Re: SVN: Zope/branches/ajung-zpt-end-game/lib/python/Products/PageTemplates/tests/testExpressions.py Fix a test that was testing wrong behaviour: primitive types
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Philipp von Weitershausen wrote:
Log message for revision 68244: Fix a test that was testing wrong behaviour: primitive types werent' called if they were the result of a path expression, while all other callables were called. This tests now tests the correct behaviour (which happens to be already implemented by the Zope 3 TALES engine)
+ assert ec.evaluate('x | python:int') == 0
Wait a minute -- that isn't a path expression there! The '|' is not supposed to change the rules for evaluation of the subexpressions! If 'int' or the like were somehow avaialbe at module scope, or via 'options' or another top-level name, then they should be called *when they are the last item in a path expression*. The test was actually correct beofore: because tne PathExpression, 'x', raises, the PythonExpression, 'int', is evalutated. The result of a subexpression is not suppoeed to be called,
assert ec.evaluate('x | string:x') == 'x' assert ec.evaluate('x | string:$one') == '1' assert ec.evaluate('x | not:exists:x')
- -- =================================================================== Tres Seaver +1 202-558-7113 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFEchs6+gerLs4ltQ4RAi9DAJ9X34bgQNfCq+RN4zj27OoU6/Et0QCfbEDn IYaH04TINHsOQk36f2f19do= =eW90 -----END PGP SIGNATURE-----
Tres Seaver wrote:
Philipp von Weitershausen wrote:
Log message for revision 68244: Fix a test that was testing wrong behaviour: primitive types werent' called if they were the result of a path expression, while all other callables were called. This tests now tests the correct behaviour (which happens to be already implemented by the Zope 3 TALES engine)
+ assert ec.evaluate('x | python:int') == 0
Wait a minute -- that isn't a path expression there!
Yes it is. This line could explicitly written as: assert ec.evaluate('path: x | python:int') == 0 The python: expression is just part of the path expression.
The '|' is not supposed to change the rules for evaluation of the subexpressions!
It doesn't change the rules for the evaluation of the subexpressions themselves. But it might easily change the rules of what happens to the result of the expression afterwards.
If 'int' or the like were somehow avaialbe at module scope, or via 'options' or another top-level name, then they should be called *when they are the last item in a path expression*.
The test was actually correct beofore: because tne PathExpression, 'x', raises, the PythonExpression, 'int', is evalutated. The result of a subexpression is not suppoeed to be called,
No, not the result of a subexpression, but the result of the overall path expression. That's how it works in both Zope 2 and Zope 3. Zope 2 just special-cases primitive types. Philipp
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Philipp von Weitershausen wrote:
Tres Seaver wrote:
Philipp von Weitershausen wrote:
Log message for revision 68244: Fix a test that was testing wrong behaviour: primitive types werent' called if they were the result of a path expression, while all other callables were called. This tests now tests the correct behaviour (which happens to be already implemented by the Zope 3 TALES engine)
+ assert ec.evaluate('x | python:int') == 0
Wait a minute -- that isn't a path expression there!
Yes it is. This line could explicitly written as:
assert ec.evaluate('path: x | python:int') == 0
The python: expression is just part of the path expression.
The '|' is not supposed to change the rules for evaluation of the subexpressions!
It doesn't change the rules for the evaluation of the subexpressions themselves. But it might easily change the rules of what happens to the result of the expression afterwards.
If 'int' or the like were somehow avaialbe at module scope, or via 'options' or another top-level name, then they should be called *when they are the last item in a path expression*.
The test was actually correct beofore: because tne PathExpression, 'x', raises, the PythonExpression, 'int', is evalutated. The result of a subexpression is not suppoeed to be called,
No, not the result of a subexpression, but the result of the overall path expression. That's how it works in both Zope 2 and Zope 3. Zope 2 just special-cases primitive types.
If a subexpression returns a callable, you're telling me that you think the "top-level" expression is supposed to call it? That's nuts! The spec[1] says:
When a TALES path expression is evaluated, it attempts to traverse each path, from left to right, until it succeeds or runs out of paths. To traverse a path, it first fetches the object stored in the variable. For each path segment, it traverses from the current object to the subobject named by the path segment.
Once a path has been successfully traversed, the resulting object is the value of the expression. If it is a callable object, such as a method or class, it is called. The semantics of traversal (and what it means to be callable) are implementation-dependent.
If a traversal step fails, evaluation immediately proceeds to the next path. If there are no further paths, an error results.
Calling the callable happens *before* attempting alternate expressions, not after. That is partly because the *call* might be what triggers the alternation. [1] http://www.zope.org/Wikis/DevSite/Projects/ZPT/TALES%20Specification%201.3 Tres. - -- =================================================================== Tres Seaver +1 202-558-7113 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFEciCT+gerLs4ltQ4RAhesAKDHpp1+XJRukZfRgp/wBrDyBA3/HACfRj4P k20UEFCRTzu++Bcyp015JDE= =nmpz -----END PGP SIGNATURE----- -- =================================================================== Tres Seaver +1 202-558-7113 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com
Tres Seaver wrote:
If a subexpression returns a callable, you're telling me that you think the "top-level" expression is supposed to call it? That's nuts! The spec[1] says:
When a TALES path expression is evaluated, it attempts to traverse each path, from left to right, until it succeeds or runs out of paths. To traverse a path, it first fetches the object stored in the variable. For each path segment, it traverses from the current object to the subobject named by the path segment.
Once a path has been successfully traversed, the resulting object is the value of the expression. If it is a callable object, such as a method or class, it is called. The semantics of traversal (and what it means to be callable) are implementation-dependent.
If a traversal step fails, evaluation immediately proceeds to the next path. If there are no further paths, an error results.
Calling the callable happens *before* attempting alternate expressions, not after. That is partly because the *call* might be what triggers the alternation.
[1] http://www.zope.org/Wikis/DevSite/Projects/ZPT/TALES%20Specification%201.3
You're absolutely right. I filed a collector issue for Zope 3: http://www.zope.org/Collectors/Zope3-dev/638 This will be fixed soon in Zope 3. Philipp
participants (2)
-
Philipp von Weitershausen -
Tres Seaver