I keep running into errors that look like this: ! LuaTeX error ...imal/tex/texmf-context/tex/context/base/lpdf-ini.lua:345: pdf.immediateobj() object in use I can't find anything specific that's causing them. For example: input{chapter1} input{chapter2} fails, but: %input{chapter1} input{chapter2} and input{chapter1} %input{chapter2} work. Or, I may make a small change, get the error, reverse the change, and continue to get the error. Maybe I need to allocate more memory somehow? what's wrong?
Michael Saunders wrote:
Or, I may make a small change, get the error, reverse the change, and continue to get the error. Maybe I need to allocate more memory somehow? what's wrong?
This problem we have seen before, and I think it was a luatex bug then. Which luatex is this ? 0.60.1 ? Also, debugging this really needs a complete test input file. Best wishes, Taco
This problem we have seen before, and I think it was a luatex bug then. Which luatex is this ? 0.60.1 ?
It was .50, so I updated to .60. I'm still getting errors with the same behavior, e.g.: % block of text block of text and block of text % block of text work, but block of text block of text doesn't. The difference is that now they are reporting differently: " ! Undefined control sequence. <argument> \@@bklocation \processaction ...-> \expandaction \!!stringa{#1 } \ifx \!!stringa \empty \l..." In addition, everything involving inferiors, superiors, numerators and denominators is now broken. For example: \definefontfeature[in][body][sinf=yes]%inferior Cs{\setff{in}133} has no effect.
Also, debugging this really needs a complete test input file.
I don't know how to make one.
Michael Saunders wrote:
In addition, everything involving inferiors, superiors, numerators and denominators is now broken. For example:
\definefontfeature[in][body][sinf=yes]%inferior Cs{\setff{in}133}
has no effect.
This depends on whether the font supports the sinf feature. The default fonts do not.
Also, debugging this really needs a complete test input file.
I don't know how to make one.
Well, if I literally run \starttext block of text block of text \stoptext It works just fine, so I need a test file to debug your bug report. Best wishes, Taco
In addition, everything involving inferiors, superiors, numerators and denominators is now broken. For example:
\definefontfeature[in][body][sinf=yes]%inferior Cs{\setff{in}133}
has no effect.
This depends on whether the font supports the sinf feature. The default fonts do not.
Mine does. It worked in 0.50. Maybe the name of the feature "sinf" changed?
Well, if I literally run
\starttext block of text block of text \stoptext
It works just fine, so I need a test file to debug your bug report.
It appears to depend on having a big enough, complicated enough text, not a specific combination of commands, so it's hard to come up with one, but I'll try. Thanks.
Michael Saunders wrote:
In addition, everything involving inferiors, superiors, numerators and denominators is now broken. For example:
\definefontfeature[in][body][sinf=yes]%inferior Cs{\setff{in}133}
has no effect.
This depends on whether the font supports the sinf feature. The default fonts do not.
Mine does. It worked in 0.50. Maybe the name of the feature "sinf" changed?
How is the [body] feature defined?
How is the [body] feature defined?
\definefontfeature[body][default][onum=yes,pnum=yes,calt=yes,protrusion=quality,expansion=quality] \definefontfeature[in][body][sinf=yes]%inferior \definefontfeature[su][body][sups=yes,ordn=yes]%superior \definefontfeature[nu][body][numr=yes]%numerator \definefontfeature[de][body][dnom=yes]%denominator in{\addff{in}123} su{\addff{su}123abc} nu{\addff{nu}123} de{\addff{de}123} The numerals all come out as proportional oldstyle, but the letters work. So, sinf, sups, numr, and dnom are failing but ordn works. I was thinking---besides updating to 0.60 yesterday, I was also trying to test Stix fonts, so I ran: mtxrun --script fonts --reload luatools --generate Maybe a change in mtxrun or luatools broke these features? About my other problem---the strange one that's hard to isolate, and which appeared this way in 0.50: ! LuaTeX error ...imal/tex/texmf-context/tex/context/base/lpdf-ini.lua:345: pdf.immediateobj() object in use and this way in 0.60: " ! Undefined control sequence. <argument> \@@bklocation \processaction ...-> \expandaction \!!stringa{#1 } \ifx \!!stringa \empty \l..." I'm still having a very hard time isolating a short example---e.g., I spent the evening discovering this: near the end of about seven pages of typeset text, I have the following string in my input file: "become more obvious as we examine the phenomena" This works: become more obvious as %we examine the phenomena and this fails: become more obvious as we %examine the phenomena I know the word "we" isn't the problem. Maybe it has something to do with the length of the text, but that doesn't seem quite right either.
