lynx-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: lynx-dev cleanup chartrans [patch]


From: Klaus Weide
Subject: Re: lynx-dev cleanup chartrans [patch]
Date: Thu, 25 Feb 1999 07:36:53 -0600 (CST)

On Thu, 25 Feb 1999, Leonid Pauzner wrote:
> 25-Feb-99 04:49 Klaus Weide wrote:
> 
> > I am a bit skeptical about this patch.  Maybe it's just me not being able
> > to throw anything away.  So I leave the decision to someone else.  But
> > consider the comments below.
> 
> >> @@ -329,20 +332,6 @@
> >>  PUBLIC CONST char ** LYCharSets[MAXCHARSETS]={
> >>         ISO_Latin1,             /* ISO Latin 1          */
> >>         SevenBitApproximations, /* ISO 8859-15 (Latin 9)*/
> 
> That is an example on "SevenBitApproximations" usage,
> we couldn't go any further with UCdomap.c "new method meets old method!"
> cleanups if we remove this reference to SevenBitApproximations.
> I mean this is not the last step in cleaning up things,
> ISO_Latin1 may be the next :-)

I guess I shouldn't complain too much if you are willing to clean up
the duplication left being by me...

> > What's special about 8859-15, to be the only one left intact here
> > besides 8859-1?  "7 Bit Approximations" would much rather deserve that
> > honor.
> Well, yes, but this got hidden the another side...

It would be more logical to put "7 Bit Approximations" in the 2nd place
though.  That would change the order in the "Display character set" option
list, but maybe that's not a bad idea anyway.

> >> @@ -394,41 +369,6 @@
> >>     *  Placeholders for Unicode tables. - FM
> >>     */
> >>    {-1,"iso-8859-15",   UCT_ENC_8BIT,0,0,0,     UCT_R_8BIT,UCT_R_ASCII},
> >> -  {-1,"cp850",         UCT_ENC_8BIT,0,
> >> -                       UCT_REP_SUPERSETOF_LAT1,
> >> -                       0,                      UCT_R_8BIT,UCT_R_ASCII},
> > [ etc - including CJK, 7-bit approx., transparent ]
> 
> > The various tables here served to provide some minimal information
> > (without taking much space) about several charsets / Display character
> > sets even in the case where chartrans table files for them were not
> > included.  Yes it's redundant; however, sometimes redundancy *may* be
> > good.
> Yes, but in this case I think this redundancy may be misguiding for other
> changes. In fact, no fields from this struct are used except mime name and
> encoding name, only UCT_REP_* _may_ be useful when we are very close to
> old-style LATIN1 charset.

Yes, most of those bits are underused...  probably even more so now.
I always liked to keep the possibility open to one day do something more
with that info, or put more detailed info in that struct (like *what kind*
of CJK encoding, or which scripts of Unicode where present in a charset's
repertoire).  But it hasn't happened, and leaving it open is not exactly
compatible with your goal of cleaning up.
 
> > Historically, that information was first listed here because
> > EXP_CHARTRANS was optional; later, because for quite some time not all
> > of the Display character sets traditionally handled by Lynx had
> > associated .tbl files (not just the CJK ones).  I always kept thinking
> > that whoever installs Lynx should be able to exclude most (or all) of
> This is still possible based on chrtrans/README.tables, isn't it?

... which no-one reads to take away, only to add. :)

> And if someone considered seriously on removing 7-bit approximations file
> what we could benefit for him, 

If you mean "what could be the benefit for him" - saving same memory space.

>                                anyone tested?

Not me, or at least not in a very long time.  Someone with a strong interest
in saving memory might want to...

>                                               It may be probably safer to edit
> entities.h and exclude most entries except ISO LATIN1 instead.

That would be another bit of memory to save.  No I don't know whether it's
worth it - and it shouldn't be done under "normal" circumstances (when
memory is not a BIG problem).

None if this should be "unsafe" of course, if that means lynx would crash
or something like that.

> > the chartrans files, and still have some minimal fallback behavior in
> > that case, at least as good as say Lynx 2.6.  I am not sure this is
> > still the case; maybe it's not wanted.  It probably hasn't been tested
> > by anyone in a long time.  An example would be the case where someone
> > wanted to not have the large 7-bit approximations file, but still have
> > 7-bit approximations available as Display character set to at least
> > deal with the "classical" ISO-8859-1 chars and entities.
> How about Euro/(TM)/Copyright/emdash/etc requests?

It's not strictly 8859-1 but with some extensions - &trade, &copy, &emdash
were "classically" covered, euro is much to new (and isn't listed in
entities.h even now, as of dev.17).

> > Well, since even Henry seems to not have tried that, it may be
> > irrelevant.  It could become relevant if one day configure allowed
> > to specify the to-be-included chartrans .tbl files.
> No problem, except the default table.

You may well be right.

   Klaus

reply via email to

[Prev in Thread] Current Thread [Next in Thread]