- 06 Ara, 2018 4 kayıt (commit)
-
-
Caolán McNamara yazdı
Change-Id: I73bc93ebb5010f3a376b3b121e1ac873f1be137a Reviewed-on: https://gerrit.libreoffice.org/64677 Tested-by: Jenkins Reviewed-by:
Caolán McNamara <caolanm@redhat.com> Tested-by:
-
Stephan Bergmann yazdı
(cherry picked from commit d1a1c9fa, adapting the pdfium block, which has only been upgraded to 3550 on libreoffice-6-2) Change-Id: Idffd274a18df14a96e61bf61ba09f837fdbbf9c7 Reviewed-on: https://gerrit.libreoffice.org/64670 Tested-by: Jenkins Reviewed-by:
Stephan Bergmann <sbergman@redhat.com>
-
Stephan Bergmann yazdı
It had originally been read as 16 bit via >> in d2000efb "#i106421#: move msfilter to filter", then accidentally changed to be read as 32 bit via >> in 9830fd36 "date/time IDL datatypes incompatible change" (and later changed to be read as 32 bit explicitly via ReadUInt32 with 15535e32 "convert SvStream::operator>> methods to ReadXXX methods"). Change-Id: I5062e67a578d182a0df2726ab8d0bae465f154f3 Reviewed-on: https://gerrit.libreoffice.org/64604Reviewed-by:
Lionel Elie Mamane <lionel@mamane.lu> Tested-by: Jenkins (cherry picked from commit f84f5b0a) Reviewed-on: https://gerrit.libreoffice.org/64635Reviewed-by:
-
Justin Luth yazdı
This reverts commit 85818da3 which caused bug 121110. Only reverting from 6.2 because I hope to fix this for the current master 6.3, but I won't backport the fix because it isn't trivial. Change-Id: Ie91b707fbb05fbc5c6806aef7d1760ab69abb277 Reviewed-on: https://gerrit.libreoffice.org/64660 Tested-by: Jenkins Reviewed-by:
Justin Luth <justin_luth@sil.org>
-
- 05 Ara, 2018 21 kayıt (commit)
-
-
Caolán McNamara yazdı
is that how it should work, or should there be an additional TogglePause created at import time to get the apparent intended pause effect Change-Id: I0573bd8d9534271f2b36d7ab3b8cdc20e0f985a1 Reviewed-on: https://gerrit.libreoffice.org/64548 Tested-by: Jenkins Reviewed-by:
Thorsten Behrens <Thorsten.Behrens@CIB.de> (cherry picked from commit 1aa10fcb) Reviewed-on: https://gerrit.libreoffice.org/64620
-
Eike Rathke yazdı
Basically this was always wrong. Original code didn't treat the r attribute as optional and resulted in an invalid -1 row for such cases, hence the RowModel wasn't applied later. Then commit af8e6f4f CommitDate: Wed May 29 22:05:02 2013 +0200 r is an optional attribute in c and r, fdo#65059 introduced a 0-based SheetDataContext::mnRow but assigned that to a 1-based RowModel::mnRow as well if r attribute was missing, and did not update SheetDataContext::mnRow if r attribute was given. commit ff56553e CommitDate: Thu Mar 13 18:39:22 2014 -0400 fdo#76032: This row index is 1-based whereas our own mnRow is 0-based. then correctly assigned the decremented r attribute to SheetDataContext::mnRow but did not touch the assignment to RowModel::mnRow if r was not given. Probably confused by the ++mnRow already there which is correct but not sufficient for the assignment. Which we now finally fix herewith.. Change-Id: Iae1563ed01b95e7852ddad4fd59374c07186b493 Reviewed-on: https://gerrit.libreoffice.org/64580 Tested-by: Jenkins Reviewed-by:
Eike Rathke <erack@redhat.com> (cherry picked from commit 76fbe494) Reviewed-on: https://gerrit.libreoffice.org/64638
-
Xisco Fauli yazdı
it's still used on Mac. Revert it until XOL is removed This reverts 25830187 This also includes https://gerrit.libreoffice.org/#/c/64596/ Thanks Stephan Change-Id: I06548a590f370618ad640724a1b9c59a3faceec2 Reviewed-on: https://gerrit.libreoffice.org/64582 Tested-by: Jenkins Reviewed-by:
Noel Grandin <noel.grandin@collabora.co.uk> (cherry picked from commit be504e8c) Reviewed-on: https://gerrit.libreoffice.org/64631Reviewed-by:
