Change-Id: Id7cf80e1da87a56dee645dc01e64dedc4a8586ab

Alt+O and Alt+C were assigned twice (Ok,Cancel), Alt+C did not jump to
the category listbox no matter how often pressed, Alt+O alternated
between fOrmat and Ok.

Still, Alt+C does not trigger Cancel!?

Change-Id: I89b027fb33f39a929cc0fc3345a08c711558505e

Change-Id: I19b7105b62464234cb72de701b07176f951a7951

Add LegacyProvider to read the old format (requires a provider Manager, to
resolve singletons' bases, to decide whether they are interface- or service-
based).

Replace registry-based reg2bin with provider-based reg2unoidl.

Change-Id: I5865e62308cc2d9c5439211ac803d84e93aab656

The XInputStream is guaranteed to be there always
and if it is not there, do not try to create it.

Change-Id: Ie133489dfcb3d9cec0c4efe78740348a1efa2423

Change-Id: I8706d08ab8abe05ea76ea65210726911e446c20b

Change-Id: Ie9bf8ea60f1c0595aa3ac2e0f9b3e6505af30e6c

The comment provided by git starts lines with '# ' (space or tab),
so warn if a line starts with # not followed by a space. It's most
likely something like '#ifdef UNX' or AOO's '#i103131#'. We already
have commits where git silently stripped off such lines.

Change-Id: Ic366d8ee64207edb8bb2fb1ef3a6a192f55872d8

...see the condition in neon's src/neon_socket.c under which
USE_NONBLOCKING_CONNECT would be enabled.  This should not have any immediate
effect in LibreOffice, as we do not call ne_set_connect_timeout.  But it might
be beneficial to selectively do so in light of rhbz#915743 "thread
deadlock/slow join in insert->hyperlink in impress," to work around neon's
disability to abort potentially blocking calls.

The enabled features are hopefully POSIX-y enough to allow for plain #ifdef UNX.
If not, we need to adapt this to more fine grained conditions.

Change-Id: I6bf0a89c1ae73d7183f515e780fff7fe0ae2fed0
Reviewed-on: https://gerrit.libreoffice.org/2492Tested-by: LibreOffice gerrit bot <gerrit@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
Tested-by: Stephan Bergmann <sbergman@redhat.com>

Change-Id: I8c23fadc550091c834a873b53dbd3eae2ba8286e

Reported upstream by Kendy (r5626), for now do not use the field
if not present, seems to be not that important.

Change-Id: Ic947ceef71c86e31fbf0bce74b064aaf91f47178

Word typically uses the Symbol font to describe bullet characters
instead of using a sane Unicode value, the previous can only be handled
if we parse the custom font set for the list.

Change-Id: I1491f07c40953949e381a035c1596c207cdc4c35

Change-Id: Icd0ba0bcbf519a15006af2e0eb176c37766345a3

Change-Id: Ib81753bbe26050e2f108c30a1e596a7c34d72fa3

Change-Id: I3a78903b5bdc81e751570c32dca93e323e8053e7

regression since 76b3e61b

Change-Id: Idfc7789b402aba0d6ab41875a9efc3b990173af0

If there is more than one widget with the same mnemonic, and one is a frame
label, then we need to sort frame labels before frame bodies in the tab
traversal order. Otherwise if the focus is in the body of a frame whose label
has that shortcut and the shortcut is pressed again, the traversal through
following widgets will encounter the frame label as the next candidate, leading
back to the starting point and not onwards to the next widget using that
shortcut.

Frame labels have type "label" in the .ui, so suck that out to designate which
widget is the frame label at load time.

Change-Id: Ie21ed87867bd0c983981a3a8f3318b3cf598c1d6

Change-Id: I57c01a8ab7ba158c160dce977e3e0e54e5b67f23

...as otherwise the SvtMatchContext_Impl thread can continue to run for
arbitrarily long, and the other thread calling Stop() and join() will block.

However, especially the WebDAV UCP does not properly support aborting commands,
see 260afe56 " neon commands cannot be aborted",
so this is not yet enough to actually fix rhbz#915743 "thread deadlock/slow
join in insert->hyperlink in impress."

Change-Id: I0da899f824763e1b3d19bb5b38d906feb690b623

...from webdav Content::getResourceType, as otherwise Content::abort would be
blocked waiting for the mutex (in code that would call abort, which will be
required to fix rhbz#915743 "thread deadlock/slow join in insert->hyperlink in
impress").  This required to get the odd reference to enum return type of
getResourceType straight.

