* Matthew Toseland <toad at amphibian.dyndns.org> [2007-08-21 23:27:49]:
> On Monday 13 August 2007 21:59, you wrote:
> > Author: nextgens
> > Date: 2007-08-13 20:59:10 +0000 (Mon, 13 Aug 2007)
> > New Revision: 14664
> >
> > Removed:
> > trunk/freenet/src/freenet/clients/http/filter/CSSTokenizerFilter.java
> > Modified:
> > trunk/freenet/src/freenet/clients/http/filter/CSSTokenizerFilter.jflex
> > Log:
> > Fix last commit, unbreak the CSS filter (hopefully).
> >
> > REVIEW IT!
> >
> > Modified:
> trunk/freenet/src/freenet/clients/http/filter/CSSTokenizerFilter.jflex
> > ===================================================================
> > --- trunk/freenet/src/freenet/clients/http/filter/CSSTokenizerFilter.jflex
> 2007-08-13 20:23:07 UTC (rev 14663)
> > +++ trunk/freenet/src/freenet/clients/http/filter/CSSTokenizerFilter.jflex
> 2007-08-13 20:59:10 UTC (rev 14664)
> > @@ -237,9 +237,9 @@
> > STRING2=\'(\\{NL}|\"|(\\\')|{NONASCII}|{ESCAPE}|[^\'])*\'
> >
> > IDENT={NMSTART}{NMCHAR}*
> > -UNOFFICIAL_IDENT="-[^0-9]"{IDENT}
> > +UNOFFICIAL_IDENT="-"{IDENT}
>
> Ok, this is just reverting it to the old version.
>
> > NAME={NMCHAR}+
> > -NUM="-"([0-9]+|[0-9]*"."[0-9]+)
> > +NUM=[-]([0-9]+|[0-9]*"."[0-9]+)
>
> Need more brackets, no?
>
> (([0-8]+)|([0-9]*"."[0-9]+)) ?
>
> Also I don't get the [-].
I changed it again in a later version : current is
NUM=(-){0,1}([0-9]+|[0-9]*"."[0-9]+)
> > +{UNOFFICIAL_IDENT} {
> > + if(debug) log("Deleted unofficial ident: "+yytext());
> > + w.write("/* " + l10n("deletedUnofficialIdent") + " */");
> > +}
> > +{UNOFFICIAL_IDENT}{W}":"{W}{REALURL} {
> > + if(debug) log("Deleted unofficial ident with url: "+yytext());
> > + w.write("/* " + l10n("deletedUnofficialIdentWithURL") + " */");
> > +}
>
> Moving the unofficial ident matching down seems sensible although *it has no
> effect at all* - jflex always takes the longest match. The last rule is the
> fallback.
Weird... I moved it because it matched things it shouldn't have.
> Lexical states look quite interesting, maybe a way to do more
> sophisticated parsing rather than simple sequential lexing which really
> doesn't work that well for CSS.
Agreed: the current solution sucks.
NextGen$
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL:
<https://emu.freenetproject.org/pipermail/devl/attachments/20070822/6e7e354b/attachment.pgp>