Xisco Faulí <xiscofauli@libreoffice.org>
-
Miklos Vajna yazdı
This restores compatibility for API users who called getByName() on the bitmap table and expected that the result can be set as a value for the GraphicURL property. The case is similar to the Writer images, which was handled in commit 587803ba (sw TextGraphicObject: handle XBitmap for GraphicURL, 2018-08-08) already. Change-Id: I3740a68989b306425d0d6c4d4e350ac163cb4170 Reviewed-on: https://gerrit.libreoffice.org/64069 Tested-by: Jenkins Reviewed-by:
Miklos Vajna <vmiklos@collabora.com> (cherry picked from commit e30f3e76) Reviewed-on: https://gerrit.libreoffice.org/64427
-
Miklos Vajna yazdı
Make sure that our normal SAL_WARN() reporting is registered during the xmlSecCheckVersionExt() invocation, just like it's registered during other libxmlsec API calls. (cherry picked from commit 099ab643) Change-Id: Ia19368b55dbae47974f48954f86b50e36d1462ff Reviewed-on: https://gerrit.libreoffice.org/64605 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
Change-Id: Ie171eea334de65b993554146c11adb95e0dbb85e Reviewed-on: https://gerrit.libreoffice.org/64245 Tested-by: Jenkins Reviewed-by:
Luboš Luňák <l.lunak@collabora.com> (cherry picked from commit 9a446225) Reviewed-on: https://gerrit.libreoffice.org/64535
-
Luboš Luňák yazdı
It's silly to just return NaN just because the function can't handle something. If nothing else, at least a proper error should be reported (or in this case, the fallback to the core function will take care of it). Change-Id: I9c971082f4c5c9836318cf63d15fa7c278274273 Reviewed-on: https://gerrit.libreoffice.org/64244 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
Change-Id: I885a1b70556311f5a0862fec24a3596dbeada5d6 Reviewed-on: https://gerrit.libreoffice.org/64243 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
sc_subsequent_filters_test in ScFiltersTest::testFunctionsODS() also checks "SERIESSUM(2;3;2;3)", which the opencl code does not handle. At least bail out gracefully instead of returning 0. Change-Id: I154dca8cc437a6225b4eb98012232d80683f0114 Reviewed-on: https://gerrit.libreoffice.org/64242 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
Otherwise the yielddisc unittest fails. Change-Id: I2ebb0fbf9d4aea6678c91d0289418991aa92dbb5 Reviewed-on: https://gerrit.libreoffice.org/64241 Tested-by: Jenkins Reviewed-by:
-
Eike Rathke yazdı
Seems these were always wrong in some locales. Change-Id: I5dfc89a607d57c19034d6fac07a07bf3c5395ed8 Reviewed-on: https://gerrit.libreoffice.org/64541Reviewed-by:
-
Eike Rathke yazdı
It didn't matter originally, but en_ZA separators changed with commit 5c10c448 CommitDate: Mon Sep 3 13:38:13 2018 +0200 tdf#119613 [*-ZA] decimal separator is comma, group separator is space and commit 464154aa CommitDate: Mon Sep 3 13:38:23 2018 +0200 Related: tdf#119613 inherit af_ZA separators from identical en_ZA so the separators didn't match the formats inherited from en_NA anymore. Change-Id: Iccf951b27b5575706e0f6bae12b5df4b53e335ac (cherry picked from commit 7f64ec49) Reviewed-on: https://gerrit.libreoffice.org/64549Reviewed-by: -
Eike Rathke yazdı
Order is 'D, MMM YYYY' instead of 'MMM, D YYYY' 'NNNN, ' displays separator twice. Change-Id: Id84af3e1bb94dad0e8405c680e63f48f38415fba (cherry picked from commit 35dd9167) Reviewed-on: https://gerrit.libreoffice.org/64540 Tested-by: Jenkins Reviewed-by:
-
Michael Weghorn yazdı
Stop reading commands from the pipe on kde5 side once the "Quit" command has been sent, in order to have the thread that is reading commands from stdin finish properly. Join the thread in the 'FilePickerIpc' destructor, rather than just deleting it while it may still be running, which resulted in 'terminate()' being called. Change-Id: Ia184987e7994cc1de0208ff2757a3cf06c8b7194 Reviewed-on: https://gerrit.libreoffice.org/63835 Tested-by: Jenkins Reviewed-by:
