Re: cm fonts status - SOLVED - v2

2003-04-02 Thread Dan Armak
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Tuesday 01 April 2003 23:21, Dan Armak wrote:
> OK, I'll look at the problematic chars tomorrow. With any luck they will 
have
> a single cause of failure and I will find it :-)

It looks like there's some problem in pfaedit, the font editor I use. (If you 
know of any other font editor I could try please tell me, I couldn't find 
anything else that handles both type1 and truetype).

If I open the cmsy font from bakoma, pfaedit converts it to its internal 
postscript format. Saving as truetype again is basically exporting from 
pfaedit's pov. Now, if I export the cmsy font agin, _making no changes_, the 
newly created font won't provide the 'broken' characters such as /leq. If I 
open the new font in pfaedit again, though, the characters are still there 
and look perfectly ok, so I don't really know what the problem is. I'm going 
to mail the pfaedit-devel list asking for help.

- -- 
Dan Armak
Gentoo Linux developer (KDE)
Matan, Israel
Public GPG key: http://www.gentoo.org/~danarmak/danarmak-gpg-public.key
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+ivqdUI2RQ41fiVERAjMIAJ4vPzEPifjZ5HQwGnIUNYvWexGcSQCdH1ps
sB5CEK88d+RROpU7LuK+FvE=
=FJnE
-END PGP SIGNATURE-



Re: cm fonts status - SOLVED - v2

2003-04-01 Thread John Levon
On Tue, Apr 01, 2003 at 11:21:23PM +0300, Dan Armak wrote:

> > Qt doesn't let us use normal X fonts when using fontconfig, at least on
> > Red Hat etc.
> 
> Ah yes, I forgot the original problem - fontconfig won't select/return the 
> bluesky fonts because they don't provide the latin1 characters... bummer.

Well, that, and they have patches to disable *any* "old-style" X fonts
for Qt/GTK+ apps.

> OK, I'll look at the problematic chars tomorrow. With any luck they will have 
> a single cause of failure and I will find it :-)

thanks a lot

regards
john


Re: cm fonts status - SOLVED - v2

2003-04-01 Thread Dan Armak
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Tuesday 01 April 2003 23:06, John Levon wrote:
> On Tue, Apr 01, 2003 at 04:56:24PM +0300, Dan Armak wrote:
> 
> > Hmm. Looking at it, it provides the mapping of the orig bluesky type1 
fonts, 
> > not that of the bakoma ones. Why can't lyx use that and the bluesky fonts 
> > natively?...
> 
> Qt doesn't let us use normal X fonts when using fontconfig, at least on
> Red Hat etc.

Ah yes, I forgot the original problem - fontconfig won't select/return the 
bluesky fonts because they don't provide the latin1 characters... bummer.

> > A full list of the problematic chars:
> > \leq \kappa \circleddash \nleqq \upuparrows
> 
> This isn't a full list :)

OK, I'll look at the problematic chars tomorrow. With any luck they will have 
a single cause of failure and I will find it :-)

- -- 
Dan Armak
Gentoo Linux developer (KDE)
Matan, Israel
Public GPG key: http://www.gentoo.org/~danarmak/danarmak-gpg-public.key
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+ifTEUI2RQ41fiVERAvsLAKCCZbvDqdNL1tSokrLPseuF9/zmvgCfZB+v
5O9WqNlVBnlyECfOtJ/Phv0=
=p0R0
-END PGP SIGNATURE-



Re: cm fonts status - SOLVED - v2

2003-04-01 Thread John Levon
On Tue, Apr 01, 2003 at 04:56:24PM +0300, Dan Armak wrote:

> Hmm. Looking at it, it provides the mapping of the orig bluesky type1 fonts, 
> not that of the bakoma ones. Why can't lyx use that and the bluesky fonts 
> natively?...

Qt doesn't let us use normal X fonts when using fontconfig, at least on
Red Hat etc.

> This is still WIP due to issue #2 below, but since I'm prolly not going to do 
> anything more with this today, I wanted to update the list. The fonts are at 
> http://cvs.gentoo.org/~danarmak/fonts-v2.tar.bz2 - testing appreciated (last 
> time it took about 10 minutes for the big problems to be found, but this time 
> I've tried to do better testing).

Looks good.

> 1. Quality loss: tolerable for on-screen editing (imo at least).

I think it's fine.

