[poppler] user-defined output device & private functions in libpoppler

[suzuki toshiya]

Dear Adam,

Thank you for encouraging comment!

Regards,
mpsuzuki

Adam Reichold wrote:
> Hello,
> 
> I am obviously not a maintainer, but here a my two cents anyway: Just
> from reading this and hence assuming that OPVP is to become generally
> relevant, I think upstreaming the OPVPOutputDevice and possibly even
> making pdftoopvp part of poppler-utils sounds like the most reasonable
> course of action and would be more or less parallel to the already
> existing PostScript output device and utility program.
> 
> Best regards, Adam.
> 
> Am 02.12.2016 um 01:37 schrieb suzuki toshiya:
>> Dear Poppler maintainers,
>>
>> I would like to hear your comments about the possibility
>> of the disclosure of functions used "in" libpoppler, like,
>> Goo, Fofi and Splash (if there was some discussion in the
>> past, please give me the pointer to there).
>>
>> It could be useful for the developers who want to make
>> their own new OutputDevice, out of libpoppler. Or, it is
>> better for the developers to propose new OutputDevice to
>> be included by future release of libpoppler, rather than
>> making external library referring private functions?
>>
>> --
>>
>> Recently, CUPS developers discussed about the future of the
>> filter "pdftoopvp". OPVP is an API set for the non-PostScript
>> printers discussed by the engineers in OpenPrinting consortium
>> under Linux Foundation (OPVP is not yet another PDL).
>>
>> Also Ghostscript supports OPVP output, some engineers
>> regarded Ghostscript is overkilling for pdf-to-opvp data flow,
>> and using Poppler would be more compact & straightforward.
>>
>> However, libpoppler has no officially-exposed APIs to make
>> user-defined OutputDevice. Thus, pdftoopvp filter links to
>> private functions in libpoppler, to make OPVPOutputDevice
>> (basically it is a diversion of SplashOutputDevice).
>>
>> Some people have concerns that linking private functions
>> might be risky, because nobody guarantees the compatibility.
>>
>> Thus, I want to ask the questions in the beginning. Is
>> there any possibility for libpoppler to expose some internal
>> functions for user-defined OutputDevice developers, or,
>> they should propose their OutputDevice to libpoppler?
>>
>> I hope to hear your opinions.
>>
>> Regards,
>> mpsuzuki
>>
>> _______________________________________________
>> poppler mailing list
>> poppler at lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/poppler
>>
>