hi,
i am just looking through the CHANGES file to find out what all the new stuff is and try it out.
- the getter-setter example fails with:
getset.pike:4:Illegal ` identifier. getset.pike:4:parse error, unexpected `TOK_IDENTIFIER', expecting `error' or `'('' getset.pike:7:parse error, unexpected `TOK_RETURN' getset.pike:10:Illegal ` identifier. getset.pike:10:parse error, unexpected `TOK_IDENTIFIER', expecting `error' or `'('' getset.pike:13:parse error, unexpected `'='' getset.pike:14:parse error, unexpected `'}'' getset.pike:14:Missing ';'. getset.pike:20:parse error, unexpected `'}'' getset.pike:20:Missing ';'. getset.pike:26:parse error, unexpected `';'', expecting `')'' getset.pike:26:Missing ')'. getset.pike:27:Missing ';'. getset.pike:29:Missing '}'. getset.pike:29:Unexpected end of file.
it looks almost like it is not working at all. (i am using the pikefarm build to test this) is there a switch that needs to be turned to activate this?
- does portable bytecode affect the performance? did the dumped bytecode contain architecture specific optimizations before? or were those always generated at runtime? in other words, where does the machinecode generation come in?
- what is the missing example in ADT.Struct supposed to demonstrate?
- what do all these FIXME: Rewrite comments mean? rewrite the code or rewrite the CHANGES entry?
greetings, martin.
- the getter-setter example fails with:
getset.pike:4:Illegal ` identifier. getset.pike:4:parse error, unexpected `TOK_IDENTIFIER', expecting `error' or `'('' getset.pike:7:parse error, unexpected `TOK_RETURN' getset.pike:10:Illegal ` identifier. getset.pike:10:parse error, unexpected `TOK_IDENTIFIER', expecting `error' or `'('' getset.pike:13:parse error, unexpected `'='' getset.pike:14:parse error, unexpected `'}'' getset.pike:14:Missing ';'. getset.pike:20:parse error, unexpected `'}'' getset.pike:20:Missing ';'. getset.pike:26:parse error, unexpected `';'', expecting `')'' getset.pike:26:Missing ')'. getset.pike:27:Missing ';'. getset.pike:29:Missing '}'. getset.pike:29:Unexpected end of file.
Right, seems someone updated the example to the new syntax without checking that the new syntax actually worked... There were some bugs in the lexer. Fixed.
- does portable bytecode affect the performance?
did the dumped bytecode contain architecture specific optimizations before? or were those always generated at runtime? in other words, where does the machinecode generation come in?
Portable bytecode causes the intermediate bytecode to be retained (and dumped). Machinecode is generated in the step after that.
- what is the missing example in ADT.Struct supposed to demonstrate?
Missing?
- what do all these FIXME: Rewrite comments mean?
rewrite the code or rewrite the CHANGES entry?
They mean that the CHANGES file ought to be fixed in that regard.
greetings, martin.
On Fri, Apr 18, 2008 at 10:35:12PM +0200, Henrik Grubbstr�m (Lysator) @ Pike (-) developers forum wrote:
Right, seems someone updated the example to the new syntax without checking that the new syntax actually worked... There were some bugs in the lexer. Fixed.
great!
Portable bytecode causes the intermediate bytecode to be retained (and dumped). Machinecode is generated in the step after that.
is the machinecode dumped as well?
- what is the missing example in ADT.Struct supposed to demonstrate?
Missing?
the ADT.Struct section in CHANGES has an FIXME: Example comment. i am trying to think of an example that could fit there. but i don't know what to focus on.
greetings, martin.
- what is the missing example in ADT.Struct supposed to demonstrate?
Missing?
the ADT.Struct section in CHANGES has an FIXME: Example comment. i am trying to think of an example that could fit there. but i don't know what to focus on.
The idea was to have an example of a Struct object using another Struct object as an Item.
Perhaps a small/partial IFF codec? (That should give opportunity to show how a containing frame (FORM) length field depends on both the lengths of its own elements and those of its sub-frames, anyway.)
pike-devel@lists.lysator.liu.se
-
Henrik Grubbstr�m (Lysator) ＠ Pike (-) developers forum -
Johan Sundstr�m (Achtung Liebe!) ＠ Pike (-) developers forum
-
Martin B�hr -
Martin Nilsson (Opera Mini - AFK!) ＠ Pike (-) developers forum