docutils-users Mailing List for Docutils: Documentation Utilities

rst2odt.py/odtwriter now has the ability to generate
ODF/oowriter table of contents.  This is enabled by
default.  If you want the old style (list) table of contents,
use the --generate-list-toc.

Since rst2odt.py/odtwriter cannot calculate page
numbers needed in an ODF/oowriter-style TOC,
you will need to open a generated
document in oowriter, right-click on the TOC,
and do an update.

There is a bit of explanation in the "Table of contents"
section of odt.txt (odt.html) in the distribution.

If you have comments, suggestions, or problems,
please let me know.

- Dave

 -- 


Dave Kuhlman
http://www.rexx.com/~dkuhlman

Looks like a sensible default.

Roberto Alsina <ra...@ne...> wrote:

>
>I am trying to improve how rst2pdf displays lists and I need some opinions:
>
>Does this look ok?
>
>http://twitpic.com/619qi
>
>For the record, here's what rst2latex does with it:
>
>http://twitpic.com/619z9
>
>And here's what rst2html does:
>
>http://twitpic.com/61a13
>
>-- 
> ("\''/").__..-''"`-. .         Roberto Alsina
> `9_ 9  )   `-. (    ).`-._.`)  KDE Developer (MFCH)
> (_Y_.)' ._   ) `._`.  " -.-'  http://lateral.netmanagers.com.ar 
>  _..`-'_..-_/ /-'_.'     The 6,855th most popular site of Slovenia   
>(l)-'' ((i).' ((!.'       according to alexa.com (27/5/2007) 
>"Our opponent is an alien starship packed with atomic bombs, I said. 
>We have a protractor. Okay, I’ll go home and see if I can scrounge up a 
>ruler and a piece of string."  — Neal Stephenson
>
>--~--~---------~--~----~------------~-------~--~----~
>You received this message because you are subscribed to the Google Groups "rst2pdf-discuss" group.
>To post to this group, send email to rst...@go...
>To unsubscribe from this group, send email to rst...@go...
>For more options, visit this group at http://groups.google.com/group/rst2pdf-discuss?hl=en
>-~----------~----~----~----~------~----~------~--~---
>

--
Sent from my Android phone with K-9. Please excuse my brevity.

On 2009-05-28, Roberto Alsina wrote:
> On Wednesday 27 May 2009 10:10:41 Guenter Milde wrote:


>> > The only serif guaranteed to be available in PDF is Times :-(

>> I'd suppose all Base35 fonts to be available
...

> That's in postscript. In PDF 1.3 (what reportlab is compatible with) you get 
> 14 (and that counts things like Times-Bold:

> Times ... , Helvetica ... , Courier ... , Symbol, ZapfDingbats

I agree that the default should stick with these (also because of the
smaller file size).  

I'd still prefer a serif font for the text (and as both Times and
Helvetica are shunned by font experts, avoidance of Times does not seem
to be a strong case).

> The rest you need to embed, and that is a whole licensing can of worms.

latex's PSNFSS package can embed free versions of the Base35 fonts. More
free fonts are available.

My suggestion would be to support font-setting from the style sheet and
provide some styles with "nice and free" fonts.
(There is also a stylesheets repository in the Sandbox where you (and
other users) can add rst2pdf style sheets.)

Günter

On Wednesday 27 May 2009 23:43:25 Alan G Isaac wrote:
> http://kb2.adobe.com/cps/322/322792.html
>
>          If you view a PDF file that contains Helvetica or
>          Times, and the font isn't embedded in the PDF file,
>          Acrobat 6.x and 5.x products substitute the font
>          using either MT Arial or MT Times New Roman, even if
>          Helvetica or Times is installed on your system.
>
> fwiw,

Amazing. It's surprising PDF works as well as it does.

-- 
 ("\''/").__..-''"`-. .         Roberto Alsina
 `9_ 9  )   `-. (    ).`-._.`)  KDE Developer (MFCH)
 (_Y_.)' ._   ) `._`.  " -.-'  http://lateral.netmanagers.com.ar 
  _..`-'_..-_/ /-'_.'     The 6,855th most popular site of Slovenia   
(l)-'' ((i).' ((!.'       according to alexa.com (27/5/2007) 
"Our opponent is an alien starship packed with atomic bombs, I said. 
We have a protractor. Okay, I’ll go home and see if I can scrounge up a 
ruler and a piece of string."  — Neal Stephenson

>> On 5/27/2009 9:08 PM Roberto Alsina apparently wrote:
>>> In PDF 1.3 (what reportlab is compatible with) you get 
>>> 14 (and that counts things like Times-Bold:

>>> Times−Roman, Times−Bold, Times−Italic, Times−BoldItalic, Helvetica, 
>>> Helvetica−Bold, Helvetica−Oblique, Helvetica−BoldOblique, Courier, 
>>> Courier−Bold, Courier−Oblique, Courier−BoldOblique, Symbol, ZapfDingbats 

> On Wednesday 27 May 2009 22:34:56 Alan G Isaac wrote:
>> Except that when you ask for Helvetica, you get Arial, 
>> iirc. 