> 2. Some chars won't draw - lyx displays a square instead. This is the major 
> thing left to be fixed.
> 
> A full list of the problematic chars:
> \leq \kappa \circleddash \nleqq \upuparrows

This isn't a full list :)

Attached is a hacked-up lyx document with the symbols from lib/symbols/
It should load into 1.3.x, just ignore the warning

The problematic ones are :

\kappa
\spadesuit
\leq / \le
\upuparrows
\circleddash
\nleqq
\backepsilon

These ones really need to be fixed. It would also be nice if we could
add the ones currently in ERT from wasy, at some point.

regards,
john

#LyX 1.3 created this file. For more info see http://www.lyx.org/
\lyxformat 223
\textclass article
\language english
\inputencoding latin1
\fontscheme default
\graphics default
\paperfontsize default
\spacing single 
\papersize Default
\paperpackage a4
\use_geometry 0
\use_amsmath 0
\use_natbib 0
\use_numerical_citations 0
\paperorientation portrait
\secnumdepth 3
\tocdepth 3
\paragraph_separation indent
\defskip medskip
\quotes_language english
\quotes_times 2
\papercolumns 1
\papersides 1
\paperpagestyle default
\tracking_changes 0
\end_header

\layout Standard


\begin_inset Formula $\acute{}$
\end_inset 


\begin_inset Formula $\bar{}$
\end_inset 


\begin_inset Formula $\breve{}$
\end_inset 


\begin_inset Formula $\check{}$
\end_inset 


\begin_inset Formula $\dddot{}$
\end_inset 


\begin_inset Formula $\ddot{}$
\end_inset 


\begin_inset Formula $\dot{}$
\end_inset 


\begin_inset Formula $\grave{}$
\end_inset 


\begin_inset Formula $\hat{}$
\end_inset 


\begin_inset Formula $\mathring{}$
\end_inset 


\begin_inset Formula $\overbrace{}$
\end_inset 


\begin_inset Formula $\overleftarrow{}$
\end_inset 


\begin_inset Formula $\overleftrightarrow{}$
\end_inset 


\begin_inset Formula $\overline{}$
\end_inset 


\begin_inset Formula $\overrightarrow{}$
\end_inset 


\begin_inset Formula $\tilde{}$
\end_inset 


\begin_inset Formula $\underbar{}$
\end_inset 


\begin_inset Formula $\underbrace{}$
\end_inset 


\begin_inset Formula $\underleftarrow{}$
\end_inset 


\begin_inset Formula $\underleftrightarrow{}$
\end_inset 


\begin_inset Formula $\underline{}$
\end_inset 


\begin_inset Formula $\underrightarrow{}$
\end_inset 


\begin_inset Formula $\vec{}$
\end_inset 


\begin_inset Formula $\widehat{}$
\end_inset 


\begin_inset Formula $\widetilde{}$
\end_inset 


\begin_inset Formula $\dots$
\end_inset 


\begin_inset Formula $\cdots$
\end_inset 


\begin_inset Formula $\ddots$
\end_inset 


\begin_inset Formula $\dotsb$
\end_inset 


\begin_inset Formula $\dotsc$
\end_inset 


\begin_inset Formula $\dotsi$
\end_inset 


\begin_inset Formula $\dotsm$
\end_inset 


\begin_inset Formula $\dotso$
\end_inset 


\begin_inset Formula $\ldots$
\end_inset 


\begin_inset Formula $\vdots$
\end_inset 


\begin_inset Formula $\lyxnegspace$
\end_inset 


\begin_inset Formula $\lyxposspace$
\end_inset 


\begin_inset Formula $\!$
\end_inset 


\begin_inset Formula $\negmedspace$
\end_inset 


\begin_inset Formula $\negthickspace$
\end_inset 


\begin_inset Formula $\,$
\end_inset 


\begin_inset Formula $\:$
\end_inset 


\begin_inset Formula $\;$
\end_inset 


\begin_inset Formula $\quad$
\end_inset 


\begin_inset Formula $\qquad$
\end_inset 


\begin_inset Formula $\alpha$
\end_inset 


\begin_inset Formula $\beta$
\end_inset 


\begin_inset Formula $\gamma$
\end_inset 


