Ghostscript development

Ralph Giles | 2 Mar 2009 09:18

Re: [gs-cvs] rev 9514 - trunk/gs/psi

On Sun, Mar 1, 2009 at 10:13 PM,  <alexcher <at> ghostscript.com> wrote:

> When /JBIG2Decode filter takes /.jbig2globalctx parameter and retains a
> pointer to it, the filter must not be allocated in global memory if the
> parameter is local. Bug 690308.

Oops. Thanks, Alex.

 -r

headers

Daniel Spannbauer | 4 Mar 2009 11:23

Embedding ttf-font in GS

Hello,

I'm using GS 8.15 on a SuSE 10.2-System.
Now, I nedd a ttf-Font in GS.
The font is named han_nom_a.ttf
So, I put these line in Fontmap.GS in /usr/share/ghostscript/8.15/lib:

/hannom-a   (/usr/share/ghostscript/fonts/han_nom_a.ttf)  ;

But when I try to put a ps thru GS, I get:

              Loading NimbusRomNo9L-Regu font from 
/usr/share/ghostscript/fonts/n021003l.pfb... 2568236 1261266 1820196 
518519 2 done.
Loading hannom-a font from /usr/share/ghostscript/fonts/han_nom_a.ttf... 
ERROR: /rangecheck in --string--
Operand stack:
    hannom-a   Font   hannom-a   518871   hannom-a   --nostringval-- 
hannom-a   (/usr/share/ghostscript/fonts/han_nom_a.ttf)   false 
--nostringval--   101098   101098
Execution stack:
    %interp_exit   .runexec2   --nostringval--   --nostringval-- 
--nostringval--   2   %stopped_push   --nostringval--   --nostringval-- 
   --nostringval--   false   1   %stopped_push   1   3   %oparray_pop 
1   3   %oparray_pop   --nostringval--   1   3   %oparray_pop   1   3 
%oparray_pop   .runexec2   --nostringval--   --nostringval-- 
--nostringval--   2   %stopped_push   --nostringval--   1   3 
%oparray_pop   2   3   %oparray_pop   --nostringval--   --nostringval-- 
   --nostringval--   --nostringval--   --nostringval--   --nostringval-- 
   false   1   %stopped_push   5   4   %oparray_pop   --nostringval--

(Continue reading)

headers

Daniel Spannbauer | 4 Mar 2009 11:35

Re: Embedding ttf-font in GS


ademmler schrieb:
> hi,
> 
> can you try to use GS 8.63. This is much newer.
> 
> greetings Alexander

Shure, I can try this, but then I have to update a huge ammount of PCs, 
so I hoped it will be possible with 8.15

regards

Daniel

> 
> Am 04.03.2009 um 11:23 schrieb Daniel Spannbauer:
> 
>> Hello,
>>
>> I'm using GS 8.15 on a SuSE 10.2-System.
>> Now, I nedd a ttf-Font in GS.
>> The font is named han_nom_a.ttf
>> So, I put these line in Fontmap.GS in /usr/share/ghostscript/8.15/lib:
>>
>> /hannom-a   (/usr/share/ghostscript/fonts/han_nom_a.ttf)  ;
>>
>> But when I try to put a ps thru GS, I get:
>>
>>

(Continue reading)

headers

Ray Johnston | 4 Mar 2009 14:44

Re: Embedding ttf-font in GS

Hello Daniel,

Actually the latest release is 8.64 (not 8.63). While you are
welcome to search the "doc/History8.htm" (or doc/Details8.htm)
for likely fixes, we can't really support 8.15. It is WAY too
old.

Best Regards,
Ray Johnston

Daniel Spannbauer wrote:

Hello,

I'm using GS 8.15 on a SuSE 10.2-System.
Now, I nedd a ttf-Font in GS.
The font is named han_nom_a.ttf
So, I put these line in Fontmap.GS in /usr/share/ghostscript/8.15/lib:

/hannom-a   (/usr/share/ghostscript/fonts/han_nom_a.ttf) ;

But when I try to put a ps thru GS, I get:

