Re: cperl-mode and fontlock

[Ilya Zakharevich]

On 2009-09-28, Ted Zlatanov <tzz@lifelogs.com> wrote:
> Ilya, it looks like Chong Yidong has synchronized 6.2 into the Emacs
> trunk.

How much does it differ from what was done a decade ago?  That time
a brave guy volunteered and converted cperl-mode to new conventions.
Only years of user experiences had shown that this introduced so many
bugs that the mode is IMO unusable.

  (Usability is the question of trust.  THE purpose of a mode is to
   use faces/paren-match/indentation to make driver errors as much
   visible as possible.  I know that "my" version has so few false
   positives that EVERY irregularity of
   highlight/paren-match/indentation deserves to be investigated.
   AFAIK, the situation is opposite with the shipped version; the "cry
   wolves" effect...)

> Can you follow up on emacs-devel and clear up any remaining issues?

Could not find it on gmane.emacs.devel...

> I'd like to make sure the Emacs version of cperl-mode doesn't
> get out of sync again, and it's good that Chong (one of the two current
> Emacs maintainers) feels strongly enough to synchronize against your
> version.

The only practically viable solution I can see is to aim for identical
byte-code between "my" version and Emacs's one...  (At least with some
"define"; the default compile may aim for distinct bytecode, but all
the corresponding places in the .el file should be easily identifyable...)

Thanks,
Ilya