On 5/27/2009 10:24 PM Roberto Alsina apparently wrote:
> Depends on the viewer. On acroread you are supposed to get real Arial, since 
> they own it. Just like on other viewers you get Times New instead of Times. 


http://kb2.adobe.com/cps/322/322792.html

         If you view a PDF file that contains Helvetica or
         Times, and the font isn't embedded in the PDF file,
         Acrobat 6.x and 5.x products substitute the font
         using either MT Arial or MT Times New Roman, even if
         Helvetica or Times is installed on your system.

fwiw,
Alan Isaac

On Wednesday 27 May 2009 23:24:15 Roberto Alsina wrote:
> On Wednesday 27 May 2009 22:34:56 Alan G Isaac wrote:
> > On 5/27/2009 9:08 PM Roberto Alsina apparently wrote:
> > > In PDF 1.3 (what reportlab is compatible with) you get
> > > 14 (and that counts things like Times-Bold:
> > >
> > > Times−Roman, Times−Bold, Times−Italic, Times−BoldItalic, Helvetica,
> > > Helvetica−Bold, Helvetica−Oblique, Helvetica−BoldOblique, Courier,
> > > Courier−Bold, Courier−Oblique, Courier−BoldOblique, Symbol,
> > > ZapfDingbats
> >
> > Except that when you ask for Helvetica, you get Arial, iirc.
>
> Depends on the viewer. On acroread you are supposed to get real Arial,
> since they own it. Just like on other viewers you get Times New instead of
> Times.

I meant real helvetica, of course, but after reading wikipedia I am not sure 
anymore.

-- 
 ("\''/").__..-''"`-. .         Roberto Alsina
 `9_ 9  )   `-. (    ).`-._.`)  KDE Developer (MFCH)
 (_Y_.)' ._   ) `._`.  " -.-'  http://lateral.netmanagers.com.ar 
  _..`-'_..-_/ /-'_.'     The 6,855th most popular site of Slovenia   
(l)-'' ((i).' ((!.'       according to alexa.com (27/5/2007) 
"Our opponent is an alien starship packed with atomic bombs, I said. 
We have a protractor. Okay, I’ll go home and see if I can scrounge up a 
ruler and a piece of string."  — Neal Stephenson

On Wednesday 27 May 2009 22:34:56 Alan G Isaac wrote:
> On 5/27/2009 9:08 PM Roberto Alsina apparently wrote:
> > In PDF 1.3 (what reportlab is compatible with) you get
> > 14 (and that counts things like Times-Bold:
> >
> > Times−Roman, Times−Bold, Times−Italic, Times−BoldItalic, Helvetica,
> > Helvetica−Bold, Helvetica−Oblique, Helvetica−BoldOblique, Courier,
> > Courier−Bold, Courier−Oblique, Courier−BoldOblique, Symbol, ZapfDingbats
>
> Except that when you ask for Helvetica, you get Arial, iirc.

Depends on the viewer. On acroread you are supposed to get real Arial, since 
they own it. Just like on other viewers you get Times New instead of Times.

-- 
 ("\''/").__..-''"`-. .         Roberto Alsina
 `9_ 9  )   `-. (    ).`-._.`)  KDE Developer (MFCH)
 (_Y_.)' ._   ) `._`.  " -.-'  http://lateral.netmanagers.com.ar 
  _..`-'_..-_/ /-'_.'     The 6,855th most popular site of Slovenia   
(l)-'' ((i).' ((!.'       according to alexa.com (27/5/2007) 
"Our opponent is an alien starship packed with atomic bombs, I said. 
We have a protractor. Okay, I’ll go home and see if I can scrounge up a 
ruler and a piece of string."  — Neal Stephenson

On 5/27/2009 9:08 PM Roberto Alsina apparently wrote:
> In PDF 1.3 (what reportlab is compatible with) you get 
> 14 (and that counts things like Times-Bold:
> 
> Times−Roman, Times−Bold, Times−Italic, Times−BoldItalic, Helvetica, 
> Helvetica−Bold, Helvetica−Oblique, Helvetica−BoldOblique, Courier, 
> Courier−Bold, Courier−Oblique, Courier−BoldOblique, Symbol, ZapfDingbats


Except that when you ask for Helvetica, you get Arial, iirc.

