There is no reason for the generic UnoApiTest to require Calc
specifically. Calc tests can/should instantiate a Calc instance.
We can create a CalcUnoApiTest for that that inherits from
UnoApiTest; however this does not seem necessary, "make sc.clean"
succeeds.
Anyway, the ScGlobals::ensure mentioned in the comment does not
seem to exist.
This allows us to eliminate some code duplication in tests
that were reimplementing UnoApiTest minus the Calc instantiation.
Change-Id: I37bea9df41e3960df0458fe689cf6c046a243617
as demonstrated by abi10075-1.doc
just register the AnchoredObjects in the SwLayoutFrm and
inform them when the SwLayoutFrm goes away.
This crash was triggered by "1e113cb7604e1509e7d598a9be329f1f7b6e9322" import
different first page header/footer from doc. But that commit is blameless.
Change-Id: Ia079cc635a81dff1ccbf740641f441aa784328a4
...to not use WildCard (in case a trusted location URI already contains an
unescaped "*"), be specific about matching only past a final "/", and rename to
isSecureMacroUri for clarification.
The check with an INET_PROT_NOT_VALID default INetURLObject in
SfxApplication::OpenDocExec_Impl ("we have to check the referer before
executing") had efficiently been dead since its inception in
14237ac4bf497decdde8b742acea23780833ba12 "#90880#: security checks corrected,"
as INET_PROT_NOT_VALID is considered secure regardless of referer anyway.
Change-Id: I03bca5e6dac89bb2aac52909aff273ea640228d8
Previously inserting a new slide required using a context menu
(via right-click) or the main menu-bar. It is now possible
to insert a blank slide by double-clicking in an empty area
of the slide sorter.
Change-Id: Id7ac5f9d3befd02d0a89ad5e1631885c0493ad3d
Reviewed-on: https://gerrit.libreoffice.org/6442
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
When nPos == USHRT_MAX, even if later we checked for that case, just for
the sake of an assert(), we already called operator[] with nPos.
(To reproduce: right click on List 1 -> New -> Cancel, results in a
crash.)
Change-Id: I465b5519000b79bd8b2eaa3b3c1ceb80a182a25a
CRITBINOM could not handle large sample quantities, e.g.
CRITBINOM(1080,0.5,0.1).
This patch fixes that and also improves the efficiency of
the function by always iterating from the end nearest to
the cumulative binomial distribution.
Change-Id: I35f965acc83e26141fb2cf5a7e3303ec0791f04a
Reviewed-on: https://gerrit.libreoffice.org/6559
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
Apparently the $$languagesref in create_directories results in the full
path to services.rdb when called from mergemodules.pm as
my $unpackdir = installer::systemactions::create_directories("cab", "");
... which apparently means that there's some variable whose name is the
empty string and whose value is the path to services.rdb.
Add some mysterious workaround suggested by mmeeks.
Change-Id: I6ba203f5cc44786969795eebab5359d5e9a8a12a
Translations may have omitted the %1 placeholder in SCSTR_COLUMN and
SCSTR_ROW, or not have it translated newly yet because it was introduced
later. In these cases append the column indicator / row number as it was
done before.
Change-Id: I331cf0c529bebea02277784ea145c202eb6f6fda
...so that it is still found after 7b3133ad882b24834f87b478d00bf4ee92d3cd63
"instsetoo_native: shouldn't need instdir in include path." And no idea why
getuid was in PLAINLIBS_URE instead of PLAINLIBS_NONE.
Change-Id: I5de2a79cbd3c0064cd7776e269f4792304ec34e3
