Age | Commit message (Collapse) | Author |
|
Change-Id: I8b424a21b7a1fcc72b607fec00cbfaa6a6961c8b
|
|
Change-Id: I1b5594173b912c8ac54bb54718ce06bee47e2ed9
|
|
Change-Id: I6e027027e78770e32aa484bb1598ebd8c2f291de
|
|
In some LibreOfficeKit use cases it will be useful to use a separate
(initially empty) user profile each time, instead of whatever the default
might turn out to be. (When using the "instdir" of a LibreOffice build tree,
the user profile is the "instdir/user" directory.)
Also add a corresponding new function to be looked up, libreofficekit_hook_2.
I did not bother with any more descriptive name. After all, "lok_init" already
is quite terse, so calling the new function "lok_init_with_user_profile" or
something similarly verbose would in my humble opinion have been
inconsistent. (And if/when we need to extend the LibreOfficeKit initialisation
function with even more parameters, the name would become really long.) But
feel free to change this if you feel like it...
Make sure to stay backward-compatible with source code calling lok_init() and
with binaries looking for only the libreofficekit_hook entry point.
Change-Id: Ifa9ce8f72c2f60554fb3431d522e5784afa8d8d3
|
|
Change-Id: I05175f0d8c37994658e0dcdf355a753bf7c840b8
|
|
Change-Id: Ib537d5e4a0c7c9ed6a6b2520512d32da6c651188
|
|
Change-Id: I93f331bd1c57bd016948e3ed784530724a1028b4
|
|
Change-Id: I248ac1d116ffabc8346fd000c2ae4b0b9c6bd89b
|
|
Change-Id: I2d4e534444678ddf8243db7b3fb6fe66af9fa24f
|
|
Change-Id: I10f7806288f817e599876f95af4998ff6f539fc3
|
|
Change-Id: Iab87cc3fb2b21b53303e46a227f3a1ae35a3ab6d
|
|
Change-Id: I82fcb95c972defde4a854d4c4cce80e6e561bc92
|
|
Change-Id: I3ea00ddec1a740cbcff9eb22727b1f8ce32bb373
|
|
Change-Id: I389a0640d6ffefa1b98a9feff86c73775a7771dc
|
|
Change-Id: Ie12264c02d178be9bc5752e7ea20b4feff409e68
|
|
The #All keyword can be prefixing a column range anyway, so needs to be
treated like all other keywords.
Change-Id: Ic64a5caeff1c6b2d116b21c0c855c7207d649a8d
|
|
Change-Id: I19ec176abf3291b2d4ad5baff06f66518680b378
|
|
Change-Id: I5009df42fcb640ae6e2808900d7d42bd6dcc8677
|
|
In this function the anchor is not set by the time CopyRange() is
called, so the check always returns false, but in worst case it's O(n^3)
wrt. the number of fly frames in the document, so it's a significant
cost to perform that check.
Before:
$ time OOO_EXIT_POST_STARTUP=1 instdir/program/soffice.bin --headless /home/vmiklos/orig.rtf
real 0m46.833s
After:
$ time OOO_EXIT_POST_STARTUP=1 instdir/program/soffice.bin --headless /home/vmiklos/orig.rtf
real 0m26.832s
Change-Id: I7ab9565634e4df7e0adc701908815f778dc59d32
Reviewed-on: https://gerrit.libreoffice.org/15197
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
|
|
This fix is regarding 9835bb562 "MM: export the SwDocShell mail merge via UNO":
In case of using the new output type MailMergeType.SHELL, the UNO
MailMergeService returned a XTextDocument instance that was not
correctly initialized. In particular after calling doc.getCurrentController()
(with doc = the returned instance) we recieved NULL. The reason was, that the
previous implementation used to create a new SwXTextDocument instance
instead of just returning the BaseModel that is already associated with
the internal SwDocShell. This is fixed in this patch.
I also tested this patch with the unit test sw/CppunitTest_sw_mailmerge
and everything seems to be fine now. doc.getCurrentController() no longer
returns NULL.
