- 20 May, 2001 1 commit
-
-
Jim Fulton authored
intrspection.
-
- 18 May, 2001 10 commits
-
-
Guido van Rossum authored
At least this now passes Evan's test_metal1.html! I had to restructure attrAction() and the macroStack() quite a bit. The macro stack now has a mostly abstract interface (a few lookups still dig directly into the stack). Weekend!
-
Guido van Rossum authored
the test input filename contains the substring "metal".
-
Guido van Rossum authored
quite easy to accidentally set off an infinite recursion here.
-
Evan Simpson authored
-
Evan Simpson authored
-
Evan Simpson authored
-
Evan Simpson authored
-
Evan Simpson authored
-
Guido van Rossum authored
-
Guido van Rossum authored
-
- 17 May, 2001 9 commits
-
-
Evan Simpson authored
-
Shane Hathaway authored
-
Shane Hathaway authored
-
Shane Hathaway authored
-
Fred Drake authored
Remove unnecessary "global" statement -- already in global context. Python 2.1 complained about this because the global was assigned to before the global statement, all in a single scope.
-
Guido van Rossum authored
-
Guido van Rossum authored
-
Guido van Rossum authored
dict of the raw attribute values. Also, put beginScope/endScope around each tag for which a tagDict is emitted, rather than only for tags using define or repeat. This affects the generated code, but doesn't affect semantics (according to the tests).
-
Guido van Rossum authored
tag.
-
- 16 May, 2001 9 commits
-
-
Guido van Rossum authored
-
Guido van Rossum authored
- Define-slot can now be nested inside fill-slot, so that a macro can call another macro and use one of its slots to fill that other macro's slots. (This is essential for sensible semantics of nested macros.) - Inside a use-macro, if the macro defines a slot that's not filled, the define-slot turns into fill-slot. (This is just a convenience feature.) To implement this, the current macro call and its slots are turned into a stack (they already had stack behavior but the stack was inaccessible) and added a variable indicating the macro currently being defined (if any). In the process, refactored startTagCommon a bit, turning the decision procedure for argument replacement into a separate method (attrAction()). Added a safeguard to the attribute-wrapping feature: if the alignment column is too far to the right (half of the wrap column), change alignment to 4 spaces. This solves the problem reported by Gregor Hoffleit with attribute wrapping in a long line containing several elements.
-
matt@zope.com authored
that the issues remain to allowing unicode objects through Zope that are not addressed herin, but are related to type('') not equalling type(u''); much special case code exists that handles strings differently that needs review to handle unicode as well.
-
Evan Simpson authored
-
Evan Simpson authored
-
matt@zope.com authored
after changes applied; tutorial functional with gadfly.
-
Guido van Rossum authored
in setLocal() was improperly indented. Evan: please, Please, PLEASE, run the test suite before checking in changes!
-
Shane Hathaway authored
-
Shane Hathaway authored
-
- 15 May, 2001 1 commit
-
-
Evan Simpson authored
-
- 14 May, 2001 4 commits
-
-
Evan Simpson authored
__*attr__ magic methods don't work when you add them after the class is defined, so add them during class definition.
-
Evan Simpson authored
-
Shane Hathaway authored
-
Evan Simpson authored
-
- 11 May, 2001 6 commits
-
-
Evan Simpson authored
-
Evan Simpson authored
-
Evan Simpson authored
Implemented alternate path '|' operator. Added 'default' builtin. Made Expression types available to python expressions as functions. Changed the way render() calls DTML.
-
Shane Hathaway authored
-
Evan Simpson authored
-
Evan Simpson authored
-