https://bugs.documentfoundation.org/show_bug.cgi?id=91252

--- Comment #9 from Armin Le Grand (CIB) <[email protected]> ---
Checked example from comment 8, it contains a polygon usin ga fill pattern. The
fill pattern is defined using geometry (not pixels), thus re-inserting to the
office and breaking up will reveal the clipped fill pattern in geometry
quality. What the browsers seem to do with fill patterns is to prepare them as
bitmap tile, then render a tiled fill of it, probably for speed reasons. That
single tile bitmap shows AA-ed edges of the contained ellipses. To make that
more obvious you may change the gradient steps to 12 or so, export to svg again
and check in the browsers.
Also a good test is to use inkscape, it also shows no moire. I do not say that
often, but in this case the SVG export is not the reason, but more the
browser's renderers.
To really make sure I checked inside the SVG file. It contains SVG fill patters
which use path objects with simple fill color and no stroke, thus the moire
really comes from some AAing of the browsers, there is *no* stroke defined for
the pattern fill.
Still leads to that the SVG export creates no SVG gradients - as do ther
exports. That is since SVG gradients do not 'reach' the exporter. We would
need:
(1) SVG gradients in primitives (already tasks for that I think)
(2) SVG (and other) exports based on primitives, *not* on metafiles like now
(3) the exporter use the higher geometry definitions

Unfortunately a lot of work, bu tat least with the primitives the migration
path is available and the current 'fallback' works as well as possible.

-- 
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
Libreoffice-bugs mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs

Reply via email to