Re: [tex4ht] Huge update of tex4ht-fonts-noncjk.tex

[Karl Berry]

for example I've tried to make fonts for `LGR` fontenc 

Looks like Eitan defined some lgr-related aliases for libertine, e.g.,
tex4ht/ht-fonts/alias/libertine/fxlri-xl-lgr.htf
contains the two lines:
---
.fxlr-xl-lgr
htfcss: fxlri-xl-lgrs font-style:italic;
---

So maybe other fonts can point to .fxlr-xl-lgr?

Also, FWIW, the gfs fonts (gfsdidot, gfsbodoni, etc.) support LGR, also
cbfonts.  They don't have htf files, but maybe htfgen could actually
work.  I know a couple of good people who can answer Greek-specific
questions (but know nothing in particular about tex4ht), if they can help.

Thanks,
Karl

Re: [tex4ht] [bug #254] tex4ht breaks

[Michal Hoftich]

> Thanks for the feedback. As for the citekeys – they’re perfectly valid for 
> bibtex/biber. I don’t say that because I don’t want to change them, but in an 
> ideal world tex4ht shouldn’t bother with them (and they shouldn’t appear 
> anywhere in the final document).
>

they form good unique identifier for link anchors. but maybe something
else should be used, as in other cross-reference stuff in tex4ht. and
as I said, I don't have access to Biblatex 3 and current configuration
for tex4ht, so I don't know whether these identifiers are currently
used.

I've find source of some problems: in biblatex fiwi style, this is
used for printing urls:

\DeclareFieldFormat{url}{$\langle$\url{#1}$\rangle$}

add this to the my.cfg file to avoid math (I know these aren't exactly
the same angles)

\DeclareFieldFormat{url}{\textless\url{#1}\textgreater}

Then some really strange problem:

After the end of document, lot of numbers appear:

  53 63 73 83 93 103 114
 135 145 155 175 185 195 205 215 246
409 419 429 439 449 4694810 49105111 5411 5511 5611
   5812 6012  6213
6413 6713  6814
 7027 7627   7930 8130
...

I have no idea where they come from, but it clearly causes problem:

System call: java -classpath /usr/local/texlive/2014/texmf-dist/tex4ht/bin/tex4h
t.jar xtpipes -i /usr/local/texlive/2014/texmf-dist/tex4ht/xtpipes/ -o dissschue
renlatex.4oo dissschuerenlatex.tmp
org.xml.sax.SAXParseException; lineNumber: 16852; columnNumber: 3; Content is no
t allowed in trailing section.

Best,
Michal

Re: [tex4ht] [bug #254] tex4ht breaks

[Simon Spiegel]

> On 08.06.2015, at 15:58, Michal Hoftich  wrote:
> 
> I don't have new TL yet, but I've got some validation errors with your 
> example:
> 
> - urls in bibliography are inserted as math, this needs to be fixed on
> tex4ht side
> - some citekeys contains invalid characters (it is `*` in particular,
> you may replace it with something safe, such as `_` for quick fix)

Thanks for the feedback. As for the citekeys – they’re perfectly valid for 
bibtex/biber. I don’t say that because I don’t want to change them, but in an 
ideal world tex4ht shouldn’t bother with them (and they shouldn’t appear 
anywhere in the final document).

Best

Simon

--
Simon Spiegel
Steinhaldenstr. 50
8002 Zürich

Telephon: ++41 44 451 5334
Mobophon: ++41 76 459 6039


http://www.simifilm.ch
http://www.utopia2016.ch

„Yesterday is history. Tomorrow is a mystery. Today is a gift. That is why they 
call it the present.“

Re: [tex4ht] [bug #254] tex4ht breaks

[Michal Hoftich]

I don't have new TL yet, but I've got some validation errors with your example:

- urls in bibliography are inserted as math, this needs to be fixed on
tex4ht side
- some citekeys contains invalid characters (it is `*` in particular,
you may replace it with something safe, such as `_` for quick fix)

Michal

On Mon, Jun 8, 2015 at 3:24 PM, Simon Spiegel  wrote:
> Follow-up Comment #5, bug #254 (project tex4ht):
>
> Ok, I have now a reduced version of my project which shows the problem. It's
> still quite big, but when I reduced it more, the problem didn't show up
> anymore. You can get the ZIP file here:
> https://dl.dropboxusercontent.com/u/35828/Test.zip
>
> Something strange on top: The problem only shows when I run biber first. So
> the problem shows up with the following sequence:
>
> `mk4ht oolatex dissschuerenlatex.tex my`
> `biber dissschuerenlatex`
> `mk4ht oolatex dissschuerenlatex.tex my?
>
> After the first mk4ht run, I get a working .odt file, after the second run, no
> new file is produced.
>
> Notice that I include several files with \include from my master file. If I
> comment the last one (Kapitel/philosophisches.tex) out, everything works fine.
> Even when I run biber.
>
> ___
>
> Reply to this item at:
>
>   
>
> ___
>   Message sent via/by Puszcza
>   http://puszcza.gnu.org.ua/
>