             Loading NimbusRomNo9L-Regu font from /usr/share/ghostscript/fonts/n021003l.pfb... 2568236 1261266 1820196 518519 2 done.
Loading hannom-a font from /usr/share/ghostscript/fonts/han_nom_a.ttf... ERROR: /rangecheck in --string--
Operand stack:
   hannom-a   Font   hannom-a   518871   hannom-a   --nostringval-- hannom-a   (/usr/share/ghostscript/fonts/han_nom_a.ttf)   false --nostringval--   101098   101098
Execution stack:
   %interp_exit   .runexec2   --nostringval--   --nostringval-- --nostringval--   2   %stopped_push   --nostringval--   --nostringval-- --nostringval--   false   1   %stopped_push   1   3   %oparray_pop 1   3   %oparray_pop   --nostringval--   1   3   %oparray_pop   1   3 %oparray_pop   .runexec2   --nostringval--   --nostringval-- --nostringval--   2   %stopped_push   --nostringval--   1   3 %oparray_pop   2   3   %oparray_pop   --nostringval--   --nostringval-- --nostringval--   --nostringval--   --nostringval--   --nostringval-- false   1   %stopped_push   5   4   %oparray_pop   --nostringval-- --nostringval--   --nostringval--   --nostringval--   --nostringval-- %array_continue   --nostringval--   --nostringval--   --nostringval-- %loop_continue   --nostringval--   --nostringval--   --nostringval-- --nostringval--   %loop_continue   --nostringval--   10 --nostringval--   --nostringval--   false   1   %stopped_push --nostringval--   --nostringval--   --nostringval--   --nostringval-- --nostringval--   %array_continue   --nostringval--   --nostringval-- --nostringval--
Dictionary stack:
   --dict:1136/3371(ro)(G)--   --dict:0/20(G)--   --dict:106/200(L)-- --dict:47/65(ro)(G)--   --dict:1136/3371(ro)(G)-- --dict:28/50(ro)(G)--   --dict:15/40(L)--
Current allocation mode is local
Current file position is 295
ESP Ghostscript 8.15.3: Unrecoverable error, exit code 1

Can anybody help me at this point?

Regards

Daniel

_______________________________________________
gs-devel mailing list
gs-devel <at> ghostscript.com
http://www.ghostscript.com/mailman/listinfo/gs-devel

headers

Paul Smedley | 21 Mar 2009 14:13

Tips on correct .ps syntax

Hi Guys,

I've been working on getting the OS/2 ports of Ghostscript and CUPS to 
the point where they can be used as a replacement for the printer system 
on OS/2, so that modern printers can be used (as nobody writes native 
printer drivers for OS/2 anymore).

The idea is to use the IBM postscript driver to generate .ps, which is 
routed to CUPS - which in the main works quite well.

Fortunately, IBM did release the source code to their postscript driver 
via the DDK, unfortunately, it does create some buggy .ps code - at 
least the code is available :)

One of the remaining problems is printing things like equations from 
Lotus Wordpro, which require a symbol font.

It seems the IBM postscript code embeds the font in the postscript code, 
but ghostscript doesn't see it correctly.

Wondering if anyone can provide some good links to some postscript code 
references, so I can try and fix the driver?

A sample file that contains just a few equation characters, but renders 
as am empty page in gsview is at http://smedley.info/test.ps

Any feedback appreciated!

Cheers,

Paul

headers

Paul Smedley | 22 Mar 2009 02:48

Re: Tips on correct .ps syntax

One finding this morning, it seems:
%%DocumentSuppliedResources:
%%+ font Times New Roman
%%+ font Symbol Set

Needs to be:
%%DocumentSuppliedFonts:
%%+ font Times New Roman
%%+ font Symbol Set

which helps with part of the problem, but it 'appears' that the 
Ghostscript copy of Times New Roman is used instead of the embedded copy 
- is there a way to 'force' the use of the supplied font - even though 
Ghostscript has a copy already?

Cheers,

Paul.

Paul Smedley wrote:
> Hi Guys,
> 
> I've been working on getting the OS/2 ports of Ghostscript and CUPS to 
> the point where they can be used as a replacement for the printer system 
> on OS/2, so that modern printers can be used (as nobody writes native 
> printer drivers for OS/2 anymore).
> 
> The idea is to use the IBM postscript driver to generate .ps, which is 
> routed to CUPS - which in the main works quite well.
> 
> Fortunately, IBM did release the source code to their postscript driver 
> via the DDK, unfortunately, it does create some buggy .ps code - at 
> least the code is available :)
> 
> One of the remaining problems is printing things like equations from 
> Lotus Wordpro, which require a symbol font.
> 
> It seems the IBM postscript code embeds the font in the postscript code, 
> but ghostscript doesn't see it correctly.
> 
> Wondering if anyone can provide some good links to some postscript code 
> references, so I can try and fix the driver?
> 
> A sample file that contains just a few equation characters, but renders 
> as am empty page in gsview is at http://smedley.info/test.ps
> 
> Any feedback appreciated!
> 
> Cheers,
> 
> Paul
> _______________________________________________
> gs-devel mailing list
> gs-devel <at> ghostscript.com
> http://www.ghostscript.com/mailman/listinfo/gs-devel

headers

Ken Sharp | 21 Mar 2009 16:55

Re: Tips on correct .ps syntax

Hi Paul,

At 22:58 21/03/2009 +0945, Paul Smedley wrote:

>It seems the IBM postscript code embeds the font in the postscript code, 
>but ghostscript doesn't see it correctly.
>
>Wondering if anyone can provide some good links to some postscript code 
>references, so I can try and fix the driver?

