Age | Commit message (Collapse) | Author |
|
Create a table with a merged cell like in the screenshot in the
bug, with a SAL_DEBUG in SwTabFrmPainter::PaintLines the following
lines are painted:
debug: paint start
1 debug: start: 2749,1488 end: 12387,1488
2 debug: start: 2749,1945 end: 7567,1945
3 debug: start: 7567,1945 end: 12387,1945
4 debug: start: 2749,2015 end: 12387,2015
5 debug: start: 2749,2542 end: 7567,2542
6 debug: start: 7567,2542 end: 12387,2542
7 debug: start: 2749,1488 end: 2749,1945
8 debug: start: 2749,1945 end: 2749,2015
9 debug: start: 2749,2015 end: 2749,2542
A debug: start: 7567,1945 end: 7567,2542
B debug: start: 12387,1488 end: 12387,1945
C debug: start: 12387,1945 end: 12387,2015
D debug: start: 12387,2015 end: 12387,2542
debug: paint end
*11111*11111*
7 B
7 B
*22222*33333*
8 A C
*44444*44444*
9 A D
9 A D
*55555*66666*
The problem is obviously that the Y coordinates of the lines 2, 3
and 4 differ; they should be on the same Y position.
The problem here is that logically horizontal lines must be painted
not centered but "below" the line, and It turns out that
SwTabFrmPainter::Insert cannot correct the positions properly to
do that, because it only looks at borders in a single cell.
When using the UI to set the borders, we get (for innner table borders)
only a bottom border in the cells, but no top borders, so the
top position of the logically vertical borders needs to be corrected
with the width of the bottom border of the cell _above_; a symmetric
correction of the bottom position to the top border of the cell below
is also necessary.
Fortunately if we just leave the positons alone in Insert then
TabFrmPainter will eliminate duplicate lines with equal positions
and so it's only necessary to correct the positions when actually
painting the line in wTabFrmPainter::PaintLines,
where we have the neighboring lines available.
Change-Id: Ia8519f6673db0f3a1ecaa68038896cac39609129
|
|
So use the avoid-ATSUI patch dependingh on (the newly added)
MAXOSX_SDK_VERSION, not MAC_OS_X_VERSION_MAX_ALLOWED.
It is not clear (to me at least) whether for instance building against
SDK version 10.7 but with MAC_OS_X_VERSION_MAX_ALLOWED=1060 ensures
the resulting binaries actually work on 10.6. Somebody should
check. If it doesn't work, we should add a check to configure.in that
the min-required and max-allowed values are not less than the SDK
version.
(If it would work even for MAC_OS_X_VERSION_MAX_ALLOWED=1040, we
would not need to require the increasingly obsolete Xcode 3 and 10.4
SDK for "official" builds.)
Change-Id: Ia7088a26024ed288db55fa734465138d44d2ae12
|
|
Change-Id: I3c1137bdb1186df773a92c34267b2324ad43ae7e
|
|
Change-Id: Ie355701716c7106ef2afd76ae8b6544a00ac474c
|
|
Change-Id: I79d71c682c07f49b70c6791953c000868bf65cd9
|
|
Change-Id: Ie8174ae4b7d2c02503f40fe1263076d924f2c9e2
|
|
When compiled with Clang on OS X, the OS X dynamic linker appears to be
confused about which FieldEntry destructor to call at runtime. In some
cases, the FieldEntry destructor in register/source/reflwrit.cxx is
called instead (both destructors have the same symbol name). This patch
avoids this problem.
Change-Id: I7d69894318cb8fda0a7c5a9b1c2ff3ca0d47a37c
|
|
some big hunks melting away now
Change-Id: Ic74870cad06a2322e98bf1a66287a9b29a7b003e
|
|
since 6d6198393e0677710191248d6f7c9ec15d0f0e0f
Change-Id: I8c08088453027ab7aa0ca964262456475f767a5f
|
|
Change-Id: I38c92dad50fd13dcd32a07feed860eb60349c080
|
|
...instead of sal_Bool, to work around MSVC warning C4805: unsafe mix of type
'sal_Bool' and type 'bool'.
Change-Id: Ia3477539ccc23aa9daec49d633f023d414f2b4bf
|
|
...where a local variable in a nested class shadows a member of an outer class.
Fixed by renaming the relevant outer class members as "m..."