[tex4ht] [bug #254] tex4ht breaks

[Simon Spiegel]

Follow-up Comment #5, bug #254 (project tex4ht):

Ok, I have now a reduced version of my project which shows the problem. It's
still quite big, but when I reduced it more, the problem didn't show up
anymore. You can get the ZIP file here:
https://dl.dropboxusercontent.com/u/35828/Test.zip

Something strange on top: The problem only shows when I run biber first. So
the problem shows up with the following sequence:

`mk4ht oolatex dissschuerenlatex.tex my`
`biber dissschuerenlatex`
`mk4ht oolatex dissschuerenlatex.tex my?

After the first mk4ht run, I get a working .odt file, after the second run, no
new file is produced.

Notice that I include several files with \include from my master file. If I
comment the last one (Kapitel/philosophisches.tex) out, everything works fine.
Even when I run biber. 

___

Reply to this item at:

  

___
  Message sent via/by Puszcza
  http://puszcza.gnu.org.ua/

Re: [tex4ht] [bug #254] tex4ht breaks

[Simon Spiegel]

Hi Bill,

> On 08.06.2015, at 14:05, William F Hammond  wrote:
> 
> Simon --
> 
> You write:
> 
>> It's a PhD thesis in film studies, so there's not any math
>> involved (at least not to my knowledge).
> 
> 1.  The log seems to suggest that TeX's math mode is somehow
> involved.  For example, in the source is math mode used for
> text super- or sub-script construction?

Yes, superscript is involved , mainly for footnote marks,
> 
> 2.  You say that mk4ht xhmlatex runs smoothly.  But have you
> validated the resulting xhtml file?  Or even run xmlwf (chez expat)
> on it?

I can open the file with a browser, but validating it with validator.w3.org 
gives me lots of errors (though from a first glance most of them don’t look 
really serious).

Best

Simon

--
Simon Spiegel
Steinhaldenstr. 50
8002 Zürich

Telephon: ++41 44 451 5334
Mobophon: ++41 76 459 6039


http://www.simifilm.ch
http://www.utopia2016.ch

„Yesterday is history. Tomorrow is a mystery. Today is a gift. That is why they 
call it the present.“

Re: [tex4ht] [bug #254] tex4ht breaks

[William F Hammond]

Michal --

You write:

>> Usually the problem can be found in the temporary xml
>> file (content.xml, perhaps) using something like xmllint,
>> and often the problem is a misconstructed piece of xml,
>> which in some cases is easy to correct by hand.
>
> maybe we should modify the oolatex script to invoke
> xmllint or tidy before it starts extracting code with
> xtpipes. but it would mean that users need to install
> these programs.

"xmlwf" (expat) might be sufficient in most cases and is not
much of an installation burden.

-- Bill

Re: [tex4ht] [bug #254] tex4ht breaks

[William F Hammond]

Simon --

You write:

> It's a PhD thesis in film studies, so there's not any math
> involved (at least not to my knowledge).

1.  The log seems to suggest that TeX's math mode is somehow
involved.  For example, in the source is math mode used for
text super- or sub-script construction?

2.  You say that mk4ht xhmlatex runs smoothly.  But have you
validated the resulting xhtml file?  Or even run xmlwf (chez expat)
on it?

-- Bill

Re: [tex4ht] Huge update of tex4ht-fonts-noncjk.tex

[Michal Hoftich]

> it seems like a good idea, for example `modern
>>

sorry, I accidentally sent the mail too soon.

modern fonts are included in their own file, we may cut the huge
tex4ht-fonts-noncjk.tex into many smaller files.

>> As for other support, it seems nothing comes easy.  For instance, tfm
>> names *-ot1.tfm (LaTeX name) should correspond to *7t.htf (fontname
>> name), but the only non-alias *7t.htf I see are:
>>   ht-fonts/unicode/adobe/courier/pcrr7t.htf - monospace
>>   ht-fonts/unicode/adobe/courier/pcrrc7t.htf - monospace, small caps?
>>   ht-fonts/unicode/adobe/times/ptmrc7t.htf - regular + small caps
>>   ht-fonts/unicode/adobe/times/ptmri7t.htf - italic
>>

it is only important that character codes corresponds, the styling
information is included in the light .htf files. But we may of cource
create new master .htf file using htfgen for some font, verify that
all characters are correct and then reference to it.


>>
>> The Cyrillic encodings *t2a, t2b, t2c, x2, might correspond (somewhat)
>> to existing *z.htf files.
>>
>> Can we not generate .htf files from .enc files?  In the new font
>> packages, there should be an enc file for every user-intended tfm,
>> defined in the .map file in the package.

we can, this is what htfgen does, but we may get too many huge files
as the result. but sometimes it is nearly impossible to find usable
.enc files, for example I've tried to make fonts for `LGR` fontenc and
couldn't find anything usable, the HTML contained different characters
than the source file.


