This one became pretty small.

Change-Id: Ic1e1e2b7afc35e5d1141a34722a6fe395832d936

Again, this version has much less branching.

Change-Id: I9138471261f2d4df14b64e603a88d4495b6ea45e

It's no longer that useful in today's pivot table design.

Change-Id: I665a7707126793dfde17a2221ba563d22a59aad1

Change-Id: I41e4c747c4662fb066746c0ca87d17ea6742080e

Change-Id: I1abdae0f924bf919fbf43e5a07f950a8a10861a9

Change-Id: Ibcbcb69c96e8bbd6bccc0f852240a2a6cef2c1c0

Change-Id: Icb5bb61108f6bb674016ce044ecbe6a366b4f9c6

Change-Id: I9b3f8ca9c5af326735ff9c95359e6a0917d6898b

TODO: Commented out for now the tests that currently fail.

(Also improved the code by calling CPPUNIT_ASSERT_EQUAL_MESSAGE.)

Change-Id: I726089ad046b97ce64fd655269b2be652e78efe9

Change-Id: I3d13ab04b126a3268352df93a7ed4a35d7a085f2

Change-Id: Ie03106c3089f6d1bc0172a2e2c702b5725e55acf
Reviewed-on: https://gerrit.libreoffice.org/1656Tested-by: Kohei Yoshida <kohei.yoshida@gmail.com>
Reviewed-by: Kohei Yoshida <kohei.yoshida@gmail.com>

Change-Id: I2376e69adf235d1e6f89296c50a9cea76041f9d7

Change-Id: Iecc66a0a26c6944bd398f79eb760694405b17e70

set the two additional frames we want to take the height of the category frame
set to preferred size of 0 in order for the set of three to all the take the
category height via the sizegroup

i.e. there are a gadzillion currency formats so show eight of them at a time

Change-Id: I4badd8852fc56553a94616ce8845aa40b22f2dc1

Change-Id: I01015184223960a5bdfb895fedc79e982a408082

and re-use it to get a better calculation of the optimal size of a
widget, rather than taking the current position of the subedit

Change-Id: I85cb3ff98f23d21d7cfdcc28188e36616a19b5e8

Change-Id: I0d1b8176c9796ae39bc692d48e4481bdedb6ff09

Obvious problems are in debug code only.

Change-Id: I45aafb4f194a5a020bb95bd96c2d92ca7ee578f8

Change-Id: Ia8874a881d2a625500bbf74997ab75457e6f6b7e

Change-Id: I2ce959942f95a123d2ec75a7df2194dff11fac64

There are several heap objects that are deleted at the end of the
method.  Returning prematurely would leak those objects.

Change-Id: If70cb3c9c02e47229d10b30f1542f7b25534d593

DataPilotUpdate() is getting more and more complicated, with so many
branches.  Since self-updating a pivot table object is common, let's
create a variant just for that.  This variant has much less branching.

Change-Id: I75b0512258a09961ef13516d8d9bc8a2939287e1

...when you set dbg = True

Change-Id: Ifc170e9336a662dce2ae59227baf3bea692eedac

as the corresponding test is otherwise seen to fail, with user being b, but I have
no idea if this is the most Python-3-ish approach to fix that, or whether more code
needs to be fixed, too.

Change-Id: Ia7fbcbca3cf578ffe1bd5ce3c7c5b709cc77317e

Change-Id: I07f0f47fcbd3d463d7831820cde338d8d59d9b68

The problem was that while SwUndoInsert removed the field characters, it
didn't remove the fieldmark, so after unding a comment insert, it wasn't
possible to re-insert a comment.

Change-Id: I1df4cbb013d94cff7ae5195427044b1e1b2d8895
Reviewed-on: https://gerrit.libreoffice.org/1646Reviewed-by: Noel Power <noel.power@suse.com>
Tested-by: Noel Power <noel.power@suse.com>

Change-Id: Ibb3f8175729cb579a687275a11884c397152fe94
Reviewed-on: https://gerrit.libreoffice.org/1645Reviewed-by: Noel Power <noel.power@suse.com>
Tested-by: Noel Power <noel.power@suse.com>

Change-Id: Ie87648368f877af56878bc83fe3d5ece66f77304
Reviewed-on: https://gerrit.libreoffice.org/1644Reviewed-by: Noel Power <noel.power@suse.com>
Tested-by: Noel Power <noel.power@suse.com>

Otherwise tests relying on layout dumps will parse the wrong dump.

Change-Id: Ib7b4bc3a231eb3082567a1d1ed38c96835d43add
Reviewed-on: https://gerrit.libreoffice.org/1643Reviewed-by: Noel Power <noel.power@suse.com>
Tested-by: Noel Power <noel.power@suse.com>

* reimplemented actual ScIfError() function for matrix cases and simpler
  pop/push of non-error values
* various smaller adaptions, see also inline comments of
  https://gerrit.libreoffice.org/1522

Change-Id: I80a8235a2cbb548c0ef993862e76741af46e30d2

Change-Id: I6403b51ac8c710ad3b8d2625e1482971f50e6b1d

Change-Id: Idff204b862fe7c5a974b7b51ba147b558bb02e67

Its usage was gone with adda1a25.

Change-Id: I055c15f5154c6f78d2314d4b502ca3db751ab891

Change-Id: I8d4f8b95909b91917b456f49d7a3ac270703cc96
Reviewed-on: https://gerrit.libreoffice.org/1640Reviewed-by: Noel Power <noel.power@suse.com>
Tested-by: Noel Power <noel.power@suse.com>

Change-Id: I6734df58b589088809c712c3906f0cee3fd6a103
Reviewed-on: https://gerrit.libreoffice.org/1639Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>
Reviewed-by: Noel Power <noel.power@suse.com>
Tested-by: Noel Power <noel.power@suse.com>

...from d19c40f4 "Work around some potential
problems with thread-unsafe getenv" that had been broken with
60628799 "Notice if SAL_LOG is changed while the
process is running."

Change-Id: Ibd6dbc9921ae1f8dee114380f01a076b0770788c

Change-Id: Ia54040c7ed02e6a2d62dd93890a7d1467dd83a40

result of incorrect merges

Change-Id: I43f5ef2afbd5390f937f1220613ffa7d1f284634

When compiling using icecream, the whole input is one file fed from stdin,
so getBufferName() is "<stdin>", so get the filename from #line directives.

Change-Id: Ie87937b10b0c7e8ddfa57a405b80fd9bde68dd0a