\begin_inset Formula $\delta$
\end_inset 


\begin_inset Formula $\epsilon$
\end_inset 


\begin_inset Formula $\zeta$
\end_inset 


\begin_inset Formula $\eta$
\end_inset 


\begin_inset Formula $\theta$
\end_inset 


\begin_inset Formula $\iota$
\end_inset 


\begin_inset Formula $\kappa$
\end_inset 


\begin_inset Formula $\lambda$
\end_inset 


\begin_inset Formula $\mu$
\end_inset 


\begin_inset Formula $\nu$
\end_inset 


\begin_inset Formula $\xi$
\end_inset 


\begin_inset Formula $\pi$
\end_inset 


\begin_inset Formula $\rho$
\end_inset 


\begin_inset Formula $\sigma$
\end_inset 


\begin_inset Formula $\tau$
\end_inset 


\begin_inset Formula $\upsilon$
\end_inset 


\begin_inset Formula $\phi$
\end_inset 


\begin_inset Formula $\chi$
\end_inset 


\begin_inset Formula $\psi$
\end

Re: cm fonts status - SOLVED - v2

2003-04-01 Thread John Levon
On Tue, Apr 01, 2003 at 04:56:24PM +0300, Dan Armak wrote:

> 4. There are still no eufm10 and wasy10 fonts, because there aren't such 
> bakoma fonts (maybe they are in the bigger bakoma collection on ctan though). 
> However AFAICS lyx doesn't need/use these fonts anywhere? I tested all the 
> characters listed in the matheditor float and they were present without these 
> fonts (except for the ones from issue 2 above, and they got drawn as squares 
> and not written in red, so they must be from the existing fonts as well).

wasy is used for (from lib/symbols) :

male   wasy  26  0 xx
female wasy  25  0 xx
currency   wasy  27  0 xx
phone  wasy  7   0 xx
recorder   wasy  6   0 xx
clock  wasy  28  0 xx
lightning  wasy  18  0 xx
pointerwasy  9   0 xx
RIGHTarrow wasy  17  0 xx
LEFTarrow  wasy  16  0 xx
UParrowwasy  75  0 xx
DOWNarrow  wasy  76  0 xx
#AC wasy  n0.5pt\wasyfamily\char58\kern0.5pt}}
#HF {\leavevmode
#   \lower0.9pt\hbox to 0pt{\kern0.5pt\wasyfamily\char58\hss}%
#   \raise0.9pt\hbox{\kern0.5pt\wasyfamily\char58\kern0.5pt}}
#VHFwasy  64  0 xx
Square wasy  50  0 xx
#CheckedBox wasy  50\hss}\hbox{\wasyfamily\char8  0 x
XBox   wasy  52  0 xx
hexagonwasy  55  0 xx
pentagon   wasy  68  0 xx
octagonwasy  56  0 xx
varhexagon wasy  57  0 xx
hexstarwasy  65  0 xx
varhexstar wasy  66  0 xx
davidsstar wasy  67  0 xx
diameter   wasy  31  0 xx
invdiameterwasy  21  0 xx
varangle   wasy  30  0 xx
wasylozengewasy  53  0 xx
kreuz  wasy  54  0 xx
smiley wasy  44  0 xx
frowniewasy  47  0 xx
blacksmileywasy  45  0 xx
sunwasy  46  0 xx
checkedwasy  8   0 xx
bell   wasy  10  0 xx
eighthnote wasy  11  0 xx
quarternotewasy  12  0 xx
halfnote   wasy  13  0 xx
fullnote   wasy  14  0 xx
twonotes   wasy  15  0 xx
brokenvert wasy  124 0 xx
ataribox   wasy  109 0 xx
wasytherefore  wasy  5   0 xx
Circle wasy  35  0 xx
CIRCLE wasy  32  0 xx
Leftcircle wasy  73  0 xx
LEFTCIRCLE wasy  71  0 xx
Rightcirclewasy  74  0 xx
RIGHTCIRCLEwasy  72  0 xx
#LEFTcircle {\hbox to
0pt{\wasyfamily\char71\hss}\hbox{\wasyfamily\char35}}
#RIGHTcircle{\hbox to
0pt{\wasyfamily\char72\hss}\hbox{\wasyfamily\char35}}

#
# wasy astronomy
#

vernal wasy  23  0 xx
ascnodewasy  19  0 xx
descnode   wasy  20  0 xx
fullmoon   wasy  35  0 xx
newmoonwasy  32  0 xx
leftmoon   wasy  36  0 xx
rightmoon  wasy  37  0 xx
astrosun   cmsy 175  0 mathbin  x
mercurywasy  39  0 xx
venus  wasy  25  0 xx
earth  wasy  38  0 xx
mars   wasy  26  0 xx
jupiterwasy  88  0 xx
saturn wasy  89  0 xx
uranus wasy  90  0 xx
neptunewasy  91  0 xx
pluto  wasy  92  0 xx