Change-Id: Ic116f8f25ab9686ef56950e97ba202a55ab81fb8
Reviewed-on: https://gerrit.libreoffice.org/15207
Tested-by: Michael Stahl <mstahl@redhat.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
|
|
Change-Id: Ic4d59228f3295e8400bcacf637ff4def7669b775
|
|
Change-Id: Ibd74c58efa7f1a24de409820655fb98b4fc13df3
|
|
Change-Id: I3f47273f7b648e8c8d261217ba3f9ea8d366b61b
|
|
Change-Id: I0d306854ed295826418ad88ea5c0c72b949eeb22
|
|
Change-Id: I0d2bcf306a789f1eaa0760c69ced427c0ec70ef8
|
|
Change-Id: I801aaa8ad9a4ff08dedd2f92b09d98c870c725b8
|
|
Change-Id: I87b5122b8f1f4e26ba16ebc1228fb77f8b7800d9
|
|
Change-Id: I2becc1cbb875270e2a1e8ed6b7cb0f9b6bdbf2f9
Signed-off-by: Stephan Bergmann <sbergman@redhat.com>
|
|
...and css::uno::makeAny<css::uno::Any>() was never meant to be used. Introduce
css::uno::toAny for the (template-code) cases that shall return an Any for both
Any and non-Any inputs.
Change-Id: Ifa977d73f1da71b2fedde7e8140b19497c4a0257
|
|
Change-Id: I7d3c36e564b3a5286ebd32527575472313119587
|
|
Regression from #i118485#.
Change-Id: Ib242ca1417f46b731d62d117132cbed6c02448fb
Reviewed-on: https://gerrit.libreoffice.org/15173
Tested-by: Jenkins <ci@libreoffice.org>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
Change-Id: Ie2c072c67ecaa0c0ec45c804ebbd4b3bcd631a13
Reviewed-on: https://gerrit.libreoffice.org/15172
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
Change-Id: I9522529d918e143aafa6ea368fdb3fa55416d58b
Reviewed-on: https://gerrit.libreoffice.org/15176
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
Change-Id: I650a5112511fd57099be3f0a84314d0577deae22
Reviewed-on: https://gerrit.libreoffice.org/15175
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
...in accordance with
<http://www.openoffice.org/udk/cpp/man/spec/textconversion.html>
Change-Id: I62013f89c421722770123db8a5794e63d3572e6b
|
|
Change-Id: I138919fff6e78ab8f751c5363ff73d6830f40222
|
|
Project: dictionaries 7991db3dd76ae9fe15faadbd25efb9bceed44e63
|
|
introduced by my commit 216fdcbdb4c17f349ffd68abd1f372bea3287fad
"convert RECALCMODE_ constants to scoped enum"
Change-Id: Ie7db053741205faeac70e3d21757774567f43ce0
|
|
Change-Id: I88fa3672a1f933ae818368c9bc400c6a845babb6
|
|
Change-Id: I083f1b7f21b18c5067b01c82fa8fec467a3d0c50
|
|
and move it inside the linguistic module, since it is not used
externally
Change-Id: I5e405bd921c090277e75e3bcff81638af0845ece
|
|
Change-Id: I402e9073f15dccb3618c06e355559b9f8ff7f763
|
|
Change-Id: I4df3b1e1886e77f7314e034550907ffca58b9ce6
|
|
The problem was that in SwViewShell::Paint(), in case
CheckInvalidForPaint() returns true, we don't paint anything, but rely
on invalidations to call Paint() again and hand out an incomplete tile.
For tiled rendering, we want to hand out complete tiles all the time, so
call CheckInvalidForPaint() early in SwViewShell::PaintTile() that will
ensure that by the time SwViewShell::Paint() is called, the layout is
ready for that tile.
Without this, when opening any larger document with the Android viewer,
about at least once out of 5 times an incomplete tile (e.g. almost
totally white) was shown for the first page, and the user had to work
around this by zooming out + in again to force a re-draw with valid
layout.
Change-Id: Ic7d3a6ddbf9f04beae9ecaa9563d3f5b4d22cc5c
|
|
Change-Id: Ice2e45a3db59adc34a160b4b410cc51c750835f2
|
|
Change-Id: I7d7d516f37cbb5e28a842bbc6bff54916d450dbb
|
|
Change-Id: I1d3c144b2438776a31a246a2d0d4fe57b0caeaa3
|
|
With the current behavior, showing the cell cursor is counter-intuitive and
counter productive - we focus more on the editing itself.
It is still possible to select just one cell though - it is the cell selection
that starts and ends in the same cell.
Change-Id: Ie6d96783d486c6fcda5b4c4f6acf91c1926cdf6c
|
|
When moving the handle outside of the cell editeng, we want to turn the text
selection into the cell selection.
This patch includes work by Henry Castro - thanks!
Change-Id: I081480b7d8b0fde8276f63c261d91bb9f2dc3b96
|
|
Change-Id: I6f28ab1e4af91b12005ddfb9e64e984c783081ff
|