w32 mule status

w32 mule status

[Edward Casey]

Does any of you know where the info reported by Options> Mule > Show all
of Mule status, is stored between emacs sessions? I have a satisfactory
setup on one machine that I am not able to emulate on another. I have
compared all the w32-* variables and they are the same on both machines.
The problem must have something to do with fontsets or extended character
mappings. Help!

Thanks,
Ed.

Re: w32 mule status

[Jason Rumney]

"Edward Casey" <ejmn@cpinternet.com> writes:

> Does any of you know where the info reported by Options> Mule > Show all
> of Mule status, is stored between emacs sessions?

It's not stored, unless you explicitly save it to a file with C-x C-w.

> I have a satisfactory setup on one machine that I am not able to
> emulate on another. I have compared all the w32-* variables and they
> are the same on both machines.  The problem must have something to
> do with fontsets or extended character mappings. Help!

You haven't described the problem, so I don't know how valid that
assumption is. I assume you wanted to compare the two files to see
what is different?

Re: w32 mule status

[Edward Casey]

<jasonr (Jason Rumney) @ f2s.com> wrote in message
news:uk6wm3pxn.fsf@jasonrumney.net...
> "Edward Casey" <ejmn@cpinternet.com> writes:
>
> > Does any of you know where the info reported by Options> Mule > Show
all
> > of Mule status, is stored between emacs sessions?
>
> It's not stored, unless you explicitly save it to a file with C-x C-w.
>
> > I have a satisfactory setup on one machine that I am not able to
> > emulate on another. I have compared all the w32-* variables and they
> > are the same on both machines.  The problem must have something to
> > do with fontsets or extended character mappings. Help!
>
> You haven't described the problem, so I don't know how valid that
> assumption is. I assume you wanted to compare the two files to see
> what is different?

Excuse my tardiness in replying but I have been trying to educate myself
enough so that I can frame the problem clearly, but without success. The
symptom is that out of 3 installations of emacnt 21.3.1 only one of them
is able to display Latin-4 characters. All of them are using the same
operating system win98, all have the same TTF fonts, all have the same
version of the macronize character routine that you wrote for me. This
routine is the only thing in my .emacs. Apparently I got one system
working by sheer luck (that is by copy-pasting and then evaluating
snippets of lisp culled from the web). If the state information is not
being saved by emacs then it is being saved by Windows since I can cold
boot each machine, set exactly the same options in the Mule environment,
and get different results. Apparently fontsets are used in both X and
ntemacs but can't be changed via S-mouse1 under the latter O.S.  I ran an
ediff on the outputs of Show all of mule status on two machines and the
differences (all under fontset default and all the indented entries
starting with [ ) were:

     [-outline-Courier
New-normal-i-normal-normal-13-97-96-96-c-80-iso8859-1]
     [-outline-Courier New-bold-r-normal-normal-13-97-96-96-c-80-iso8859-1
]
     [-outline-Courier
New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-1]
latin-iso8859-1  -*-iso8859-1
     [-outline-Courier
New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-1]
latin-iso8859-2  -*-iso8859-2
     [-outline-Courier
New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-2]
latin-iso8859-3  -*-iso8859-3
latin-iso8859-4  -*-iso8859-4
     [-outline-Courier
New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-4]
thai-tis620  -*-*-*-tis620-*
     [-outline-Tahoma-normal-r-normal-normal-13-97-96-96-p-60-tis620*-*]
greek-iso8859-7  -*-*-*-iso8859-7
     [-outline-Courier
New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-7]
arabic-iso8859-6 -*-*-*-iso8859-6
hebrew-iso8859-8 -*-*-*-iso8859-8
     [-outline-Courier
New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-8]
katakana-jisx0201 -*-*-*-jisx0208-sjis
     [-outline-Caslon-bold-r-normal-normal-13-97-96-96-p-80-jisx0208-sjis]
latin-jisx0201  -*-*-*-jisx0208-sjis
cyrillic-iso8859-5 -*-*-*-iso8859-5
     [-outline-Courier
New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-5]
latin-iso8859-9  -*-iso8859-9
     [-outline-Courier
New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-9]
latin-iso8859-15 -*-iso8859-15
latin-iso8859-14 -*-iso8859-14
japanese-jisx0208-1978 -*-*-*-jisx0208-sjis
chinese-gb2312  -*-*-*-gb2312.1980-*
     [-outline-SimSun-normal-r-normal-normal-13-97-96-96-p-70-gb2312*-*]
japanese-jisx0208 -*-*-*-jisx0208-sjis
     [-outline-Caslon-bold-r-normal-normal-13-97-96-96-p-80-jisx0208-sjis]
korean-ksc5601  -*-*-*-ksc5601.1989-*
     [-outline-Arial Unicode
MS-normal-r-normal-normal-13-97-96-96-p-60-ksc5601.1987*-*]
japanese-jisx0212 -*-*-*-jisx0212-*
chinese-cns11643-1 -*-*-*-cns11643.1992-1
chinese-cns11643-2 -*-*-*-cns11643.1992-2
japanese-jisx0213-1 -*-*-*-jisx0213.2000-1
chinese-big5-1  -*-*-*-big5-*
     [-outline-MingLiU-normal-r-normal-normal-13-97-96-96-c-70-big5*-*]
chinese-big5-2  -*-*-*-big5-*
chinese-sisheng  -*-sisheng_cwnn-*
ipa   -*-muleipa-*
vietnamese-viscii-lower -*-viscii1.1-*
vietnamese-viscii-upper -*-viscii1.1-*
arabic-digit  -*-*-*-mulearabic-0
arabic-1-column  -*-*-*-mulearabic-1
ascii-right-to-left -*-iso8859-1
lao   -*-*-*-mulelao-1
arabic-2-column  -*-*-*-mulearabic-2
indian-is13194  -*-*-*-is13194-devanagari
indian-1-column  -*-*-*-muleindian-1
tibetan-1-column -*-*-*-muletibetan-1
mule-unicode-2500-33ff -*-iso10646-1
mule-unicode-e000-ffff -*-iso10646-1
mule-unicode-0100-24ff -*-iso10646-1
     [-outline-Courier
New-normal-r-normal-normal-13-97-96-96-c-80-iso10646-1]

