27 Mar
2018
27 Mar
'18
2:07 p.m.
Hello all,
Hello Mohamed,
There has been two tools suggested in this thread: flyingsaucer and weasyprint.
Can you find more tools?
Can you try to see how an architecture would look like with these two tools and
incorporate that in your proposal?
flyingsaucer should be quite easy to integrate since it’s in java. So it’d be probably a
pom.xml change or an extension with the relevant pom.xml…
paul
On 26 Mar 2018, at 9:55, Vincent Massol wrote:
Hi Ludo,
On 26 Mar 2018, at 09:16, Ludovic Dubost
<ludovic(a)xwiki.com> wrote:
On Sat, Mar 24, 2018 at 10:06 PM, Vincent Massol <vincent(a)massol.net> wrote:
This is a major limitation of a latex based export for XWiki. This makes it
very hard to export any macros that would produce HTML + CSS and any HTML
that the user would create in XWiki.
Yes this is what I was mentioning re CSS. On the HTML side, it’s not exactly true since
we can parse the HTML with the XWiki HTML parser which generates and XDOM and then render
that XDOM. It won’t be perfect though. For example there are HTML elements that are not
supported by our HTML parser (ex: <FORM> elements).
On 24 Mar 2018, at 21:58, Paul Libbrecht
<paul(a)hoplahup.net> wrote:
Hello Mohammed,
have you googled for paged-media html to css converters?
Surely an option is to let it be done by the browser but there must also
be
engines.
We have evaluated this in the past and there are lots of limitations, see
https://markmail.org/message/ztcwibiuoqfjcnjo
E.g. I think that phantomJS of weasyprint can do
that. However, I
haven’t found yet in java (which would simplify things).
Note that phantomjs is dead now:
https://www.puzzle.ch/blog/articles/2018/02/12/phantomjs-
is-dead-long-live-headless-browsers
As Vincent says, print with LaTeX in the middle
is a way to get
high-quality but there are many losses too: it is really hard to
get CSS
rules to be all implemented in TeX.
Yes indeed, that’s very hard. CSS shouldn’t be used as a way to style the
LaTeX output. The LaTeX exporter itself should provide its own way of
controlling the style of the output. This is what I do in the LaTeX
exporter. Basically I provide some default styles (sometimes with some
config options) and the user has the ability to control exactly the styles
he/she wants applied if the default style is not enough. It’s not trivial
though and will take a bit of time if you need a heavily styled document.
The current XML-FO based export supports a
limited set of HTML + CSS. Also
latex does not provide us with a java pdf export.
One thing has to be clear: I’m absolutely not pushing for having the LaTeX export as a
replacement of our XSL-FO approach. For some reason you seem to be hinting at that which
is not my opinion for several reasons. I mentioned LaTeX here because it’s important to
know all the technologies that exist to produce a PDF and that’s one we have, that’s all.
It’ll be interesting at some point to draw a Pro/Cons table on xwiki.org to compare the
various export options with their limits.
The CSS paged media standard has this advantage
of bringing to the table
HTML + CSS support and support of CSS for the general document output
(header, footer, etc..). Now of course we need to find the right libraries
for that. It would be nice to have an experiment based on this to see how
far we can go with css pages media.
Definitely. That’s actually the purpose of this GSOC Ludo! :)
TBH I was the one pushing for this experiment initially when I found about the nice
result of flyingsaucer… So I’m as eager as you to see what we can get with paged CSS.
It's important to consider the full needs if
we want to compare
technologies. The latex export makes nice high quality output but currently
only for the basic syntax elements that we validate for that output.
Regarding the quality of the output, yes it’ll be fun to compare what we get with various
inputs when using the 3 technologies:
* XSL-FO
* LaTeX
* Paged CSS (see https://print-css.rocks/intro.html#what-is-css-paged-media)
Thanks
-Vincent
>
> Ludovic
>
>
>
>>
>> Thanks
>> -Vincent
>>
>>>
>>> I’m wondering if CSSbox could do the job.
>>>
>>> paul
>>>
>>> On 24 Mar 2018, at 20:51, Mohamed Ashraf wrote:
>>>
>>>> Yes this is part of GSOC project
>>>>
>>>> Sent from my iPhone
>>>>
>>>>> On Mar 24, 2018, at 9:29 PM, Vincent Massol
<vincent(a)massol.net>
>> wrote:
>>>>>
>>>>> Hi Mohamed,
>>>>>
>>>>>> On 24 Mar 2018, at 19:12, Mohamed Ashraf
<morybtf(a)gmail.com> wrote:
>>>>>>
>>>>>> Currently, the PDF export of XWiki is implemented based on XSL-FO
and
>>>>>> transformation of XHTML to FO. This poses a couple of problems,
mainly
>>>>>> related to the current level of support of FO from libraries
>> implementing
>>>>>> FO to PDF transformation, as well as the limitations of
automatized
>>>>>> transformation of XHTML to FO. The problems are mainly related
to
>> styling
>>>>>> limitations, auto-layouting, etc.
>>>>>>
>>>>>> The idea is to try to replace this with a pure XHTML & CSS
(paged CSS)
>>>>>> export, using an open source library for producing PDFs out of
this
>>>>>> ,
>>>>>
>>>>> Sure, but which one?
>>>>>
>>>>> The only alternative I know is flying saucer (which is dead:
>> https://github.com/flyingsaucerproject/flyingsaucer). Is that what you
>> mean?
>>>>>
>>>>> Do you know a maintained fork of it? One that I know is used by a
>> competing wiki: https://bitbucket.org/atlassian/xhtmlrenderer-atlassian
>>>>>
>>>>> Are you doing this as part of this GSOC project:
>> http://dev.xwiki.org/xwiki/bin/view/GoogleSummerOfCode/
>> ImplementPDFexportwithXHTMLpagedCSS ?
>>>>>
>>>>> Thanks
>>>>> -Vincent
>>>>>
>>>>>> and I will see LaTeX ,
>>>>>> thanks
>>>>>>
>>>>>> 2018-03-24 19:52 GMT+02:00 Vincent Massol
<vincent(a)massol.net>et>:
>>>>>>
>>>>>>> Hi Mohamed,
>>>>>>>
>>>>>>>> On 24 Mar 2018, at 18:44, Mohamed Ashraf
<morybtf(a)gmail.com> wrote:
>>>>>>>>
>>>>>>>> If we should replacing the XSL-FO which we use to export
PDF file
>> out of
>>>>>>>> XML,
>>>>>>>> with XML and CSS only with open-source library ,
>>>>>>>>
>>>>>>>> and I think * ”CSS Paged Media “ *
>>>>>>>>
>>>>>>>> is this good enough to do that ,
>>>>>>>> or there are any suggestion
>>>>>>>
>>>>>>> Sorry but I don’t understand your question. Why would you
want toi
>> replace
>>>>>>> XSL-FO in your XWiki install?
>>>>>>>
>>>>>>> If you’d like to contribute to XWiki dev, then could you
provide more
>>>>>>> context and explain why you want to replace XSL-FO and by
what.
>>>>>>>
>>>>>>> You may also be interested by the LaTeX exporter which can be
used to
>>>>>>> generate PDFs:
http://extensions.xwiki.org/xwiki/bin/view/Extension/
>> LaTeX/
>>>>>>>
>>>>>>> Thanks
>>>>>>> -Vincent