Change-Id: I6caa33914ff1060e08f66eadde879fb7b4e3e7cc
|
|
Change-Id: Idf923e4a1e81a2e2e9c7107e4b6725bd9eb33e38
|
|
Change-Id: Ib7d7860cd422b28b6987f94da0a9c664ff7dfde1
|
|
Change-Id: I51caaca911f71788940260bd63d3854526d7473f
|
|
Change-Id: Id384aeda51bcbba9aefdb242db386a30ae991c30
|
|
Change-Id: I127776bc0a1583107591c9a7abb7c2a5036c5839
|
|
Change-Id: Icf73d7fd29900a59652b32060e0d66db5c75ad32
|
|
Change-Id: I04cda5ba5b248ef7938d36b5af6003ac89642e42
|
|
Change-Id: Ie2c3dae7ebe3181cb8bb85869882a4eed4a07c14
|
|
Change-Id: I7c4cb1a5f31ba9e37a280af2243a13c57914cb2f
|
|
|
|
Earlier fixes to AquaSalGraphics::GetGlyphBoundRect() are
ineffective, since this method is not even called from ATSLayout.
Change-Id: I877fe5711167f6e28ac7a8153826878359f8ed42
|
|
This is a partial revert of 0ec45dc41dd4872482b817387f253d6fab6bf20c ,
mainly for tinderbox nightlies. The .zip for dev-install is built
only in dev-install target. But IMO the build target is broken
this way too, I don't see why the .msi has to be built already there
and not in some install or make-msi or whatever target (it's pointless
in build target for a developer build).
Change-Id: Ifd63066499b67fa446127193b243d893d497b451
|
|
Change-Id: I086596bc54cada121f5eaaf8f4c0f12e836b3ac9
|
|
Change-Id: I437559c82d99adba2221f7511a7b879682b4b67d
|
|
Change-Id: I63843c335e4c0c3f7ddd15a13853c83db3218649
|
|
Change-Id: I663127446e7d9a38693ce46a7a42b5bec8cb8d51
|
|
Change-Id: I7b90d049cb1a809c028fdf37577f3b8c8b2c3835
|
|
Change-Id: I042c773f071d336c7d402ed7bca2a33b3a998d50
|
|
Change-Id: I15337276f21febba44b84d3618fedfc08013a3da
|
|
Change-Id: Ie66069a37c70b43cbee34130b5b63967a1c9f033
|
|
Change-Id: Iabfb50454c573ac9c01aba1b309822c4ef6fcc73
|
|
like dev-install except that it does not use symbolic link even on linux
Change-Id: I5d8e96caea5070e50765d6cc498c7120747ab1a1
|
|
Change-Id: I1577c1bb8a3dc498b647b930d405c90f80eafaf9
|
|
Change-Id: Ib14cf2731a451226deb4aa27fd05420a0daaba4f
|
|
Change-Id: I6aeb437a034c8c9e5ab7583d608d098912068f72
|
|
Change-Id: Ie1a9227189ec1f3c2de1b81bec99b4267a5fee1f
|
|
we use perl to determine the absolute path of argument thought to be
fiels. but sometime they are not file at all and that call 'fail'.
The code is meant to deal with such case, but
on some platform that produce a scary message on stderr in the
middle of the build log.
This hide these messages.
Change-Id: I52d43e0b26847ab091d76fd446a05c4d84836a77
|
|
Though it is unclear why everything is wrapped into SmExpressionNode. But still
it is so, the patch will work.
Change-Id: Ib62f1a148c737ad110f659f73181ed51e02b8594
|
|
Change-Id: I37816eba909caa71aca95c5181daeea1751b6c5a
|
|
This reverts commit 4e5c55afee36d02bdfe4c0fb09e3b42e9eadb6c5.
|
|
This reverts commit 59b32886a48954f9dad96b3329d172ac73aaa065.
|
|
Change-Id: Iaf58915ab5c56ef5707f3a90a308c8b57cdc4fe9
|
|
Change-Id: Ib58a48c3cbd8aaba38f053bf0a646d6c267f7909
|
|
Change-Id: I2e258ad84aafd17121f86bdc8b8b1aa24b24e709
|
|
Change-Id: I71f923112e932b9cbee9f084581b7a5cdd878510
|
|
Change-Id: Ia53fcbde6b3ff4d38b477069a1f956c9aef84f78
|
|
Change-Id: I97a554169637740aab4282b303a32f476f294e2f
|
|
Change-Id: I215a33d6a1e38ad3edcb92e160e90f7d04db08f1
|