The classic reference is the Adobe PostScript Language Reference Manual 
(Red Book) which is available on the adobe web site. Unfortunately Adobe 
move everything on this site frequently, so links rapidly go stale, however 
this seems to work today:

http://partners.adobe.com/public/developer/en/ps/psrefman.pdf

The Green and Blue books are also available, Don Lancaster's site is useful 
(http://www.tinaja.com/post01.asp)

>A sample file that contains just a few equation characters, but renders as 
>am empty page in gsview is at http://smedley.info/test.ps

I looked at the job, and I can't see any equation characters in it. It uses 
TimesNewRoman to write 'log25' and does some clipping, doesn't seem to do 
much else. Also the job renders blank, because all the x/y values are 
rather too large, given the CTM in force at the time. This is true for GS 
and Acrobat Distiller. In effect all the objects are drawn off the page or 
clipped out.

I could be mistaken of course, I haven't taken the time to debug the file 
thoroughly.

                                     Ken

headers

Paul Smedley | 23 Mar 2009 01:57

Re: Tips on correct .ps syntax

Hi Ken,

On Sat, 21 Mar 2009 15:55:01 UTC, Ken Sharp <ken.sharp <at> artifex.com> 
wrote:

> At 22:58 21/03/2009 +0945, Paul Smedley wrote:
> 
> 
> >It seems the IBM postscript code embeds the font in the postscript code, 
> >but ghostscript doesn't see it correctly.
> >
> >Wondering if anyone can provide some good links to some postscript code 
> >references, so I can try and fix the driver?
> 
> The classic reference is the Adobe PostScript Language Reference Manual 
> (Red Book) which is available on the adobe web site. Unfortunately Adobe 
> move everything on this site frequently, so links rapidly go stale, however 
> this seems to work today:
> 
> http://partners.adobe.com/public/developer/en/ps/psrefman.pdf
> 
> The Green and Blue books are also available, Don Lancaster's site is useful 
> (http://www.tinaja.com/post01.asp)
Thanks for the links ;)

> >A sample file that contains just a few equation characters, but renders as 
> >am empty page in gsview is at http://smedley.info/test.ps
> 
> I looked at the job, and I can't see any equation characters in it. It uses 
> TimesNewRoman to write 'log25' and does some clipping, doesn't seem to do 
> much else. Also the job renders blank, because all the x/y values are 
> rather too large, given the CTM in force at the time. This is true for GS 
> and Acrobat Distiller. In effect all the objects are drawn off the page or 
> clipped out.
When I say equation characters I really mean that any document created
using the equation editor in Lotus Wordpro is likely to show the 
problem with blank characters.  In Wordpro at least, these equations 
are objects in the document, so maybe this is where the code is 
getting screwed up in translating that to a position on the page in 
the postscript code.

Can you point me to the too large x/y values? I assume CTM = 
cleartomark?

> I could be mistaken of course, I haven't taken the time to debug the file 
> thoroughly.
You've given me some more to look at - thanks!
--

-- 
Cheers,

Paul.

headers

Ken Sharp | 23 Mar 2009 09:16

Re: Re: Tips on correct .ps syntax

At 00:57 23/03/2009 +0000, Paul Smedley wrote:

> > I looked at the job, and I can't see any equation characters in it. It 
> uses
> > TimesNewRoman to write 'log25' and does some clipping, doesn't seem to do
> > much else. Also the job renders blank, because all the x/y values are
> > rather too large, given the CTM in force at the time. This is true for GS
> > and Acrobat Distiller. In effect all the objects are drawn off the page or
> > clipped out.
>When I say equation characters I really mean that any document created
>using the equation editor in Lotus Wordpro is likely to show the
>problem with blank characters.

Oh, right, I was looking for math symbols 

>   In Wordpro at least, these equations
>are objects in the document, so maybe this is where the code is
>getting screwed up in translating that to a position on the page in
>the postscript code.

Tha would seem very likely, I could tell better if there were some 
non-equation stuff in the file too.

>Can you point me to the too large x/y values?

Well pretty much all of them really, here are some examples:

600 7676 m
857 7676 l
857 7822 l
600 7822 l cp
clip n d2w

...
1440 18514 m
(l) t

If the CTM (see below) was 'properly' scaled, these could be valid numbers, 
but at the time these are executed the CTM is ~ [1 0 0 0.5 0 0], which 
means the x co-ordinates are as you see them, and the y co-ordinates are 
divided by 2. Note that the device-specific CUPS code sets the resolution 
to non-square [720 360]. It looks like the PostScript job expects a square 
600 dpi resolution though:

/dpi 600 def /dpiY 600 def 0 0 translate 72 dpi div 72 dpiY div scale
/mdevice matrix currentmatrix def

Since PS nominally uses 72 points to the inch, this means that the 
rectangular clip path is positioned at ~ lower left = 600,3836 upper right 
= 857, 3911 or 8.3, 53.2 and 11.9, 54.3 inches. The 't' is located at 1440, 
9257 or 20,128.5 inches. All are well of the physical media (unless you use 
very large paper 

>I assume CTM =
>cleartomark?