Alan Isaac

On Wednesday 27 May 2009 10:10:41 Guenter Milde wrote:
> On 2009-05-27, Roberto Alsina wrote:
> > On Wednesday 27 May 2009 03:36:45 Guenter Milde wrote:
> >> On 2009-05-27, Roberto Alsina wrote:
> >> > I am trying to improve how rst2pdf displays lists and I need some
> >> > opinions:
> >>
> >> * It would be easier to evaluate from the PDF file.
> >
> > http://lateral.netmanagers.com.ar/static/test_lists.pdf (not the same
> > pdf, but a superset), done from this: http://is.gd/GJoq
>
> The indentation of nested lists is too large. In LaTeX, the indentation
> also scales (smaller with increasing level, but I don't remember the
> details).

I'll try to google that.

> >> * To be better comparable, you might use rest2latex with the option
> >>   ``--stylesheet=parskip`` (separate paragraphs by vertical space).
> >
> > I get
> >
> > ! LaTeX Error: File `parskip.tex' not found.
>
> In Docutils version 05 you need to specify the extension of LaTex packages:
> ``--stylesheet=parskip.sty`` In the SVN version the extension defaults to
> "sty".

I think I should switch to trunk already :-D

> >> * I would prefer a bullet size between the rst2pdf and LaTeX examples.
> >
> > Yeah, me too. I am drawing it using the unicode bullet character, which
> > seems to be awful small.
> >
> > Don't have much of a choice unless I start doing graphics as bullets :-(
>
> How about Character '●' (9679, 0x25CF) 25CF	BLACK CIRCLE?
> It might be too large.

Anyway it's not in the default font set :-P
Not only are there few fonts, you are not guaranteed to have any unicode 
coverage whatsoever (what you have depends on what font your PDF viewer is 
actually using, and if it's acroread, on what localized version you have).

> Another option would be to scale the bullet.

I guess. I could also use a SVG bullet. Maybe there is something in Symbol or 
ZapfDingbats.

> >> * Side note: For longer text (especially on paper), I prefer serif
> >> fonts. DejaVu is a free serif font that is also optimised for viewing on
> >> screen.
> >
> > The only serif guaranteed to be available in PDF is Times :-(
>
> I'd suppose all Base35 fonts to be available, including the
> serif fonts Bookman, New Century Schoolbook, Palatino, Times, Utopia.

That's in postscript. In PDF 1.3 (what reportlab is compatible with) you get 
14 (and that counts things like Times-Bold:

Times−Roman, Times−Bold, Times−Italic, Times−BoldItalic, Helvetica, 
Helvetica−Bold, Helvetica−Oblique, Helvetica−BoldOblique, Courier, 
Courier−Bold, Courier−Oblique, Courier−BoldOblique, Symbol, ZapfDingbats

The rest you need to embed, and that is a whole licensing can of worms.

-- 
 ("\''/").__..-''"`-. .         Roberto Alsina
 `9_ 9  )   `-. (    ).`-._.`)  KDE Developer (MFCH)
 (_Y_.)' ._   ) `._`.  " -.-'  http://lateral.netmanagers.com.ar 
  _..`-'_..-_/ /-'_.'     The 6,855th most popular site of Slovenia   
(l)-'' ((i).' ((!.'       according to alexa.com (27/5/2007) 
"Our opponent is an alien starship packed with atomic bombs, I said. 
We have a protractor. Okay, I’ll go home and see if I can scrounge up a 
ruler and a piece of string."  — Neal Stephenson

On Wed, May 27, 2009 at 15:39, Stefan Merten <sm...@oe...> wrote:
> Isn't it possible to make the ``.. only::`` directive have a similar
> semantic?

Probably. Talk to Georg on the Sphinx list; it's not in core Docutils.
Perhaps it should be (along with a lot besides).

-- 
David Goodger <http://python.net/~goodger>

Hi Guenter and all!

