Em Mon, 28 Nov 2016 10:43:15 +0100 Markus Heiser <markus.heiser@xxxxxxxxxxx> escreveu: > Am 28.11.2016 um 10:16 schrieb Jani Nikula <jani.nikula@xxxxxxxxx>: > > >> Since this is for "figure" AND "image" (inline) I thought it is > >> better to replace. But I'am open to change this .. concrete ideas? > > > > Please don't shadow Sphinx directives. Otherwise people will look for > > Sphinx documentation for the directive, and will be confused that it > > doesn't do what the documentation says. And we'll be able to add our own > > directive options to our own extensions without adding to the confusion. > > OK, I will use kernel-figure mentioned by Daniel. For inline images > I will use kernel-image. While I would prefer to keep using the original directives (provided of course that everything there would be supported), I'm OK with using a different tag. Yet, if we're willing to use kernel-figure/kernel-image, the best would be to generate a warning if some file uses figure/image, as we don't want ReST files using the old ones. > > [snip] > > >>> - We don't want to do the conversion unconditionally, right? SVG works > >>> just fine for HTML output, we just need to do it for latex, as I > >>> understand it. > >> > >> OK, by this explanation, you see, we have to convert in the writer > >> phase, but this will bring some other drawbacks. E.g. we have > >> to touch all writers (html, latex2e etc.), this means we have to write > >> our own builders for for each output format. If we wan't to be academic > >> this might be the best solution, but IMO maintaining own builder > >> is not what we want, so I choose the more *pragmatic* solution, > >> creating all potential formats in he reading phase ... any ideas ? > > > > Are we just doing an unnecessary conversion when we're only building > > HTML, or will HTML also use the converted files while it could use the > > source format directly? >From what I understood, it is not optimized neither for PDF or HTML outputs. For HTML, it will produce the uneeded *.pdf images from svg; for PDF/LaTeX, it will produce an intermediate *.svg files from Graphviz, instead of using "dot -Tpdf". It would be good (at least long term), if we could get rid of those unneeded conversions. As we add more images, this could be making the documentation build a lot slower than it needs to be. > It's only about unnecessary conversion. BTW I just remember: if graphviz > is not available, the DOT language is inserted as literal-include. This > can't be implemented in the write phase, so I think we have to accept, > that a (unused) PDF is generated from a SVG even if only HTML is build. Thanks, Mauro -- To unsubscribe from this list: send the line "unsubscribe linux-doc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
