Age | Commit message (Collapse) | Author |
|
Change-Id: Ide189c5f3ace0a5a7478ae1a70c315fe3e461667
|
|
For Windows, superseded by Windows Installer patching (i.e., creating
.msp files), which is something completely different. (And quite hard
to get working... but still a saner approach, I think.)
For Linux, many distros use delta RPMs or similar, so no home-grown
LO-specific patching mechanism is needed.
Remove the -patch and -patchinc command-line options to
make_installer.pl and all code that was invoked only when using those.
Remove the PATCH and PATCH_ONLY flags in scp2.
Remove the patchmsi.dll Windows Installer custom action.
Change-Id: I09e949e601a969f88eff60067faa2352f4f89537
Reviewed-on: https://gerrit.libreoffice.org/1605
Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Miklos Vajna <vmiklos@suse.cz>
|
|
Moved from translations' custom makefile.
Change-Id: Ife58bfa3ee607b06617d7db80408afb453b9a10c
Reviewed-on: https://gerrit.libreoffice.org/1622
Reviewed-by: Miklos Vajna <vmiklos@suse.cz>
Tested-by: Miklos Vajna <vmiklos@suse.cz>
|
|
Change-Id: I467ad9307536e6cdd32ebcdd40539bd5f8614c85
Conflicts:
dbaccess/source/core/dataaccess/documentcontainer.cxx
|
|
The asm code loads values into parameter-passing registers r0-r3.
(That is one of the very purposes of the asm snippet.) We need to tell
the compiler that. The compiler does not analyze the asm snippet and
has no idea by itself what it does.
Otherwise the compiler might well put one of the input values to the
asm snippet, like the "pmethod" (the value of the pMethod variable)
into one of those registers, so that when that value then is used in
the asm snippet, *after* r0-r3 have already been modified, it
obviously is totally unrelated to pMethod any more, and the result is
that the code jumps into hyperspace.
Apparently this has worked purely by luck, or thanks to GCC
conservatively avoiding using the r0-r3 parameter-passing registers in
this way. The problem was noticed when using the same code with Clang.
The above analysis tentatively confirmed by Caolán and Jani Monoses,
who wrote the code.
Change-Id: I3018c2e2ccb83e7a71144425fa404ad28bb955d6
|
|
Change-Id: I2b02b62311c00fc47892ea7e24d8bc01ac4fb149
|
|
Change-Id: Ibd92f5ebe3079dd0b75a0f429efc68842a4305ec
|
|
Change-Id: I00d127dbb6151e3100be931079cbf8907ac69ddd
|
|
What happens here is that the xCursor->goLeft( 1, true ) selects the
entire inserted table, which is then deleted. That causes annoyances,
such as an assertion from ~SwIndexReg because as collateral damage the
SwTxtNode that SwReader::Read's pUndoPam points to is also deleted, as
well as missing tables in the document.
DomainMapper_Impl::RemoveLastParagraph() can avoid this by not using
setString() but instead disposing the paragraph, like it already does in
other cases. But don't do it for new (loaded) documents, because it may
kill flys in that case, which i strangely couldn't reproduce when
pasting as RTF.
Change-Id: I87302b8916e39ca42bc4fedfee40ac38fadd0939
|
|
These do not occur as such but while a node is being destroyed it is
still in the nodes array and may have some base class types; without
giving these a value gdb will just stop printing the nodes array then.
Change-Id: If78c4f633312cf8bad0048b704fe20bd09c7c28f
|
|
My self-built 4.0.1 on Linux handles this fine, but 4.0.1 is not used
for LO building except for MacOSX. This specific operand combination
is rather unlikely, so this shouldn't actually cause any problem
in practice, hopefully, but I'd still like to check this where possible.
|
|
|
|
On mingw sal_Unicode is not wchar_t (see sal/types.h).
|
|
::rtl:: prefix removed from OUString as well
Change-Id: I4de29a9c487ff2b87fdb5efa8e75aaf5b88402b0
|
|
Moreover, according to MSDN, wcscpy() in C++ is a template that doesn't
need the length passed, so this shouldn't make actually anything worse.
|
|
|
|
|
|
Change-Id: Ia49f0c8982e901fdb869d451cc3f272f629afc81
Reviewed-on: https://gerrit.libreoffice.org/1384
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
This appears to be broken at least since
bde5b815873450dc9520c4dd300829eb5bd72daa "INTEGRATION: CWS jsc21 (1.11.2); FILE
MERGED: 2008/05/27 16:59:34 jsc 1.11.2.1: #i88797# adapt for new structure"?
Change-Id: I6385ca10cdded13e6d1aaf08d353f5029843d38d
|
|
Change-Id: Ife2d92131f7cd758628ae686cde734dab56e4499
|
|
During the import, a cached matrix value only has an empty matrix
of correct geometry, plus the token type of the top-left cell. The
rest of the elements are imported as hybrid values. For now, this
seems to do the trick.
In the future we may want to change it to fully populate the matrix
cache value during the import, and skip setting the hybrid values for
the non-top-left elements.
This commit also make several other trivial changes:
* Mark pRawToken mutable so that we can mark those IsFoo() methods
const.
* Move the ScCompiler instance from static instance to member of
ScXMLImport. Since we don't need the instance to persist once the
import is over, this is more appropriate.