Last week (9 days ago) Guenter Milde wrote:
> On 2009-05-15, Stefan Merten wrote:
>> Yesterday David Goodger wrote:
>>> On Wed, May 13, 2009 at 20:11, Michael P. Soulier
>>> <mso...@di...> wrote:
>>>> So, I'm wondering if I can somehow set conditionals in the source
>>>> based on the output type.
>> [...]
>>> Not currently, no. It has been discussed before, but not implemented.
> 
>> I just read the Sphinx manual (BTW: cool!) and there is the ``only``
>> directive::
> 
> ...
> 
>> The only concern I see at the moment is that you can't do something
>> like::
> 
>>   .. only:: html
> 
>>      Title for HTML
>>      ==============
> 
>>   Continue here in all formats.
> 
> Yes, section titles in an rst document can not be nested inside directives.
> This keeps the document tree straight.
> 
> There are several workarounds.
> 
> * For the title, use the "title" directive ``.. title:: ``.
> 
> * For section-like headings (e.g. inside a side-bar) use "rubric".

Yeah, but they break the logic because sections are no longer sections
:-( .

> * For complete conditional sections, try to include them from a file::
> 
>>   .. only:: html
> 
>        .. include:: conditional-html.txt

This would work but it is too cumbersome if you have lots of sections
which need such conditional inclusion. You have to have an own file
for for each section then.

I just looked into this problem from a different angle. This time I
thought about using the ``.. class::`` directive for such a
conditional. When checking through the documentation I was reminded
that the ``.. class::`` directive actually supports::

  .. class:: notInHTML

  Section not for HTML
  ====================

In fact this attaches the class to the whole section - which is
probably what you want with such a construct.

Isn't it possible to make the ``.. only::`` directive have a similar
semantic?


						Grüße

						Stefan

Roberto Alsina wrote:
> http://twitpic.com/619qi
> 
> rst2latex does with it:
> http://twitpic.com/619z9
> 
> rst2html does:
> http://twitpic.com/61a13

This is slightly misleading,
because you use an indented paragraph
in LaTeX but not in the other two.
Also, the choice of font affects how
things look.  Anyway, roughly guessing,
it looks like your indent is a little
bigger than desirable.  I'd also increase
the bullet size slightly.

Cheers,
Alan Isaac

On 2009-05-27, Roberto Alsina wrote:
> On Wednesday 27 May 2009 03:36:45 Guenter Milde wrote:
>> On 2009-05-27, Roberto Alsina wrote:
>> > I am trying to improve how rst2pdf displays lists and I need some
>> > opinions:

>> * It would be easier to evaluate from the PDF file.

> http://lateral.netmanagers.com.ar/static/test_lists.pdf (not the same
> pdf, but a superset), done from this: http://is.gd/GJoq

The indentation of nested lists is too large. In LaTeX, the indentation
also scales (smaller with increasing level, but I don't remember the
details).

>> * To be better comparable, you might use rest2latex with the option
>>   ``--stylesheet=parskip`` (separate paragraphs by vertical space).

> I get 

> ! LaTeX Error: File `parskip.tex' not found.

In Docutils version 05 you need to specify the extension of LaTex packages:
``--stylesheet=parskip.sty`` In the SVN version the extension defaults to
"sty".

>> * I would prefer a bullet size between the rst2pdf and LaTeX examples.

> Yeah, me too. I am drawing it using the unicode bullet character, which
> seems to be awful small.

> Don't have much of a choice unless I start doing graphics as bullets :-(

How about Character '●' (9679, 0x25CF) 25CF	BLACK CIRCLE? 
It might be too large.

Another option would be to scale the bullet.

>> * Side note: For longer text (especially on paper), I prefer serif fonts.
>>   DejaVu is a free serif font that is also optimised for viewing on
>>   screen.

> The only serif guaranteed to be available in PDF is Times :-(

I'd suppose all Base35 fonts to be available, including the
serif fonts Bookman, New Century Schoolbook, Palatino, Times, Utopia.

Günter

On Wednesday 27 May 2009 03:36:45 Guenter Milde wrote:
> On 2009-05-27, Roberto Alsina wrote:
> > I am trying to improve how rst2pdf displays lists and I need some
> > opinions:
> >
> > Does this look ok?
> >
> > http://twitpic.com/619qi
> >
> > For the record, here's what rst2latex does with it:
> >
> > http://twitpic.com/619z9
> >
> > And here's what rst2html does:
> >
> > http://twitpic.com/61a13
>
> * It would be easier to evaluate from the PDF file.

http://lateral.netmanagers.com.ar/static/test_lists.pdf (not the same pdf, but 
a superset), done from this: http://is.gd/GJoq

> * To be better comparable, you might use rest2latex with the option
>   ``--stylesheet=parskip`` (separate paragraphs by vertical space).