On 30-5-2010 1:24, Michael Saunders wrote:
" ! Undefined control sequence. <argument> \@@bklocation \processaction ...-> \expandaction \!!stringa{#1 } \ifx \!!stringa \empty \l..."
in str-flt.mkiv there is a section \def\@@bknumbering {\floatsharedparameter\c!numbering } % global one \def\@@bkspaceafter {\floatsharedparameter\c!spaceafter } % global one \def\@@bkspacebefore{\floatsharedparameter\c!spacebefore} % global one \def\@@bknbottom {\floatsharedparameter\c!nbottom } % global one \def\@@bkntop {\floatsharedparameter\c!ntop } % global one \def\@@bknlines {\floatsharedparameter\c!nlines } % global one \def\@@bkmargin {\floatsharedparameter\c!margin } % global one \def\@@bkcache {\floatsharedparameter\c!cache } % global one can you add: \def\@@bklocation {\floatsharedparameter\c!location } and then remake the formats the bk namespace has been replaced (and this section is a temporary solution in the transition) Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
Thanks, I added the line to the \@@bk block in str-flt.mkiv, but I
don't know how to remake the formats. I tried:
mtxrun --script fonts --reload
luatools --generate
texexec --make --all
texexec --make --all --pdftex
texexec --make --all --xetex
luatools --generate
luatools --ini --compile --verbose cont-en
none of which had any effect on the output. How do I remake the formats?
On Sun, May 30, 2010 at 6:35 AM, Hans Hagen
On 30-5-2010 1:24, Michael Saunders wrote:
" ! Undefined control sequence. <argument> \@@bklocation \processaction ...-> \expandaction \!!stringa{#1 } \ifx \!!stringa \empty \l..."
in str-flt.mkiv there is a section
\def\@@bknumbering {\floatsharedparameter\c!numbering } % global one \def\@@bkspaceafter {\floatsharedparameter\c!spaceafter } % global one \def\@@bkspacebefore{\floatsharedparameter\c!spacebefore} % global one \def\@@bknbottom {\floatsharedparameter\c!nbottom } % global one \def\@@bkntop {\floatsharedparameter\c!ntop } % global one \def\@@bknlines {\floatsharedparameter\c!nlines } % global one \def\@@bkmargin {\floatsharedparameter\c!margin } % global one \def\@@bkcache {\floatsharedparameter\c!cache } % global one
can you add:
\def\@@bklocation {\floatsharedparameter\c!location }
and then remake the formats
the bk namespace has been replaced (and this section is a temporary solution in the transition)
Hans
----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
Please add script=latn, that works for Palatino Nova, and I think there possibly was in change in that a while back (some months).
Thanks, I have added script=latn to \definefontfeature[body], but those Open Type features still aren't back to normal. I should have mentioned that I had: \definefontfeature[default][mode=node,language=dflt,script=latn,kern=yes,liga=yes,tlig=yes,trep=yes] \definefontfeature[body][default][onum=yes,pnum=yes,calt=yes,protrusion=quality,expansion=quality] But I seem to remember when I set this up that argument 1 never actually inherited from argument 2 when I tested it.
Don't be too sure they are related, they could be totally independent problems.
Ahh, of course you are right. I see that:
" ! Undefined control sequence. <argument> \@@bklocation...
must be related to the missing line from str-flt.mkiv. I have added it, though, and I _think_ I remade the formats, but I get no change in behavior.
Michael Saunders wrote:
Please add script=latn, that works for Palatino Nova, and I think there possibly was in change in that a while back (some months).
Thanks, I have added script=latn to \definefontfeature[body], but those Open Type features still aren't back to normal. I should have mentioned that I had: \definefontfeature[default][mode=node,language=dflt,script=latn,kern=yes,liga=yes,tlig=yes,trep=yes] \definefontfeature[body][default][onum=yes,pnum=yes,calt=yes,protrusion=quality,expansion=quality]
There is no test file, and I still do not know which font you are using. Attempting to find the problem is closer to tea-gazing than to debugging at the moment. Font features cannot be debugged independent of the actual font. All I can say is that your original example, after addition of 'script=latn', works with Palatino Nova (which is the only font I have that supports those features)
must be related to the missing line from str-flt.mkiv. I have added it, though, and I _think_ I remade the formats, but I get no change in behavior.
The context --make en should have taken care of that, but you can check: the third or fourth line in the log file is something like this: ConTeXt ver: 2010.05.30 17:08 MKIV fmt: 2010.5.30 int: english/english The date after fmt: is the date on which the format was generated. Best wishes, Taco
There is no test file, and I still do not know which font you are using. Attempting to find the problem is closer to tea-gazing than to debugging at the moment. Font features cannot be debugged independent of the actual font.
Thanks, I understand. I had been using GaramondPremrPro (which has all four features and worked with 0.50). You gave me the idea to try GillSansMTPro, which also has the four features. I got a very interesting result---for it, numr and dnom work, but sinf and sups do not. I don't know how I could prepare a test file for you short of sending you my fonts and typescripts (which I'd be willing to do).
On 30-5-2010 10:36, Michael Saunders wrote:
There is no test file, and I still do not know which font you are using. Attempting to find the problem is closer to tea-gazing than to debugging at the moment. Font features cannot be debugged independent of the actual font.
