Hi! On Tue, Oct 20, 2020 at 6:42 AM 積丹尼 Dan Jacobson wrote: > > >>>>> "LI" == Lars Ingebrigtsen writes: > LI> 積丹尼 Dan Jacobson writes: > > >> OK but that doesn't explain the inconsistencies within the same language seen. > > LI> The $ is the same colour in Perl mode, so you'll have to be more > LI> specific about what inconsistencies you're referring to. > > All I know is in perl mode, > $$, $@, $bla > BB, BB, BTTT (Black, Tan) > > So it should be > BT, BT, BTTT like Shell-script mode, > or > BB, BB, BBBB > or > TT, TT, TTTT. I think what we have now makes sense. I've tried to attach a screenshot of a -Q session but with (setq cperl-hairy t) In essence, what I think I see is that cperl will color the whole array or hash variable including the @ or % sigil when it used as such, however, when a sigil is used to dereference only sigils is colored for the hash or array - the scalar being derefered is colored per scalar (e.g. tan). This is perhaps the most immediately obvious difference between perl and cperl-mode and probably a big part of why I think that cperl provides much better visual feedback than perl-mode. (As an aside, I've long wondered if there are people who perl-mode to cperl-mode and if disagreement about this drives that preference.) One thing I could wish for is that scalars used within hash and array access would be colored just as they are otherwise, e.g. using the tan color for the $scalar in $arr[$scalar] and $hash{$scalar}. Otherwise I'm pretty happy; I've not found a better Perl editing experience than Emacs cperl-mode. Thanks for your work on this feature! Corwin