I get 

! LaTeX Error: File `parskip.tex' not found.

> * I would prefer a bullet size between the rst2pdf and LaTeX examples.

Yeah, me too. I am drawing it using the unicode bullet character, which seems 
to be awful small.

Don't have much of a choice unless I start doing graphics as bullets :-(

> * Consider supporting the "compact" option and class argument (from
>   rst2html) as some lists look better compact and others need the space.

Yes, that's a good idea, added to the tracker.

> * Side note: For longer text (especially on paper), I prefer serif fonts.
>   DejaVu is a free serif font that is also optimised for viewing on
>   screen.

The only serif guaranteed to be available in PDF is Times :-(

rst2pdf supports TTF font embedding, and I intend to provide a set of 
stylesheets supporting different default fonts so you can do a 
 ``-s times`` in the next release.

-- 
 ("\''/").__..-''"`-. .         Roberto Alsina
 `9_ 9  )   `-. (    ).`-._.`)  KDE Developer (MFCH)
 (_Y_.)' ._   ) `._`.  " -.-'  http://lateral.netmanagers.com.ar 
  _..`-'_..-_/ /-'_.'     The 6,855th most popular site of Slovenia   
(l)-'' ((i).' ((!.'       according to alexa.com (27/5/2007) 
"Our opponent is an alien starship packed with atomic bombs, I said. 
We have a protractor. Okay, I’ll go home and see if I can scrounge up a 
ruler and a piece of string."  — Neal Stephenson

On 2009-05-26, Roberto Alsina wrote:
> On Tuesday 26 May 2009 04:00:31 Guenter Milde wrote:

> Ok, will change rst2pdf's behaviour to be more like pdflatex, using the
> DPI setting only with the "natural size".

Fine. (Mark that this is exactly what you specified at the start of this
thread.)

>> > because 72DPI means almost every document that works on rst2html breaks
>> > with rst2pdf :-(

>> Use whatever suits you (or your users) best. As it is a configuration
>> value, it can easily be changed in a config file as well.

Typical monitor resolutions are around 100 DPI nowadays. Add the longer
viewing distance and better resolution of a printout and + the desire to
save paper and you get a rangen between 100 and 300 for a "size-keeping
PDF resolution setting".

>> Maybe latex2pdf could provide this config value as well, replacing the
>> "cryptic" LaTeX commands to configure this in a style sheet?

> It could also be added to rst2latex, maybe?

I added it to the todo list
http://docutils.sf.net/docutils/docs/dev/todo.html
but 

* I still look for a better name,
* don't know whether the setting should override a style sheet setting or
  vice versa, so maybe it is simpler to keep it as a "style sheet option" in
  rst2latex.

Günter

On 2009-05-27, Roberto Alsina wrote:
> I am trying to improve how rst2pdf displays lists and I need some opinions:

> Does this look ok?

> http://twitpic.com/619qi

> For the record, here's what rst2latex does with it:

> http://twitpic.com/619z9

> And here's what rst2html does:

> http://twitpic.com/61a13

* It would be easier to evaluate from the PDF file.

* To be better comparable, you might use rest2latex with the option
  ``--stylesheet=parskip`` (separate paragraphs by vertical space).

* I would prefer a bullet size between the rst2pdf and LaTeX examples.

* Consider supporting the "compact" option and class argument (from
  rst2html) as some lists look better compact and others need the space.

* Side note: For longer text (especially on paper), I prefer serif fonts.
  DejaVu is a free serif font that is also optimised for viewing on
  screen.
  
Günter

I am trying to improve how rst2pdf displays lists and I need some opinions:

Does this look ok?

http://twitpic.com/619qi

For the record, here's what rst2latex does with it:

http://twitpic.com/619z9

And here's what rst2html does:

http://twitpic.com/61a13

-- 
 ("\''/").__..-''"`-. .         Roberto Alsina
 `9_ 9  )   `-. (    ).`-._.`)  KDE Developer (MFCH)
 (_Y_.)' ._   ) `._`.  " -.-'  http://lateral.netmanagers.com.ar 
  _..`-'_..-_/ /-'_.'     The 6,855th most popular site of Slovenia   
(l)-'' ((i).' ((!.'       according to alexa.com (27/5/2007) 
"Our opponent is an alien starship packed with atomic bombs, I said. 
We have a protractor. Okay, I’ll go home and see if I can scrounge up a 
ruler and a piece of string."  — Neal Stephenson