Also, propagate information about !shouldAccessNetworkAfterException from
getResourceType out to getPropertyValues, to avoid further calls that would
again block/fail.

Change-Id: I8b9d43a61eb4078acb90079c4eb7aa98a59a8983

See <http://lists.manyfish.co.uk/pipermail/neon/2013-February/001533.html> "Re:
About ne_set_read_timeout" for why ne_close_connection is not a solution here
(and currently would not work anyway, for the thread calling abort would block
on the aGlobalNeonMutex introduced with 510da291
"forced to make certain neon api calls thread-safe due to cups" until the
blocking thread was done with its neon call) and "a new, dedicated, thread-safe
ne_session_* API call will be needed to allow aborting of a running request."

So enable DAVResourceAccess::abort again and instead disable the underlying
NeonSession::abort.  i#106766 "Crash when cancelling extension update" does not
give any good rationale why DAVResourceAccess::abort had to be disabled, so it
might well have been because the underlying NeonSession::abort was bogus to
begin with.

Change-Id: If8293b93a960907726208bb2f93c375d83357ed3

...keeping all the defines that had not yet been coverd as undefined for now.
(HAVE_SOCKS_H curiously has disappeared from neon 0.29.5 config.h.in even though
src/ne_socket.c still references it, but we had it undefined anyway.)

In general, this hard-coded config.h is a bad idea, though, of course.

Change-Id: I5337f1849b776b62676cee7794917742bcf35356

It is similar change that we did for the release builds.

Also fixes make-dev install on Windows.

Change-Id: I7d9a4bc2e7b14e79516acdfb90c709dbde1378ab

Change-Id: I3ef372b07f2bacc2b9bbbb034e3d61b9cc8a3677
Reviewed-on: https://gerrit.libreoffice.org/2486Reviewed-by: Andras Timar <atimar@suse.com>
Tested-by: Andras Timar <atimar@suse.com>

Hey,
there are more translation patches to come in the next days...

All of my past & future contributions to LibreOffice may be
   licensed under the MPL/LGPLv3+ dual license.

Regards,
Stefan

Project: dictionaries  7e4239060266bf238b5e6692ed10d548c37572d5

Change-Id: I13c1665b0d4b266256d225c1f7b213d18f7ca0ba

not the associated entry data, which isn't set for these simple
1 to 1 position -> data comboboxes

Change-Id: I0c73786bca00684cedc4672d0ad5e61d0c1f9bd7

#i103131# in doubt treat unicode variation selectors as invisible(cherry picked from commit 179f88df)

Other changes include:
in bridges/test/testcomp.cxx duplicate 'using namespace ::test' removed,
'using namespace ::rtl' replaced with 'using ::rtl::OUString' and
'using ::rtl::OUStringToOString'

Change-Id: I67a5952afd305adeb07f728c9074620ecba8e9fc
Reviewed-on: https://gerrit.libreoffice.org/2475Reviewed-by: Andras Timar <atimar@suse.com>
Tested-by: Andras Timar <atimar@suse.com>

Change-Id: Ia2302ee8acdf638bf001f0a2663e09688c21c505

Change-Id: Ia9d120348787bd566a3c9049f7adaf76648d096e

Conflicts:
	vcl/win/source/gdi/winlayout.cxx

Change-Id: I564e3d2186578041a70f329810173b82c8f1ce5b

According to the header file ww8struc.hxx the fHidden member is at bit 1
(cherry picked from commit a428571a)

Change-Id: Idb9fda589b309bd850ba2fdd68e7170c28d4c31a
Reviewed-on: https://gerrit.libreoffice.org/2448Tested-by: Olivier Hallot <olivier.hallot@alta.org.br>
Reviewed-by: Olivier Hallot <olivier.hallot@alta.org.br>

This is a build fix for Windows, otherwise packaging ODK fails with complaints
about lines in .ddf files being longer than 256 characters, caused by doxygen
using very long names for complex types (such as the rtl stringutils.hxx
helpers). It'd be possibly cleaner to avoid somehow generating docs for these
internal types at all, but this should do.

Change-Id: Ie9bbeb38ec9cc2fa0377709e75abb0338b20bab5

Change-Id: I652653ae8a99b8486e893c65b6c82b6ea6e4faec