la 09.06.2005 21:39 Glenn Morris skribis:
Daniel Pfeiffer wrote:
Like I said, I'll solve this more intelligently.
I really don't think this is the time to be doing this. Emacs is
supposed to be a feature freeze state so it can be released sometime.
Admittedly, it's not clear how anyone is supposed to recognize this...
Nothing fancy, just statements like ifdef at bol, or $(patsubst ...) et
al, would mean gmake, .ifdef at bol bsdmake...
I'm sure your new make code has lots of improvements under the hood,
but:
i) the surface appearance puts me off it completely. Yes, I'm shallow.
The thing is that font-lock is the best parser Emacs has. And finding
the right colons for rules needs some tricky parsing. So I put the two
together (like I had done for compile). Everything else would have
been reinventing the wheel. But while font-lock allows additional
properties, it will always also set the face...
ii) the timing is bad.
iii) the lack of consultation is bad.
Granted! But from my experiance many developers don't half understand
advanced makefiles and hardly mess with them. Thus the impact here is
way smaller than changes to, say, cc-mode. Not really an excuse, but
...
I have nothing more to say, so I'm shutting up about it after this.
That's the nice thing about freedom. :-) And you'll still have
correct recognition of targets.
The "you can customize it to look how you want" argument is not a
valid way to justify default behaviour that no-one apart from you
seems to like.
As a counter proposal, why not put the gaudier make font-locking
(together with that font-lock-negation stuff that no-one else liked
either), in your .emacs? Discuss bringing it back after 22.1 comes
out.
It seems that in both cases RMS is taking the attributes out of these
faces, leaving them only for those who want to customize them. Doesn't
that solve this issue nicely?
I don't know why, but when I took over makefile mode, single
character variables were already highlighted differently from
parenthesized multiletter ones.
I don't see this in Emacs-21.3 (I don't have an old CVS version to
test). There, $(FOO), $^, and $@ look exactly the same. In the current
CVS, they all look different.
I can't speak for your eyes, just for the code: I just randomly picked
out 1.88 of Mon Sep 20 and it already did long vars in
font-lock-variable-name-face and short vars in
font-lock-constant-face. The 10 years old 1.23 didn't font-lock single
char vars at all.
Hey cut it out! If it bothers you, why didn't you complain or do
something about this 10 years ago? On the one hand I get flamed for
having changed make-mode, and then right afterwards I also get
flamed for not having changed it.
I don't think I'm flaming you. I'm just saying I don't agree with you,
and I haven't seen anyone who does.
It's ok to discuss the things I did, but not why I didn't do something
else, just because I admit it struck me too.
coralament / best Grötens / liebe Grüße / best regards / elkorajn salutojn
Daniel Pfeiffer
--
lerne / learn / apprends / lär dig / ucz się Esperanto:
http://lernu.net/
|