Thanks, I understand. I had been using GaramondPremrPro (which has all four features and worked with 0.50). You gave me the idea to try GillSansMTPro, which also has the four features. I got a very interesting result---for it, numr and dnom work, but sinf and sups do not. I don't know how I could prepare a test file for you short of sending you my fonts and typescripts (which I'd be willing to do).
you can try mtxrun --script font --list --info garamondpremrpr or context --trackers=otf.features yourfile to see what happens with features ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
you can try
mtxrun --script font --list --info garamondpremrpr ... to see what happens with features
Thanks. This seems to confirm that the features are there: MTXrun | dnom cyrl dflt MTXrun | grek dflt MTXrun | latn aze crt deu dflt fra ita mol rom srb trk MTXrun | numr cyrl dflt MTXrun | grek dflt MTXrun | latn aze crt deu dflt fra ita mol rom srb trk MTXrun | ordn cyrl dflt MTXrun | grek dflt MTXrun | latn aze crt deu dflt fra ita mol rom srb trk MTXrun | sinf cyrl dflt MTXrun | grek dflt MTXrun | latn aze crt deu dflt fra ita mol rom srb trk MTXrun | sups cyrl dflt MTXrun | grek dflt MTXrun | latn aze crt deu dflt fra ita mol rom srb trk but still, \definefontfeature[body][default][script=latn,onum=yes,pnum=yes,calt=yes,protrusion=quality,expansion=quality] \definefontfeature[in][body][sinf=yes]%inferior \definefontfeature[su][body][sups=yes,ordn=yes]%superior \definefontfeature[nu][body][numr=yes]%numerator \definefontfeature[de][body][dnom=yes]%denominator in {\addff{in}123} su {\addff{su}123} nu {\addff{nu}123} de {\addff{de}123} produces only proportional oldstyle figures.
On 30-5-2010 11:24, Michael Saunders wrote:
you can try
mtxrun --script font --list --info garamondpremrpr .... to see what happens with features
Thanks. This seems to confirm that the features are there:
MTXrun | dnom cyrl dflt MTXrun | grek dflt MTXrun | latn aze crt deu dflt fra ita mol rom srb trk MTXrun | numr cyrl dflt MTXrun | grek dflt MTXrun | latn aze crt deu dflt fra ita mol rom srb trk MTXrun | ordn cyrl dflt MTXrun | grek dflt MTXrun | latn aze crt deu dflt fra ita mol rom srb trk MTXrun | sinf cyrl dflt MTXrun | grek dflt MTXrun | latn aze crt deu dflt fra ita mol rom srb trk MTXrun | sups cyrl dflt MTXrun | grek dflt MTXrun | latn aze crt deu dflt fra ita mol rom srb trk
but still,
\definefontfeature[body][default][script=latn,onum=yes,pnum=yes,calt=yes,protrusion=quality,expansion=quality] \definefontfeature[in][body][sinf=yes]%inferior \definefontfeature[su][body][sups=yes,ordn=yes]%superior \definefontfeature[nu][body][numr=yes]%numerator \definefontfeature[de][body][dnom=yes]%denominator in {\addff{in}123} su {\addff{su}123} nu {\addff{nu}123} de {\addff{de}123}
produces only proportional oldstyle figures.
add mode=node to the 'body' feature definitions Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
On 31-5-2010 12:08, Michael Saunders wrote:
add mode=node to the 'body' feature definitions
I added it. I get the same results.
you can try \usemodule[fnt-20] \setvariables [otftracker] [font=somefont, features=default, size=48pt, sample=your short text] and see what is applied ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
On 30-5-2010 11:24, Michael Saunders wrote:
\definefontfeature[body][default][script=latn,onum=yes,pnum=yes,calt=yes,protrusion=quality,expansion=quality] \definefontfeature[in][body][sinf=yes]%inferior \definefontfeature[su][body][sups=yes,ordn=yes]%superior \definefontfeature[nu][body][numr=yes]%numerator \definefontfeature[de][body][dnom=yes]%denominator in {\addff{in}123} su {\addff{su}123} nu {\addff{nu}123} de {\addff{de}123}
produces only proportional oldstyle figures.
i have a fix for it but first idris has to do some testing with his fonts as this mechanism was primarily made for very complex setups and i want to make sure that nothing breaks .. so no beta today, maybe later this week Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
Michael Saunders wrote:
How is the [body] feature defined?
\definefontfeature[body][default][onum=yes,pnum=yes,calt=yes,protrusion=quality,expansion=quality]
Please add script=latn, that works for Palatino Nova, and I think there possibly was in change in that a while back (some months).
About my other problem---the strange one that's hard to isolate, and which appeared this way in 0.50:
! LuaTeX error ...imal/tex/texmf-context/tex/context/base/lpdf-ini.lua:345: pdf.immediateobj() object in use
and this way in 0.60:
" ! Undefined control sequence. <argument> \@@bklocation \processaction ...-> \expandaction \!!stringa{#1 } \ifx \!!stringa \empty \l..."
Don't be too sure they are related, they could be totally independent problems. Best wishes, Taco
participants (4)
-
Hans Hagen -
Michael Saunders -
taco -
Taco Hoekwater