loops.
186 (codename Yasuzō, released on November 23rd 2008)
188 - fixed a bug with nested loops and the special loop variable. Before the
189 change an inner loop overwrote the loop variable from the outer one after
192 - fixed a bug with the i18n extension that caused the explicit pluralization
193 block to look up the wrong variable.
195 - fixed a limitation in the lexer that made ``{{ foo.0.0 }}`` impossible.
197 - index based subscribing of variables with a constant value returns an
198 undefined object now instead of raising an index error. This was a bug
199 caused by eager optimizing.
201 - the i18n extension looks up `foo.ugettext` now followed by `foo.gettext`
202 if an translations object is installed. This makes dealing with custom
203 translations classes easier.
205 - fixed a confusing behavior with conditional extending. loops were partially
206 executed under some conditions even though they were not part of a visible
209 - added `sort` filter that works like `dictsort` but for arbitrary sequences.
211 - fixed a bug with empty statements in macros.
213 - implemented a bytecode cache system. (:ref:`bytecode-cache`)
215 - the template context is now weakref-able
217 - inclusions and imports "with context" forward all variables now, not only
220 - added a cycle helper called `cycler`.
222 - added a joining helper called `joiner`.
224 - added a `compile_expression` method to the environment that allows compiling
225 of Jinja expressions into callable Python objects.
227 - fixed an escaping bug in urlize
231 (codename jinjavitus, released on July 17th 2008)
233 - the subscribing of objects (looking up attributes and items) changed from
234 slightly. It's now possible to give attributes or items a higher priority
235 by either using dot-notation lookup or the bracket syntax. This also
236 changed the AST slightly. `Subscript` is gone and was replaced with
237 :class:`~jinja2.nodes.Getitem` and :class:`~jinja2.nodes.Getattr`.
239 For more information see :ref:`the implementation details <notes-on-subscriptions>`.
241 - added support for preprocessing and token stream filtering for extensions.
242 This would allow extensions to allow simplified gettext calls in template
243 data and something similar.
245 - added :meth:`jinja2.environment.TemplateStream.dump`.
247 - added missing support for implicit string literal concatenation.
248 ``{{ "foo" "bar" }}`` is equivalent to ``{{ "foobar" }}``
250 - `else` is optional for conditional expressions. If not given it evaluates
253 - improved error reporting for undefined values by providing a position.
255 - `filesizeformat` filter uses decimal prefixes now per default and can be
256 set to binary mode with the second parameter.
258 - fixed bug in finalizer
262 (no codename, released on June 9th 2008)
264 - first release of Jinja2