Michael Weghorn <m.weghorn@posteo.de> (cherry picked from commit 93815c2b) Reviewed-on: https://gerrit.libreoffice.org/64587Reviewed-by:
-
Caolán McNamara yazdı
Change-Id: Iedb2b11be0e008501273925d6ae7137b28c3b912 Reviewed-on: https://gerrit.libreoffice.org/64538 Tested-by: Jenkins Reviewed-by:
-
Caolán McNamara yazdı
if we want to hide the text direction widget, we should hide the rest of the elements too. For now show them all. Change-Id: Icb2ef69c50eb2335170dda9bfcfcb947ce4d6d19 Reviewed-on: https://gerrit.libreoffice.org/64553 Tested-by: Jenkins Reviewed-by:
-
Caolán McNamara yazdı
Change-Id: I579cef26964aebc89bf6c4786725b4bf85fb4fe5 Reviewed-on: https://gerrit.libreoffice.org/64556 Tested-by: Jenkins Reviewed-by:
-
Caolán McNamara yazdı
Change-Id: I98cdda7e384dce35f02d20d5b98b61168d34f786 Reviewed-on: https://gerrit.libreoffice.org/64554 Tested-by: Jenkins Reviewed-by:
-
Jan-Marek Glogowski yazdı
I'm not exacly sure when and why the font cache is invalidated, but as a result this also invalidates the cached layouted glyphs. So just check the glyph layout cache state before trying to use it. But actually the status bar should update it's cache, so I'm not sure where to really put the fix. At least this won't crash LO anymore. Change-Id: I5f3ff8b9d56808af74d1419e878819d6019cd893 Reviewed-on: https://gerrit.libreoffice.org/64529 Tested-by: Jenkins Reviewed-by:
Jan-Marek Glogowski <glogow@fbihome.de> (cherry picked from commit 9e2e6c27) Reviewed-on: https://gerrit.libreoffice.org/64572
-
heiko tietze yazdı
Now 20 MRU plus... Change-Id: I069021bf1f42da793f7ec79a7e0eacbea519bc14 Reviewed-on: https://gerrit.libreoffice.org/64515 Tested-by: Jenkins Reviewed-by:
Heiko Tietze <tietze.heiko@gmail.com> (cherry picked from commit 271a767d) Reviewed-on: https://gerrit.libreoffice.org/64530
-
Noel Grandin yazdı
regression from commit 39e6bae2 loplugin:useuniqueptr in cui numpages Change-Id: I77e16880d0590793545da38b7a3c8c5d2626d8f5 Reviewed-on: https://gerrit.libreoffice.org/64441 Tested-by: Jenkins Reviewed-by:
-
- 04 Ara, 2018 15 kayıt (commit)
-
-
andreas kainz yazdı
Change-Id: I10f801de49cdf3d5fb37657ee88842223bf50570 Reviewed-on: https://gerrit.libreoffice.org/64385 Tested-by: Jenkins Reviewed-by:
andreas_kainz <kainz.a@gmail.com> (cherry picked from commit 5f6b5c98) Reviewed-on: https://gerrit.libreoffice.org/64512
-
Caolán McNamara yazdı
Change-Id: Ie6bccd2781fbbdc5f4d5dc2eb6903191aafe8265 Reviewed-on: https://gerrit.libreoffice.org/64520 Tested-by: Jenkins Reviewed-by:
-
Caolán McNamara yazdı
Change-Id: Iba217a2998de57bf963872f9b8c1cf38694eb935 Reviewed-on: https://gerrit.libreoffice.org/64507 Tested-by: Jenkins Reviewed-by:
-
Miklos Vajna yazdı
Regression from commit d6206295 (Use constants for ST_LineNumberRestart values, 2014-10-01) where I forgot to adapt the constructor's initializer list to also use the new constants. (cherry picked from commit dc8a8248) Change-Id: I7113d23cceb3a00f0d466b63c99a18bf3dfd0ee3 Reviewed-on: https://gerrit.libreoffice.org/64517 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
149a4d75 removed NeedParallelReduction() even though the code should be there. I'm not sure about GenSlidingWindowFunction(), since the base class is a template, but better revert that one too. Also revert that part of 73e6a797 that removed the subsequently no longer used functions. Change-Id: I932ffd58f9528ec840c4575078f8356640eb5420 Reviewed-on: https://gerrit.libreoffice.org/64249 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
