404f16e9 is embarrassing crap; the real
problem there is that the paragraph that should be removed is _inside_
the frame, but in sf_68e4f217c107f6eada21adfe6e3c9c57-82055-minimized.doc
the MoveInsideFly() is not called so a non-empty body paragraph was removed.

Change-Id: I5d4443804c52a5fa30ddb51609ccbbee07abc5aa

Change-Id: If189a60ae48a64cde109e52903c9a4b8b9fde85a

Those are supposed to obey the current setting for "Conversion from text to
number", and the OpenCL code does not currently take that setting into
consideration.

Change-Id: Iff6eeda6238c78f4aeb8eb27898ce7ea243b9dea

Also changed the description to match the meaning (after
2021275f).

Change-Id: Ide4b4f094495c178252c3167d52012f46f26b025

I believe the intent was to "retarget unresolved pLogCluster[n] to a glyph inside the cluster".
Unfortunately in the loop that detects clusters there was a typo and we are only indexing
element 0 of the array:

    // retarget unresolved pLogCluster[n] to a glyph inside the cluster
    // TODO: better do it while the deleted-glyph markers are still there
    for( n = 0; n < nCharCount; ++n )
        if( (p = pLogCluster[0]) >= 0 )
            break;

That just doesn't make any sense, I believe we should be accessing pLogCluster[n].
If not, then why not just do:

    p = pLogCluster[0];
    n = nCharCount - 1;

Change-Id: I9d8873541b5c794071d69c0f63df88f17a352904
Reviewed-on: https://gerrit.libreoffice.org/13876Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com>

Change-Id: Id856d8dbf437acb7e528e60b4d1b012102ff0557

Change-Id: I05923b7df753fdbc23f849c89f4f9f52a07915c7

Change-Id: I7701368e152ce20adbc05802c7c6518183919d1f

Change-Id: I30b51510ac5fbfb3c1dd09e0f86a246a8b516f1e

Change-Id: I8d23c11decca864337ccc5288165058a7c21cd4e

Change-Id: I3f50d3bb85bebdece692e069de73febc2ce63b47

... if the merged range ends behind the visble range.

Change-Id: Ib5e9ae57299614f42c9200da5641922dfd90ab60

Change-Id: I9d05a3971bd0dd14654d0b0e2a80a263d360505b

Change-Id: I86dd6a96f69d9bedeabeebc04f953b9f22476543

Change-Id: I1ca574f96ad84aa263bdcecbe549687fbf10d0d8

Change-Id: I018cebeeeee09c294ab576dd3cb5e7d6c703f5f2

When left most cell is not in the visible window, Visible Range
was not checked merged cells.

Change-Id: I3c375c0f584e29444472224c74442bbd3d97b3ac
Reviewed-on: https://gerrit.libreoffice.org/13832Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>

regression from

commit e0cce521
Author:     Armin Le Grand <alg@apache.org>
AuthorDate: Thu Jan 10 16:28:40 2013 +0000
Commit:     Caolán McNamara <caolanm@redhat.com>
CommitDate: Thu Jun 13 14:50:46 2013 +0100
    Resolves: #i121504# Support for alpha channel in clipboard for all systems
    (cherry picked from commit ef3931ff)

which blindly just bulldozed out the bMSOFormat branch

Change-Id: Iec354f1fb585f0803b9df472bc9ec9e103aa5847

Was missing from 82bc764bc9, without this opengl scaling wasn't used for these.

Change-Id: If853435b12383e799afe2f057c43263b2d2297fa

Change-Id: I52590f60882682943dda354599ab4ed762383b51

As a side effect, this prevents the lines from looking too wide,
because of generic polygon drawing being a bit too generous with AA.

Change-Id: I17314c39fd57e33ecbd10b8a8785c600bdc5b212

I'm somewhat confused by why there needs to be a separate line width
for X and Y, but apparently there is, and the latter shouldn't be just plain 0
(otherwise a number of drawPolyLine() implementations either skip using
a simpler path for the usual case of them being equal, or even plain to refuse
work at all and cause a fall back).
And I hope this doesn't lead to finding out that some of those implementation
are actually buggy.

Change-Id: I2dbbd1539c4a96d41935cce9ae6565872e2a459b

Change-Id: I66a04febdbfa673e0883ab6f574bb7768cad7953

118529d4 and 01190389 might have been technically correct, but
'mProgramIsSolidLineColor' is clearly a misnomer at best, and they'd
have made it even more likely that this would break yet again.

Change-Id: I1f01663e2abc0b1b0e557ae7241637a96e1a402a

Change-Id: Ifa69f3272ebda2a61ac00d2affb8aebd4524f0fc

...cf. 52ce5239 "vcl: assert when LogicToPixel
overflows."  Looks safest to use platform-independent SAL_MAX_INT32 rather than
INT_MAX.  Detected when running CppunitTest_sw_uiwriter with
-fsanitize=singed-integer-overflow.

Change-Id: I7df98237811a304e475d6b2db8805bc390ca6232

Change-Id: I55b69f153c0656d34a230c4328bc7af4bab40447

Change-Id: I250a683faf7d201b9fea8eb0fffd9cd49da7f79f

Change-Id: I2681b99c913b15c632dd4c9ec84922fa282692d7

Change-Id: Ia60194f9789324bc484bfa609c6eb92572b8554d

Use a so-called counter-based random number generator. Code from Random123,
http://www.deshawresearch.com/resources_random123.html.

Change-Id: Id47f84ef18eada64dcf47762a61ec3856c71760e

Change-Id: I12d230176ef1ea232ac9a401fbbebce6d8c058a7

and dump the ones that nothing is listening to

Change-Id: I253ef284df785812a439dd160edba1b07fdbaac4

and drop DATACHANGED_DATETIME because no-one is using it

Change-Id: Id5ac9a7fbba0e35501ed82e5252f66858621f7ff

and their default implementations are empty, so just delete them

Change-Id: Ibae2f92c3326ad46c4b6ef462b5b7b62ad63f0d8

Note that getPackageName() also throws NameNotFoundException, so in the
unlikely situations in case:

- package info (class containing the package version) is not found or
- the package version is not in an "a/b" form

We still just don't show anything.

Also, mark the new TextView as android:textIsSelectable, so it's
possible to copy&paste the version for bugreport purposes.

Change-Id: I63b53cca4126da17bfbda0293d7c98e8524ef41a

Change-Id: I922ce1f735df236d2bdad33820ed98f18fc85963

Change-Id: Ia65eebd153251985fe736e8f278d46b22a06ec83

Change-Id: I7a11b0eea71e96f8bd2b307539f1964acb240216

Change-Id: Id910c1905d23f16cae5b6b551c74193d46cff256