Is it possible that the "functions unify-8859-on-encoding-mode and
unify-8859-on-decoding-mode to turn on conversion between UTF-8 and
ISO-8859" (from your answer to the Cyrillic font problem) could be part of
the solution? I assume that the entry:

  [-outline-Courier New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-4]

is the only one relevant to my immediate purposes. For instance, I
certainly don't need Thai, but somehow I got it.

Thanks,

Ed

Re: w32 mule status

[Jason Rumney]

"Edward Casey" <ejmn@cpinternet.com> writes:

> I ran an ediff on the outputs of Show all of mule status on two
> machines and the differences (all under fontset default and all the
> indented entries starting with [ ) were:
>
>      [-outline-Courier
> New-normal-i-normal-normal-13-97-96-96-c-80-iso8859-1]

The differences being in the fontset suggest that the fonts are either
different on the machine that works, or the font setup is somehow
different. From the format you chose to post the "diffs" in, it is not
clear what the actual differences are you are seeing.

> Is it possible that the "functions unify-8859-on-encoding-mode and
> unify-8859-on-decoding-mode to turn on conversion between UTF-8 and
> ISO-8859" (from your answer to the Cyrillic font problem) could be part of
> the solution?

Only if you want Emacs to decode UTF-8 macroned characters as Latin-4.

> I assume that the entry:
>   [-outline-Courier New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-4]
>
> is the only one relevant to my immediate purposes.

Right.

Re: w32 mule status

[Edward Casey]

<jasonr (Jason Rumney) @ f2s.com> wrote in message
news:uzn5evckt.fsf@jasonrumney.net...
> "Edward Casey" <ejmn@cpinternet.com> writes:
>
> > I ran an ediff on the outputs of Show all of mule status on two
> > machines and the differences (all under fontset default and all the
> > indented entries starting with [ ) were:
> >
> >      [-outline-Courier
> > New-normal-i-normal-normal-13-97-96-96-c-80-iso8859-1]

This is an artifact of word-wrap. The lines continue ... Courier New...

>
> The differences being in the fontset suggest that the fonts are either
> different on the machine that works, or the font setup is somehow
> different. From the format you chose to post the "diffs" in, it is not
> clear what the actual differences are you are seeing.

The differences are just in the lines beginning with TAB [. These lines
are missing from the fontset on the machine that won't display Latin-4
characters.  But on all machines, when I run M-x describe-fontset, I read
that the frame is using a font and not a fontset.

>
> > Is it possible that the "functions unify-8859-on-encoding-mode and
> > unify-8859-on-decoding-mode to turn on conversion between UTF-8 and
> > ISO-8859" (from your answer to the Cyrillic font problem) could be
part of
> > the solution?
>
> Only if you want Emacs to decode UTF-8 macroned characters as Latin-4.
>
> > I assume that the entry:
> >   [-outline-Courier
New-normal-r-normal-normal-13-97-96-96-c-80-iso8859-4]
> >
> > is the only one relevant to my immediate purposes.
>
> Right.

