Fixed point representation couldn't store 0.65 exactly, so avoid doing
the divide by 100 first.
I will look at moving this into libcss's fixed point header and doing
it in a way that avoids arithmetic overflow, but for now this fixes
el reg layout.
This changes the LOG macro to be varadic removing the need for all
callsites to have double bracketing and allows for future improvement
on how we use the logging macros.
The callsites were changed with coccinelle and the changes checked by
hand. Compile tested for several frontends but not all.
A formatting annotation has also been added which allows the compiler
to check the parameters and types passed to the logging.
By hiding all but the form selection menu option structure from code
outside of render this reduces the API to the absolute minimum to
support this feature.
The netsurf core is driven from numerous operation tables most of
which are now set through a common netsurf_register() interface. The
font and plotting interfaces are currently separate and unlike all the
other operation tables are modified for differing contexts.
This change moves the font operations alongside all the other
operations table and remove unnecessary interaction with the renderers
font internals. Further this also removes the need for css internals
to be visible in frontends.
Values returned by nscss_len2px are actually libcss fixed-points,
so convert the result to pixels. Also reduce the amount by which
the top of descendant bboxes are extended to 3/4 of the text
height, which is the maximum possible amount that the text can
protrude above the box when line-height is reduced.
This allows the HTML layout code to wrap unicode text correctly.
For example on our Japanese Welcome page, the characters will be wrapped
if the platform's nsfont_split implementation implements the Unicode line
breaking algorithm.
A number of things wrong here:
1) The BOX_TABLE generated for the img shouldn't ever have been associated with the object (it should have been associated with the implied BOX_TABLE_CELL)
2) The layout engine's decision to treat BOX_TABLE/BOX_INLINE_CONTAINER with associated objects differently from non-replaced TABLE/INLINE_CONTAINERs seems wrong (looking though SVN history, it appears that this was introduced as part of the very first frames implementation, since replaced (twice!))
3) Now we fetch objects in parallel with building the box tree, we're open to this kind of race condition.
svn path=/trunk/netsurf/; revision=13128