>>
>> Anyway. Just stray thoughts.  Thanks so much for all your work,
>> Karl


Best,

Michal

Re: [tex4ht] Huge update of tex4ht-fonts-noncjk.tex

[Michal Hoftich]

Hi Karl,

>
> Fantastic.  Committed (r150).  I took the liberty of renaming the lua
> and tex to t1fontmake.lua and t1fontsample.tex, hoping that will help
> find them in the future.  I also added most of your email message to
> tex4ht-fonts-noncjk.tex as commentary in that chapter.
>
> I'll install all the new "light" htf files in TL when updates resume there.
>

great :)

> Huge update
>
> Indeed.  It would be nice to split up the huge source files into smaller
> pieces.  It gets hard to figure out what is in them.
>
> For instance, looking at where the \Chapter commands are in noncjk, it
> looks like the "Adobe" chapter deserves its own file.  And your new
> "Converted" chapter could be its own file too.  Maybe ...

it seems like a good idea, for example `modern
>
> As for other support, it seems nothing comes easy.  For instance, tfm
> names *-ot1.tfm (LaTeX name) should correspond to *7t.htf (fontname
> name), but the only non-alias *7t.htf I see are:
>   ht-fonts/unicode/adobe/courier/pcrr7t.htf - monospace
>   ht-fonts/unicode/adobe/courier/pcrrc7t.htf - monospace, small caps?
>   ht-fonts/unicode/adobe/times/ptmrc7t.htf - regular + small caps
>   ht-fonts/unicode/adobe/times/ptmri7t.htf - italic
>
> Why italic?  No idea.  And these might well be specially for the Adobe
> base35 fonts, despite the name.  And since there are no comments
> explaining what the encoding actually is, it will be laborious to
> discover where the differences are.
>
> The Cyrillic encodings *t2a, t2b, t2c, x2, might correspond (somewhat)
> to existing *z.htf files.
>
> Can we not generate .htf files from .enc files?  In the new font
> packages, there should be an enc file for every user-intended tfm,
> defined in the .map file in the package.
>
> Anyway. Just stray thoughts.  Thanks so much for all your work,
> Karl

Re: [tex4ht] [bug #254] tex4ht breaks

[Michal Hoftich]

On Mon, Jun 8, 2015 at 11:01 AM, Harri Kiiskinen  wrote:
> This sound very familiar, and is a problem I've always had with teX4HT: the 
> compilation to odt fails at the last stage with the construction of the final 
> file.
>
> Usually the problem can be found in the temporary xml file (content.xml, 
> perhaps) using something like xmllint, and often the problem is a 
> misconstructed piece of xml, which in some cases is easy to correct by hand.

maybe we should modify the oolatex script to invoke xmllint or tidy
before it starts extracting code with xtpipes. but it would mean that
users need to install these programs.

Best regards,
Michal

Re: [tex4ht] [bug #254] tex4ht breaks

[Harri Kiiskinen]

This sound very familiar, and is a problem I've always had with teX4HT: the 
compilation to odt fails at the last stage with the construction of the final 
file.

Usually the problem can be found in the temporary xml file (content.xml, 
perhaps) using something like xmllint, and often the problem is a 
misconstructed piece of xml, which in some cases is easy to correct by hand.

I usually try to solve the problem by removing as much of the packages I 
possibly can, especially the font packages are often a problem, and a 
compilation using the default fonts has a better chance of working.

Best,

Harri Kiiskinen

> -Original Message-
> From: tex4ht [mailto:tex4ht-boun...@tug.org] On Behalf Of Simon Spiegel
> Sent: 8. kesäkuuta 2015 10:30
> To: CV Radhakrishnan; Michal Hoftich; Simon Spiegel; tex4ht@tug.org
> Subject: [tex4ht] [bug #254] tex4ht breaks
> 
> Follow-up Comment #4, bug #254 (project tex4ht):
> 
> It's a PhD thesis in film studies, so there's not any math involved (at least 
> not to my
> knowledge).
> 
> I'll see whether I can post it. It's really quite a beast, with lots of 
> figures, a huge .bib
> file and other stuff.
> 
> ___
> 
> Reply to this item at:
> 
>   
> 
> ___
>   Message sent via/by Puszcza
>   http://puszcza.gnu.org.ua/

[tex4ht] [bug #254] tex4ht breaks

[Simon Spiegel]

Follow-up Comment #4, bug #254 (project tex4ht):

It's a PhD thesis in film studies, so there's not any math involved (at least
not to my knowledge).

I'll see whether I can post it. It's really quite a beast, with lots of
figures, a huge .bib file and other stuff.

___

Reply to this item at:

  

___
  Message sent via/by Puszcza
  http://puszcza.gnu.org.ua/