Specification, released in December 2010.
Substantive changes to the language specification since 1.1:
px:doctype constants,
reflecting availability in current implementation.
Specification, released in June 2004.
Substantive changes to the language specification since 1.0:
None, False and
True which would make them behave in
unexpected ways. For instance False= 0 is
okay but False= "hello" is not. This allows
PXTL implementations to perform some performance optimisations,
pretending these three identifiers are in fact literals (which they might
become in some future version of Python).
Third draft of PXTL, released in July 2003.
Substantive changes to the language specification since beta specification:
px_ PI to px with
an optional colon beginning the data to indicate a code block. This is
necessary as DOM says any whitespace between the target name and data is
not provided to the application (so it's not possible to find out the
indentation level of a block without the leading colon).
px_mark and px_code
pseudo-PI; they can now only be used in a PI in element content, not pseudo-PI
in attribute values. This is for compatibility with DOMs (esp. minidom) that
do not fully implement attribute childNodes.
px_name encoding.
future attribute, principally for controlling
nested scopes.
call element as the new way to call
subtemplates, rather than a function call, as subtemplates should only
be callable in element content.
import src attribute to be a URI instead of
local filepath; this ties in nicely with the new DOM 3 stuff.
global directive,
which as a not-a-proper-statement can't generally be emulated.
pxtl.write to inside
px_code PIs; it is not sensible to have
side-effect output in a PI that itself produces output, and
would cause extra overhead.
Second draft of the specification, released January 2003 for general public discussion.
Substantive changes to the spec since first draft:
onjssl PI as it’s possible to encode
equally well for script blocks and event handlers with the same
algorithm. So jssl (renamed jstr)
now does both;
cstr PI on the off-chance anyone wants
to write dynamic stylesheets;
par, markup
and comment PIs on the somewhat dubious
grounds that having four letters is nice;
orif element and attribute as an analogue
to anif. It remains to be seen whether this
will actually be as useful as anif;
switch and
when elements, since Python programmers are
probably used to using elif for this sort of task;
while element as it’s a Python flow control
statement that could conceivably be useful, especially with the post-test
option;
include element with the more Pythonic
import;
var attribute in for
and def as it doesn’t read very well in
def and import.
Unfinished first draft of PXTL was released for limited comment July 2002.