Kein-Hong Man wrote:
Hi all,
herbert breunung wrote:
[snip]
yes i have some interest so help out because the perl lexer ist most
important to me since
my editor (proton-ce.sf.net) is primarely for perl monks ans web
developers. you have an homepage with some informations
about your progress with the lexer?
To summarize my plan of action for the list:
- I will make some time for LexPerl in Q3, keep the list informed
of progress (if any), and work with others who are interested in
fixing LexPerl.
- Ironing out all the glitches will require some serious study of
the Perl parser. Will probably be stuck with this for a while.
- If a simplified solution can fit into the current codebase, a
fix can be coded quickly. Such a solution may not be foolproof or
futureproof, but a temporary fix may be useful.
- If the solution is non-trivial, then parsing is the probable
requirement and it might be better to go work on Eric's codebase.
HTH,
I might as well speak up -- we're still using the codebase you mentioned
in several different situations. If I could have just hooked the Perl
parser into the lexer framework, I would have long ago, but that doesn't
work for editing buffers, which usually do not contain a valid Perl string.
You absolutely need to recover from syntax errors as quickly as possible,
and not be too clever. If this means the user has forgotten to close a
quote, the rest of the buffer should be colored to reflect that.
Kein-Hong has covered the most troublesome parts, which come down to the
use of barewords and ambiguous operators (qw@/ ? <<@ are the main culprits).
I'm not sure it's possible to craft a 100% accurate Perl 5 lexer. When a user
reports a bug, there's always a been a workaround, that actually makes the code
more self-explanatory.
- Eric
_______________________________________________
Scintilla-interest mailing list
[email protected]
http://mailman.lyra.org/mailman/listinfo/scintilla-interest