Leandro Lucarella, el 26 de mayo a las 12:03 me escribiste:
> Leandro Lucarella, el 20 de mayo a las 11:51 me escribiste:
> > > i don't know of a good solution for that problem...
> > > 1) use fixed width font (easy to implement as option, but doesn't look
> > >    so nice)
> > 
> > I'd like to use this solution (I find much uglier to have big dead spaces
> > than using a fixed width font). I even tried to use a fixed width font
> > (changing the produced SVG by hand) but the result was the same.
> > 
> > A simple hack to lower the calculated width of text would do for me too,
> > but I couldn't find where to change that.
> 
> I could "emulate" this by using the 'aspect' option. Setting it to ~0.5
> make the diagrams looks much better. There is still a little difference
> between long and short strings, but it looks much better. If I change the
> font in the produced SVG to "mono", it looks even better (with an aspect
> of ~0.55). The only problem is the fill patterns, they are "compressed"
> and look ugly.

I just realized that the suggestion is in the README.txt =/

> Is there any easy way to hack aafigure to only "compress" the text? And
> maybe to calculate the text width using constant character width and
> render using a monospace font?

I was thinking a little about the problem and I think, at least for
monospaced fonts, the problem of the "compressed" patterns can be solved
by being more close to the monospace fonts aspect ratio. A monospace font
is about 2 times wider than taller, so if each character is translated to
2 "horizontal cells" in the image "grid", the aspect ratio can be kept to
1 so patterns are not compressed and labels don't get wider than they are.
For example, if you need to draw a filled square, you should write:

XXXX            XX
XXXX instead of XX
 or
OOOO instead of OO
OOOO            OO

What do you think?

-- 
Leandro Lucarella (luca) | Blog colectivo: http://www.mazziblog.com.ar/blog/
----------------------------------------------------------------------------
GPG Key: 5F5A8D05 (F8CD F9A7 BF00 5431 4145  104C 949E BFB6 5F5A 8D05)
----------------------------------------------------------------------------

Leandro Lucarella, el 20 de mayo a las 11:51 me escribiste:
> > i don't know of a good solution for that problem...
> > 1) use fixed width font (easy to implement as option, but doesn't look
> >    so nice)
> 
> I'd like to use this solution (I find much uglier to have big dead spaces
> than using a fixed width font). I even tried to use a fixed width font
> (changing the produced SVG by hand) but the result was the same.
> 
> A simple hack to lower the calculated width of text would do for me too,
> but I couldn't find where to change that.

I could "emulate" this by using the 'aspect' option. Setting it to ~0.5
make the diagrams looks much better. There is still a little difference
between long and short strings, but it looks much better. If I change the
font in the produced SVG to "mono", it looks even better (with an aspect
of ~0.55). The only problem is the fill patterns, they are "compressed"
and look ugly.

Is there any easy way to hack aafigure to only "compress" the text? And
maybe to calculate the text width using constant character width and
render using a monospace font?

Thanks!

-- 
Leandro Lucarella (luca) | Blog colectivo: http://www.mazziblog.com.ar/blog/
----------------------------------------------------------------------------
GPG Key: 5F5A8D05 (F8CD F9A7 BF00 5431 4145  104C 949E BFB6 5F5A 8D05)
----------------------------------------------------------------------------

On Tuesday 26 May 2009 04:00:31 Guenter Milde wrote:
> > 2) rst2pdf uses a default DPI setting (300 instead of 72)
>
> This is OK, in my view.
>
> My stress is on "comparable" in the sense of:
>
> * there is a configurable DPI setting
> * the images "print-resolution" will override this (if no size is
>   specified).

Ok, will change rst2pdf's behaviour to be more like pdflatex, using the DPI 
setting only with the "natural size".

> > I could ignore the image's DPI too, that's not terribly important it
> > seems. The second... well, having it set to 72 has been the main cause
> > of "rst2pdf is broken" mails so far, thus the change :-D
> >
> > But really, I could change it back, specially since there has never
> > been a release that *didn't* use 72DPI as default. Not happily, though
> > because 72DPI means almost every document that works on rst2html breaks
> > with rst2pdf :-(
>
> Use whatever suits you (or your users) best. As it is a configuration
> value, it can easily be changed in a config file as well.
>
> Maybe latex2pdf could provide this config value as well, replacing the
> "cryptic" LaTeX commands to configure this in a style sheet?

