Borderline display with direct paint and with primitive direct
paint has quite some errors in the current state. Started to
unify usages, check deeper with creation/usage.
borderline: deep changes to BorderLine
Found basic error in determining the offset values
for BorderLinePrimitive creation, these were not
centered on the lines. Corrected that. This makes
it possible to remove the formally used clipping
which seems to have been used to correct that. Also
allows to go back to a 'normal' decomposition that
creates line primitives as expected. That again can
then be painted quite normally.
Also added view-dependent case to the decomposition
to guarantee a gap of one discrete unit (pixel).
Removed the direct painter, too. Checked and corrected
stroking.
borderline: Adapted previews to primitives
Added code to use the primitive representation in
all dialogs and apps using tables. The edit views
use these mostly, so the preview should do that,
too. Currently missing is a good visualization of
diagonals, but this is also true for edit views.
Checked all apps and table usages to not get worse
borderline: correct line dash visualization
When a dashed line is used, a factor of 10.0 was applied in the
original coded, added that. Also the orientation of vertical
borders was inverted since it was simpler to exchange Start/End,
but this also mirrors the line dash visualisation, corrected that
Change-Id: I4c1b380a76cb37389fab1259a53fb7cc9da982d1
e95e246d5563360617a2a2213e4d5ec7d0e736b9
62369b4de58fb0264aeb710ec6983ceddca5701d
77418cc6c84ebb0632f8c3448976e82ce612d6b6
b4eb28dc86ce05eb89b26517167305b994158ef8
borderline: adapt cppunittest and clang
This is what it is actually is, we dropped support for server-side fonts
for a long time now. Renamed also a few related classes, but left
ServerFontLayout* ones as they will go away soonish.
Change-Id: I68a6dad51b6972368b7bf85a0b9c8089cc12740e
Reviewed-on: https://gerrit.libreoffice.org/30390
Reviewed-by: Khaled Hosny <khaledhosny@eglug.org>
Tested-by: Khaled Hosny <khaledhosny@eglug.org>
By accident, the plugin had only looked into non-member functions until now.
However, for the vast majority of newly found cases, it is unclear to me whether
the relevant code needs fixing or not, so I put all those into a giant whitelist
for now. Maybe the plugin's sginal/noise ration just isn't worth it, anyway.
Change-Id: I7997a0ebdb4211af208285611656a40ead97d935
...so that isZeroConstant doesn't trigger an assert inside Clang's
isCXX11ConstantExpr when expr is sizeof(x) with x being dependent on a template
argument.
Change-Id: I6bab46e64cc085d597db25994d8bfdc66417fe83