I still think that the parameters for the fontsets have to be stored
somewhere. What does M-x describe-fontset read? I read somewhere that I
could see extended characters by appending to default-frame-alist but in
the "good" emacs installation I read:

default-frame-alist's value is
((tool-bar-lines . 0)
 (menu-bar-lines . 1))

Since I messed up one of the machines I have 2 fontsets called
fonset-default with different settings. This leads me to believe that the
(create-fontset-from-fontset-spec) function adds the definitions to a list
somewhere.

Re: w32 mule status

[Jason Rumney]

"Edward Casey" <ejmn@cpinternet.com> writes:

> The differences are just in the lines beginning with TAB [. These lines
> are missing from the fontset on the machine that won't display Latin-4
> characters.

So that machine does not have those fonts. Since this is Windows 98,
the default fonts for the US version are Latin-1 only. There is
another set of default fonts, which used to be available from the MS
website, and is also installed by various MS software packages. This
set supports many more encodings. If you do not look closely, it may
appear that the fonts are the same on both machines, but in the font
selection dialog, you have a "Script" dropdown selection, which will
be different on the machines that work compared to the one that
doesn't.

> I still think that the parameters for the fontsets have to be stored
> somewhere.

Only if you set them.

> What does M-x describe-fontset read?

It describes the fontset you specify. But this is a red herring. Emacs
21 displays all characters regardless of whether you use a font or a
fontset. The fontset gives you finer control over which fonts are used
to display which characters.

> Since I messed up one of the machines I have 2 fontsets called
> fonset-default with different settings. This leads me to believe that the
> (create-fontset-from-fontset-spec) function adds the definitions to a list
> somewhere.

If you put that in .emacs, it is saved in .emacs. There is nothing
magic going on behind the scenes.

Re: w32 mule status

[Edward Casey]

<jasonr (Jason Rumney) @ f2s.com> wrote in message
news:ur7qquggp.fsf@jasonrumney.net...
> "Edward Casey" <ejmn@cpinternet.com> writes:
>
> > The differences are just in the lines beginning with TAB [. These
lines
> > are missing from the fontset on the machine that won't display Latin-4
> > characters.
>
> So that machine does not have those fonts. Since this is Windows 98,
> the default fonts for the US version are Latin-1 only. There is
> another set of default fonts, which used to be available from the MS
> website, and is also installed by various MS software packages. This
> set supports many more encodings. If you do not look closely, it may
> appear that the fonts are the same on both machines, but in the font
> selection dialog, you have a "Script" dropdown selection, which will
> be different on the machines that work compared to the one that
> doesn't.
>

You are right in that the original Win98 installation didn't have them but
I've gotten them either from the web or from other software packages. They
work fine under 98 on one machine and not at all on the other two. The
exact same font files are on all 3 machines. The only thing that two
non-working installations have in common is that they're not connected to
the internet and 1 (the laptop) has no modem. Although MS can justly be
faulted for it's ad hoc software design history, I can't believe, for
instance, that a setting in Outlook Express can have any effect on emacs
font display capabilities.
If I run C-h h on the good machine with font Arial Unicode I see glyphs
for Russian, Korean, four flavors of Chinese, etc. Only a 25% of the lines
have empty rectangles on them. If I change the font in this buffer to
Palatino Linotype I see almost as many foreign glyphs (JIS Chinese is
missing). In fact, on the "bad" machines, with 'list character sets' I can
get most of the glyphs to display except for Latin-3 and Latin-4. ???
Finally, if I evaluate this form on all machines:

(if (eq window-system 'w32)        ; emacs 20 & up
    (defun insert-x-style-font()
      "Insert a string in the X format which describes a font the user can
select from the Windows font selector."
      (interactive)
      (insert (prin1-to-string (w32-select-font)))))

I get exactly the same X format string.


> > I still think that the parameters for the fontsets have to be stored
> > somewhere.
>
> Only if you set them.
>
> > What does M-x describe-fontset read?
>
> It describes the fontset you specify. But this is a red herring. Emacs
> 21 displays all characters regardless of whether you use a font or a
> fontset. The fontset gives you finer control over which fonts are used
> to display which characters.

This is good to know. Even if I knew how to make a fontset with a Latin-4
font specified in it, I probably would't be able to see the glyphs until I
could resolve the other mysterious problem.

>
> > Since I messed up one of the machines I have 2 fontsets called
> > fonset-default with different settings. This leads me to believe that
the
> > (create-fontset-from-fontset-spec) function adds the definitions to a
list
> > somewhere.
>
> If you put that in .emacs, it is saved in .emacs. There is nothing
> magic going on behind the scenes.

That at least is a relief. After exiting and restarting emacs I no longer
had 2 fontsets called default, only 1 together with 'standard: 13dot.'

Thanks again for your help, Mr. Rumney. Since MS Windows is probably
off-topic here, I'll keep plugging away at solving the problem and
learning emacs and I won't bother you for a while, at least not until I
lose patience again.