It could also be added to rst2latex, maybe?

-- 
 ("\''/").__..-''"`-. .         Roberto Alsina
 `9_ 9  )   `-. (    ).`-._.`)  KDE Developer (MFCH)
 (_Y_.)' ._   ) `._`.  " -.-'  http://lateral.netmanagers.com.ar 
  _..`-'_..-_/ /-'_.'     The 6,855th most popular site of Slovenia   
(l)-'' ((i).' ((!.'       according to alexa.com (27/5/2007) 
"Our opponent is an alien starship packed with atomic bombs, I said. 
We have a protractor. Okay, I’ll go home and see if I can scrounge up a 
ruler and a piece of string."  — Neal Stephenson

On 2009-05-25, Roberto Alsina wrote:
> On Monday 25 May 2009 18:08:05 Guenter Milde wrote:

>> In my view, it is desirable to have a comparable treatment of px in
>> rst2latex and rst2pdf. Given that pdflatex is an established tool ...

> The only differences right  now seem to be:

> 1) pdflatex ignores the image's DPI setting

Actually, pdflatex uses the images "print-resolution" setting as well --
if the "natural size" of the image is used.

If the directive specifies a width or height argument in px, the global
resolution setting is used. 

To find out, I used Gimp to edit biohazard.png (specify a printing-size
of 1 inch and save to biohazard-16dpi.png) and added to the
image-unit-test the lines::

  .. image:: images/biohazard-16dpi.png
  
  16 x 16 image with print-resolution set to 16 DPI.
  
  All the above images should be exactly 1 inch wide.
  
  .. image:: images/biohazard-16dpi.png
     :width: 16px
  
  Image with print-resolution set to 16 DPI and width-option 16 px.
  
  The above image would be exactly 1 inch wide, if the "image-specified"
  printing resolution is used also for external width setting.
  With pdflatex, the global resolution is used, so it is tiny.


> 2) rst2pdf uses a default DPI setting (300 instead of 72)

This is OK, in my view. 

My stress is on "comparable" in the sense of: 

* there is a configurable DPI setting
* the images "print-resolution" will override this (if no size is
  specified).

> I could ignore the image's DPI too, that's not terribly important it
> seems. The second... well, having it set to 72 has been the main cause
> of "rst2pdf is broken" mails so far, thus the change :-D

> But really, I could change it back, specially since there has never
> been a release that *didn't* use 72DPI as default. Not happily, though
> because 72DPI means almost every document that works on rst2html breaks
> with rst2pdf :-(

Use whatever suits you (or your users) best. As it is a configuration
value, it can easily be changed in a config file as well.

Maybe latex2pdf could provide this config value as well, replacing the
"cryptic" LaTeX commands to configure this in a style sheet?

Günter

On Monday 25 May 2009 18:08:05 Guenter Milde wrote:
> In my view, it is desirable to have a comparable treatment of px in
> rst2latex and rst2pdf. Given that pdflatex is an established tool ...

The only differences right  now seem to be:

1) pdflatex ignores the image's DPI setting
2) rst2pdf uses a default DPI setting (300 instead of 72)

I could ignore the image's DPI too, that's not terribly important it seems.
The second... well, having it set to 72 has been the main cause of "rst2pdf is 
broken" mails so far, thus the change :-D

But really, I could change it back, specially since there has never been a 
release that *didn't* use 72DPI as default. Not happily, though because 72DPI 
means almost every document that works on rst2html breaks with rst2pdf :-(

-- 
 ("\''/").__..-''"`-. .         Roberto Alsina
 `9_ 9  )   `-. (    ).`-._.`)  KDE Developer (MFCH)
 (_Y_.)' ._   ) `._`.  " -.-'  http://lateral.netmanagers.com.ar 
  _..`-'_..-_/ /-'_.'     The 6,855th most popular site of Slovenia   
(l)-'' ((i).' ((!.'       according to alexa.com (27/5/2007) 
"Our opponent is an alien starship packed with atomic bombs, I said. 
We have a protractor. Okay, I’ll go home and see if I can scrounge up a 
ruler and a piece of string."  — Neal Stephenson

On 2009-05-25, Alan G Isaac wrote:
> On 5/25/2009 2:58 AM Guenter Milde apparently wrote:

>> 1. All writers support px. 

> As you noticed, I unsed the term "unambiguously defined".

Which is too vague to base a decision on whether to use or dicard a
lenght given in pixels given the fact that px is a *relative* unit (like
em and ex).