#
# wasy zodiac
#
aries  wasy  23  0 xx
taurus wasy  93  0 xx
gemini wasy  94  0 xx
cancer wasy  95  0 xx
leowasy  19  0 xx
virgo  wasy  96  0 xx
libra  wasy  97  0 xx
scorpiowasy  98  0 xx
sagittariuswasy  99  0 xx
capricornuswasy  100 0 xx
aquarius   wasy  101 0 xx
pisces wasy  102 0 xx
conjunctionwasy  86  0 xx
opposition wasy  87  0 xx
# APL characters
APLstarwasy 

Re: cm fonts status - SOLVED - v2

2003-04-01 Thread Dan Armak
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Monday 31 March 2003 19:32, John Levon wrote:
> On Mon, Mar 31, 2003 at 07:30:26PM +0300, Dan Armak wrote:
> 
> > 1) exists in cmmi10 (both bakoma and bluesky-derived versions). So 
apparently 
> > lyx isn't using that font here. (BTW I'd really appreciate an explanation 
of 
> > what characters it takes from what font, since most of them are present in 
> > several.)
> 
> Look at lib/symbols file.

Hmm. Looking at it, it provides the mapping of the orig bluesky type1 fonts, 
not that of the bakoma ones. Why can't lyx use that and the bluesky fonts 
natively?...

Anyway I decided I was beaten on the path of trying to automagically convert 
the bluesky fonts to ttf ones that look like bakoma's. And having already 
spent about 20 hours on that, I decided I could spare another hour or two 
doing the same thing by hand.
So I've taken the bakoma fonts, removed all glyphs, and copied over the 
bluesky glyphs, changing the font properties as needed. In the end nothing 
remains of the bakoma version but the way the characters are arranged, which 
is what we want. This should be ok wrt bakoma's copyright (IANAL)...

This is still WIP due to issue #2 below, but since I'm prolly not going to do 
anything more with this today, I wanted to update the list. The fonts are at 
http://cvs.gentoo.org/~danarmak/fonts-v2.tar.bz2 - testing appreciated (last 
time it took about 10 minutes for the big problems to be found, but this time 
I've tried to do better testing).

Known issues:

1. Quality loss: tolerable for on-screen editing (imo at least).
Details: I converted the glyphs from postscript to truetype. That's apparently 
a lossy conversion (unlike the other way around) - which I only found out 
when it was too late. I converted to opentype (which has the same spline 
format as type1) which recovered most of the quality (but is still worse than 
the printed bluesky). So it looks worse than the bakoma stuff (due to the 
lossy conversion I suppose), but tolerable for editing (as long as you don't 
write too many formulas :-). Probably something could be done if I started 
all over again... (The font is also smaller than the bakoma one - try using a 
bigger fontsize. Can we get a separate fontsize setting for the mathed?)

Also, the hinting went mad for some reason and most chars were over-hinted 
into smudged oblivion. I told pfaedit to remove all hints/instructions and 
autohint. Now it's usable - but I probably lost some of the custom hinting 
instructions that make bluesky look good (in particular most of the Latin 
alphabet in cmr10 had custom stuff). Again, tolerable (for me) since this is 
just for editing.

2. Some chars won't draw - lyx displays a square instead. This is the major 
thing left to be fixed.

Looking at the kappa char for example (from cmmi10), every other char in the 
font draws OK and AFAICS there's absolutely no difference between it and the 
bakoma one, in the kappa char's properties or otherwise. I really don't know 
what to do about this kappa problem...

A full list of the problematic chars:
\leq \kappa \circleddash \nleqq \upuparrows

3. At some point the characters lost their names and got the names of the 
unicode characters at their encoding positions instead. This doesn't affect 
lyx which draws by encoding value not name (I wish I'd known that sooner), 
but it makes further editing of the fonts icky. If there's sufficiently good 
reason, I can create them again, and hopefully keep the names that time.

4. There are still no eufm10 and wasy10 fonts, because there aren't such 
bakoma fonts (maybe they are in the bigger bakoma collection on ctan though). 
However AFAICS lyx doesn't need/use these fonts anywhere? I tested all the 
characters listed in the matheditor float and they were present without these 
fonts (except for the ones from issue 2 above, and they got drawn as squares 
and not written in red, so they must be from the existing fonts as well).

- -- 
Dan Armak
Gentoo Linux developer (KDE)
Matan, Israel
Public GPG key: http://www.gentoo.org/~danarmak/danarmak-gpg-public.key
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+iZqMUI2RQ41fiVERAjNRAJ4sp8d1ld7StP838ZQDyreQVSD1ZgCbBoZI
6grXoK17v5YZubRNfpVHpdA=
=WV20
-END PGP SIGNATURE-