Hi Jan,

Thanks for looking into this. Of the 3 options you propose, option (2) 
seems the most preferable.

Thanks,

Chris

On 22/02/2019 22:30, Jan Tosovsky wrote:
> On 2019-02-20 Chris wrote:
>> On 09/02/2019 18:25, Jan Tosovsky wrote:
>>> there is a proposal to enhanced painting of table borders to avoid visual
>>> inconsistencies in some renderers (Acrobat Reader, Chrome):
>>> https://issues.apache.org/jira/browse/FOP-2536
>>>
>>> That technique renders another set of shapes over the existing ones. Not
>>> everyone wants this as a default behaviour. After applying the patch it can
>>> be activated via RendererOption 'overpaint-table-borders'. It means it can
>>> be set programmatically, but not via configuration file or cmdline options.
>>>
>>> I personally do not need this as a cmdline switch, but having this option in
>>> the configuration file is IMHO a must for general audience.
>>>
>>> So the questions:
>>> (1) option name (is boolean 'overpaint-table-borders' Ok or something else
>>> should be prefered, e.g. 'table-border-rendering-mode' with the set of
>>> String values, initially 'standard' and 'overpaint')
>>>
>>> (2) If cmdline switch would be useful, what the name should be used
>>>
>> 'table-border-rendering-mode' with the set of
>> String values, initially 'standard' and 'overpaint'
> Ok, thanks.
>
> I've implemented this option for PDF renderer only. After parsing the 
> configuration file this option is passed via RendererOptions map into the 
> TableLayoutManager where it controls if borders are overpainted or not.
>
> My intention was to write a layout test case. However, when layout test is 
> executed, the PDF renderer configuration is ignored as the MIME type is now 
> application/X-fop-intermediate-format, where this option is missing.
>
> So there are some options now:
> (1) making this option global (like target-resolution)
> (2) duplicating this option also for Intermediate renderer (AFAIK there is no 
> other option yet)
> (3) accepting that the overpainting is not covered by any layout test 
> (enabling is on user own risk anyway)
>
> Thanks for your inputs,
>
> Jan
>
>
>
> .
>

Reply via email to