The fact that the resolution of the viewing device (or printer) is
generally not known at the time of the translation is valid for all
writers.

> But I agree this introduced some ambiguity and is
> quite an unfortunate part of CSS.  Currently identical
> widths specified as an element attribute or alternatively
> in a style-sheet have different meanings!  And it is
> worse than this, as the CSS recommendation is to scale
> differently for each media type.  

In my view this is just the nature of a relative unit: it will "scale"
depending on something else. If you don't like it, don't use it.

> Sheesh: too clevel by half.

??

...

> My only problem is that the specification is "abnormal".
> (I.e., it is a writer option, not a directive option.)
> When coupled with the proposal that writers have a default
> conversion, it becomes in my opinion too implicit.

I suggest to learn from the people that already spent time and energy
in solving the problem of "how to interpret the lengh unit px" for
different media types:

CSS recommends 
  + not to use the actual resolution if it differs widely from the
    "typical" screen resolution (prevent the surprise of HUGE or tiny
    images with some output media),
    
  + try to keep the "perceived size" roughly constant.
    
  + to scale so that an integer number of "output pixels" corresponds to 
    one "px". This will avoid alias (quantisation artifacts).

    (However, it requires knowledge of the actual "output resolution"
    which is not given for PDF generation.)
    
pdflatex 
  + uses a pre-set DPI value instead of the unknown "output resolution",
  + allows configureation of this value by the user.

In my view, it is desirable to have a comparable treatment of px in
rst2latex and rst2pdf. Given that pdflatex is an established tool ...

Günter

On 2009-05-25, Roberto Alsina wrote:
> On Sunday 24 May 2009 17:57:30 Guenter Milde wrote:
>> On 2009-05-21, Roberto Alsina wrote:

>> For some lengths, is is (or was until recently) even the only valid unit
>> (see the description of the actual lenght argument). There is also at
>> least one case of a "percent" length without unit (i.e. 45 -> 45 %).

> Is there a chance of that being required to have a unit?
> Or else, can you tell me where it is so I can catch it ;-)

Sorry, I mixed this up, the an interger value for the percentage is
required for the "scale" argument of a figure or image. In the SVN
version, it takes now an (optional) percent sign.

Günter

On Monday 25 May 2009 11:56:42 Alan G Isaac wrote:
> On 5/25/2009 2:58 AM Guenter Milde apparently wrote:
> > 1. All writers support px.
>
> As you noticed, I unsed the term "unambiguously defined".
>
> > 2. px is ambiguous in *every* writer:
> >    HTML uses the screen resolution but CSS defines 1 px
> >    = 1/96 in,
> >    so what is the size of a pixel for a html4css1 writer?
>
> I do not think this is quite correct.
> http://www.w3.org/TR/CSS21/syndata.html#length-units
> There is nothing more but a recommendation to rescale.
> (I.e., not a redefinition.)
> But I agree this introduced some ambiguity and is
> quite an unfortunate part of CSS.  Currently identical
> widths specified as an element attribute or alternatively
> in a style-sheet have different meanings!  And it is
> worse than this, as the CSS recommendation is to scale
> differently for each media type.  Sheesh: too clevel by half.
>
> > d) if you want an image that is exacly 48 pixels on the
> > screen when viewed with a HTML browser and at the same
> > time some exact fixed value
> >    in page based media (printout or PDF), explicitely specify both, 48px
> >    (but not just 48) and a conversion value from px to a static length
> > unit.
>
> I actually largely agree with this: you are treating dpi
> metadata as intended only for page-based media.  I believe
> this is correct.  However, I believe it conflicts with
> Robert's understanding or preferences.

Hmmm... I am trating PDF as print media and using the image's metadata if 
available. I am not sure if I am right or wrong on *how* I am doing it, 
though.


-- 
 ("\''/").__..-''"`-. .         Roberto Alsina
 `9_ 9  )   `-. (    ).`-._.`)  KDE Developer (MFCH)
 (_Y_.)' ._   ) `._`.  " -.-'  http://lateral.netmanagers.com.ar 
  _..`-'_..-_/ /-'_.'     The 6,855th most popular site of Slovenia   
(l)-'' ((i).' ((!.'       according to alexa.com (27/5/2007) 
"Our opponent is an alien starship packed with atomic bombs, I said. 
We have a protractor. Okay, I’ll go home and see if I can scrounge up a 
ruler and a piece of string."  — Neal Stephenson