Re: [Patch] Add file variables to all lisp files

[Katsumi Yamaoka <yamaoka@jpl.org>]

>>>>> Tassilo Horn wrote:
> Katsumi Yamaoka <yamaoka@jpl.org> writes:

>> Please note that XEmacs ignores the Local Variables section in the end
>> of a file if the file size is larger than 3000 bytes.

> Is that intended or a bug nobody cares of?

I guess it is intended to make loading of files fast.  When
loading a file XEmacs reads the first 3000 bytes of it first in
the raw mode to decide the coding system and then decodes and
reads it again.

>> Because of this, in the Japanese locale XEmacs 21.4 recognizes
>> gnus-sum.el to be encoded by the shift_jis coding system.  21.5
>> ignores it either (seems always to set `buffer-file-coding-system' to
>> `iso-8859-1-unix', though).

> But that would be the case without the local variables, too, right?

Yes, misjudging of the coding system always happens regardless
of the existence of the coding cookie at the bottom of a file.
If only the coding cookie were recognized.

>>> +;; Local Variables:
>>> +;; coding: iso-8859-1
>>> +;; indent-tabs-mode: t
>>> +;; tab-width: 8
>>> +;; End:
>>
>> I'm not sure whether such cares are really necessary.

> As long as it helps some people and doesn't handicap others, I think
> it's the right thing to do.

I don't need it but I'm not against to adding it.  But to put it
mildly, it looks like a notice board of a school where many bad
boys and girls are. ;-)

> BTW: Is it normal that CVS is so incredibly slow if you have a lot of
> modifications in your working copy? Here a cvs up takes about 4 minutes,
> 3:50 in which it seems to do nothing and the last 10 seconds it outputs
> the updated files and directories.

`cvs update' takes less than four seconds here in Tokyo now.