Change-Id: I1abde03c0fcd91b02ef4dbf8b5526f7965eaf19c
|
|
Change-Id: I2a4381cbe1e716cd2de1fb596053646d1a74d4b0
Signed-off-by: Marcos Paulo de Souza <marcos.souza.org@gmai.com>
Reviewed-on: https://gerrit.libreoffice.org/1572
Reviewed-by: Luboš Luňák <l.lunak@suse.cz>
Tested-by: Luboš Luňák <l.lunak@suse.cz>
|
|
Partly reverts d7a5ec62e91ce3dc5b784815254218f16181f676
An empty locale was treated as LANGUAGE_NONE, effectively being
interpreted as absence or undetermined or multiple or all depending on
context. Restore this behavior as it was an undocumented feature of the
API.
Change-Id: If4b1641e776d10dea0d3037f7a62725c2b8d612c
|
|
Change-Id: I67301dbd5b77cc5439f80b6bf4c241ca953216fd
Reviewed-on: https://gerrit.libreoffice.org/1595
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: Iafccfaa572b945d8d75d926ccc1935320b9d70d9
|
|
Change-Id: Ib03ec188ba30d122ff687937209b989d418fe7ac
|
|
Change-Id: Ib83c02d53fbae6a70781d9d9d7ab015b016c72a8
|
|
In getCrsrPropertyValue etc.; also add a unit test for the problem.
Change-Id: Ibd459a43393c39b4fed9fb89aae4a5f7bacff007
|
|
Change-Id: I67042b5d689457921a928454c9051f0402be17e6
|
|
Regression from 8063e36115a11ddf3db05928db9287947beee74d.
|
|
Change-Id: If6cc5f6d5e502c1fc5da210179263be63fc8c8a6
|
|
It was discovered that on a Russian Windows, by default only Russian and
German dictionaries would be installed, but no English dictionary. This is
because since e2fac98819c00b4fb50f9de9d0f32d20092f3191 "fdo#53009: For msi
installer, only default-select a subset of dictionaries" only those dictionaries
mentioned in setup_native/source/packinfo/spellchecker_selection.txt were
default-selected, but it apparently was a misunderstanding that
spellcheker_selection.txt for many locales does not list "en" even though an
English dictionary should always be default-installed.
Quoting Andras Timar <atimar@suse.com>: "AFAIK en-US localization with English
dictionaries is always installed in all 3 supported platforms by default. On
Windows default install should always select en-US and English dictionaries. If
it does not do so under certain circumstances, then it is a bug in my opinion.
We removed en-US entry from the spellchecker selection file for this very
reason. It would have only bloated Mac and Linux langpacks."
So this fix simply makes sure that "en" is included in every line of the
generated spellchecker_selection.hxx that is used by
setup_native/source/win32/customactions/sellang/sellang.cxx to determine which
dictionaries to select by default.
Change-Id: I3db68585dc6269a7db2f6cbf6b79796dc09d05e8
|
|
Change-Id: I073b9601c82a464f705df30cfd7e760c212eb7ca
|
|
Change-Id: I860fb2c91d6843a1d5345ab2cf9270473f73f5bc
|
|
Change-Id: Ie07d41e5dcb09977d89752e724fcbabd6c5d7df1
|
|
Change-Id: I4d914de5ec1d614e30825ad6bd231545896dc130
|
|
|
|
|
|
|
|
Change-Id: Iaa2841b245ccd3de513296a10e33a93bc1cb68a2
|
|
For setting permanent defaults via spadmin we don't want to resolve the
default backend on saving "default", this happens during the print
initialization stuff.
But when we tweak the printer settings for a single print job and reset back to
default, we need to resolve that to what backend we want
Change-Id: Iac2370624b6d248c9658fb0fbc2f2d449849d81a
|
|
|
|
Change-Id: Ib12a0726883ee1aaac87d4d02c7bae595f7f9e39
|
|
Change-Id: Id7c030410db0a67c22f2cbaaed417c4371bca361
|
|
Change-Id: I4fe95db707cdaafb220435f281cce88284afd26d
|
|
rework EventBox to inherit from VclBin and rebase SwFrmCtrlWindow on top of it
Change-Id: I25f037b33bf244d3d39f57bfe11cabfaf992bf1c
|
|
when we really want no horizontal scrollbar, and
want to be in online layout mode for previews
Change-Id: I2d0e279a6f112ad7edbd6b523ea07b6dd2061ebc
|
|
split the very opaque SwFrmCtrlWindow into an
EventBox which works like GtkEventBox, this
a) contains a EventBoxHelper window which EventBox ensures is
its first child
b) EventBox ensures that all its children are the same size
as the EventBox
c) The EventBoxHelper is transparent and in front of the
rest of the children, so its intercepts all commands
d) EventBoxHelper forwards command to the EventBox
so you can inherit from EventBox and stick a child in there
and the upshot is that the child doesn't get Commands, but
the class inherited from EventBox gets them instead
Change-Id: I783cd6bbd194e28a1744147e8175cba7895158be
|
|
Change-Id: I34fe1d6ab2e36b3c3e7e9199872c56aff184bf66
|
|
Change-Id: Ib42e171ba9f8c0d30739decdd729bc36bc9c777a
|