[poppler] Some patches WAS Re: Poppler 0.19.1 (0.20 Beta 1) released

[Thomas Freitag]

Am 26.03.2012 20:39, schrieb Albert Astals Cid:
> El Dilluns, 26 de març de 2012, a les 20:30:35, Thomas Freitag va escriure:
>> Am 26.03.2012 19:47, schrieb Albert Astals Cid:
>>> El Diumenge, 25 de març de 2012, a les 22:34:20, Thomas Freitag va
> escriure:
>>>> Am 19.03.2012 08:05, schrieb Thomas Freitag:
>>>>> Am 16.03.2012 16:08, schrieb Carlos Garcia Campos:
>>>>>> Excerpts from Albert Astals Cid's message of jue mar 15 22:23:15 +0100
>>> 2012:
>>>>>>> Available from
>>>>>>> http://poppler.freedesktop.org/poppler-0.19.1.tar.gz
>>>>>>>
>>>>>>> WARNING: This is a unstable release, it is actually 0.20 Beta 1
>>>>>>> release,
>>>>>>> it
>>>>>>> should work like any release from the 0.18 branch, but do not blame us
>>>>>>> if it turns your cat into a lion and eats your breakfast.
>>>>>>>
>>>>>>> Changes against the 0.19.0 branch:
>>>>>>>
>>>>>>> core:
>>>>>>>     * Improve CJK suport in PSOutputDev
>>>>>>>     * CJK substitute implementation on WIndows platforms
>>>>>>>     * Do not crash on malformed files with 0 bits in the color map of
>>>>>>>     an
>>>>>>>     image
>>>>>>>     * Fix regression in some PSOutputDev array sizing
>>>>>>>     * Improvements to Annotation editing
>>>>>>>     * Fix logic error in Rendition parsing code (Bug #47063)
>>>>>>>     * Minor API changes to SplashOutputDev (Bug #46622)
>>>>>>>     * Fix mismatch in some functions declarations
>>>>>>>     * Update poppler copyright year
>>>>>>>
>>>>>>> utils:
>>>>>>>     * pdftops: Fix -passfonts regression. (Bug #46744)
>>>>>>>     * pdffonts: List the encoding of each font. (Bug #46888)
>>>>>>>     * pdftohtml: Add possibilty of controlling word breaks percentage.
>>>>>>>     (Bug #47022)>>>
>>>>>>>
>>>>>>> qt4:
>>>>>>>     * Support for LinkMovie object (Bug #40561)
>>>>>>>     * Support for Media Rendition
>>>>>>>
>>>>>>> glib:
>>>>>>>     * Add poppler_fonts_iter_get_encoding
>>>>>>>     * Improvements to the demo
>>>>>>>
>>>>>>> Testing, patches and bug reports welcome.
>>>>>>>
>>>>>>> My TODO list is:
>>>>>>>     * Work with Fabio in his Annot Qt4 patches
>>>>>>>     * Decide something with Adrian in the "Pdf that uses the symbol
>>>>>>>     font"
>>>>>>>     thread * Review 47186 pdftohtml: mask images are not extracted,
>>>>>>>     unless they are JPEG * Review 32340 pdftohtml: renders images
>>>>>>>     vertically flipped
>>>>>>>
>>>>>>> Remember feature freeze is in 2 weeks with Beta 2
>>>>>>>
>>>>>>> If anyone has anything else that thinks it is waiting on me for
>>>>>>> further
>>>>>>> feedback, tellme, because it's not on my list
>>>>>> The xpdf303 merge introduced some regressions in the cairo backend,
>>>>>> I'll try to find some time this weekend to send more details to the
>>>>>> list.
>>>> I just finished the patch for these regressions, they had differents
>>>> reasons 1. In CharCodeToUnicode::mapToUnicode the identity support was
>>>> missing 2. The new algorithms for axial and radial shading caused
>>>> problems in cairo. I revert these changes but removed the examination of
>>>> hidden content (this is already done in the calling function)
>>>> 3. The examination of optional hidden content in showing text was wrong:
>>>> of course hidden text should not be shown, but text parameters like text
>>>> position in the state must be changed!
>>>> 4. Searching and finding fonts especially with base14 fonts should be
>>>> more exact than just looking at the base14 name (i.e. fixed width and so
>>>> on) when using fontconfig. I implement that to find the best font
>>>> fitting to the needs.
>>> Don't know which of them but applying it makes me lose half of the letter
>>> in bug-poppler28632.pdf
>> Can You explain it a little bit more? I saw especially this one as on
>> enhancement. I send You in private (it's nearal 1 MByte) the output with
>> cairo and splash with this patch, the ouptput with splash with the head
>> revisiion and the ouput with r92ce79.
> You can reproduce the problem. Have a look at your files and see that bug-
> poppler28632-spash-patch-1.png has much fewer characters than bug-
> poppler28632-spash-head.1.png (e.g. first column, last row, the 'a' is
> missing)
But at least You should agree, that in this column the output is the 
same as with r92ce79 :-) Damnish, I looked just at the head row. I take 
a look, what fontconfig decides with this PDF ...

Thomas
>
> Albert
>
>> Thomas
>>
>>> Can you please have a look?
>>>
>>> Albert
>>>
>>>> Cheers,
>>>> Thomas
>>>>
>>>>>>> Albert
>>>>>> Thanks,
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> poppler mailing list
>>>>>> poppler at lists.freedesktop.org
>>>>>> http://lists.freedesktop.org/mailman/listinfo/poppler
>>>>> _______________________________________________
>>>>> poppler mailing list
>>>>> poppler at lists.freedesktop.org
>>>>> http://lists.freedesktop.org/mailman/listinfo/poppler
>>> _______________________________________________
>>> poppler mailing list
>>> poppler at lists.freedesktop.org
>>> http://lists.freedesktop.org/mailman/listinfo/poppler
>>>
>>> .
>> _______________________________________________
>> poppler mailing list
>> poppler at lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/poppler
> _______________________________________________
> poppler mailing list
> poppler at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/poppler
>
> .
>