Nope, the Current Transformation Matrix. This underlies the whole 
PostScript imaging model, its the way user space (which is very large and 
effectively infinitely sub-divisible) is mapped to device space (which has 
discrete pixels). Because PS uses a matrix, you can achieve all kinds of 
useful effects with it.

> > I could be mistaken of course, I haven't taken the time to debug the file
> > thoroughly.
>You've given me some more to look at - thanks!

Since you mentioned above that the equations are separate objects, it looks 
to me like the 'w2d' and 'd2w' routines are possibly where the problem 
arises. w2d creates a 'mxworld' matrix, and sets the CTM to the identity 
([1 0 0 1 0 0] d2w uses the matrix created in w2d. The names hint at 
transforming between two 'spaces'.

If you can create a PostScript file which contains a small amount of 
working text as well as the missing 'equations' it might be easier to tell 
what's happening.

                             Ken

headers

Paul Smedley | 23 Mar 2009 09:49

Re: Re: Tips on correct .ps syntax

Hi Ken,

On Mon, 23 Mar 2009 08:16:32 UTC, Ken Sharp <ken.sharp <at> artifex.com> 
wrote:

> At 00:57 23/03/2009 +0000, Paul Smedley wrote:
> Oh, right, I was looking for math symbols 

Sorry :)

<lots of really useful stuff snipped but saved :) >

> If you can create a PostScript file which contains a small amount of 
> working text as well as the missing 'equations' it might be easier to tell 
> what's happening.

My pleasure - http://smedley.info/test.jpg shows a screen capture of 
what it _should_ look like.

http://smedley.info/test.ps is the generated postscript

Even the text that does appear is much further down the page than it 
should be - and the TimesNewRoman in the 'equation' doesn't appear.  
Before my latest fixes, the equation symbols didn't appear either :)

Thanks again for your input, I just need to digest it now! :)

--

-- 
Cheers,

Paul.

6	May 2013
24	April 2013
6	March 2013
18	February 2013
20	January 2013
10	December 2012
10	November 2012
30	October 2012
14	September 2012
17	August 2012
17	July 2012
36	June 2012
9	May 2012
11	April 2012
29	March 2012
17	February 2012
13	January 2012
10	December 2011
16	November 2011
18	October 2011
12	September 2011
19	August 2011
7	July 2011
11	June 2011
30	May 2011
25	April 2011
27	March 2011
24	February 2011
9	January 2011
22	December 2010
5	November 2010
5	October 2010
36	September 2010
13	August 2010
7	July 2010
12	June 2010
7	May 2010
20	April 2010
32	March 2010
8	February 2010
2	January 2010
13	December 2009
28	November 2009
35	October 2009
25	September 2009
13	August 2009
38	July 2009
39	June 2009
107	May 2009
53	April 2009
39	March 2009
85	February 2009
148	January 2009
17	December 2008
9	November 2008
12	October 2008
51	September 2008
47	August 2008
41	July 2008
10	June 2008
23	May 2008
13	April 2008
20	March 2008
14	February 2008
25	January 2008
14	December 2007
84	November 2007
38	October 2007
25	September 2007
57	August 2007
36	July 2007
29	June 2007
79	May 2007
38	April 2007
12	March 2007
10	February 2007
13	January 2007
2	November 2006
4	October 2006
4	September 2006
1	August 2006
5	July 2006
7	June 2006
9	May 2006
9	April 2006
10	March 2006
17	February 2006
15	January 2006
9	December 2005
8	November 2005
7	October 2005
9	September 2005
7	August 2005
3	July 2005
1	June 2005
15	May 2005
13	April 2005
19	March 2005
17	February 2005
15	January 2005
21	December 2004
23	November 2004
13	October 2004
11	September 2004
20	August 2004
10	July 2004
10	June 2004
19	May 2004
18	April 2004
40	March 2004
59	February 2004
41	January 2004
59	December 2003
62	November 2003
25	October 2003
14	September 2003
38	August 2003
24	July 2003
43	June 2003
42	May 2003
74	April 2003
52	March 2003
43	February 2003
47	January 2003
54	December 2002
83	November 2002
91	October 2002
60	September 2002
63	August 2002
55	July 2002
34	June 2002

Mon	Tue	Wed	Thu	Fri	Sat	Sun

						1
2	3	4	5	6	7	8
9	10	11	12	13	14	15
16	17	18	19	20	21	22
23	24	25	26	27	28	29
30	31