Both of these are trivial and quite common mathematical operations. Unary minus is a bit complicated in that unary '-' is the same symbol as binary '-' and so there needs to be a hack to expand the '-' from the list to both of the opcodes ocSub and ocNegSub. Change-Id: I556e20e161c22dc89919afffbf904c1ef1d552ff Reviewed-on: https://gerrit.libreoffice.org/64362 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
Some error checking, but also simply use "if(cond) code;" rather than "if(!cond) try-to-set-benign-values-that-will-break-things-nevertheless". Change-Id: Ic0c296273ba4c174dba5e96eadeeee4005988142 Reviewed-on: https://gerrit.libreoffice.org/64342 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
The data vector may not have all values (if last cells are empty). In that case GetRefRowSize() is larger than GetArrayLength(). Other cases in this function had a similar check, this one was missing it for some reason (with the 'gid0' removed, since it's a fixed position). Change-Id: I01692b51d6ed22ee6db419508cacf556fb9e644d Reviewed-on: https://gerrit.libreoffice.org/64237 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
Test::testFormulaRefUpdateRange could trigger this, leading to recursion that wasn't handled properly by the code, since it wasn't expected to happen at late time (ScDependantsCalculator should have already caught it). This is all caused by the fact that FetchVectorRefArray() fetches also all rows before the given rows (to make the caching simpler I suppose). But that fetching could lead to Interpret() calls. Therefore, make ScDependantsCalculator in OpenCL mode check also all rows above. Change-Id: Iaecc105663df21b01443759287cec605470d34a5 Reviewed-on: https://gerrit.libreoffice.org/64236 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
These functions were removed in 149a4d75, although they quite clearly do serve a purpose, and then e9586cf0 removed the no longer used variables. Without this, ParallelReductionVectorRef class simply does not work. E.g. SUM(B1:B250) would not sum all the cells, but just one (testcase is Test::testFormulaRefUpdateRange with OpenCL forced). Change-Id: Iad1aeba2fbae2b4fc1d4236030adbb9ce6d972a1 Reviewed-on: https://gerrit.libreoffice.org/64235 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
PRODUCT() when there are no arguments (or all cells are empty) is 0. Change-Id: I4bcb9afe84d08833526255da0c61f6847d68ea36 Reviewed-on: https://gerrit.libreoffice.org/64232 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
Table operations do some tricks with reusing a cell with different values, and this confuses opencl handling, which operates on a group as a whole. Possibly relevant only with SC_FORCE_CALCULATION=opencl, sc_subsequent_filters_test fails in testDataTableOneVarXLSX(). Change-Id: I55e0d238061c8a45a2388ccf6c769ff6a1fddbcc Reviewed-on: https://gerrit.libreoffice.org/64230 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
No idea why somebody thought pow() would be better than a mere division. Without this the unittest fails because of a slightly different result. Change-Id: Ibbde8b872049be310ac98389634aab6eab154ca6 Reviewed-on: https://gerrit.libreoffice.org/64229 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
Otherwise the sqrtpi unittest fails in opencl mode. This value is less precise, but consistency should be more important here, not to mention that float doesn't have that high precision anyway. Change-Id: I8e85b7029d33932c81009e5d48e0fe2164c68dcf Reviewed-on: https://gerrit.libreoffice.org/64228 Tested-by: Jenkins Reviewed-by:
-
Luboš Luňák yazdı
That's how the scaddins implementation works, and the unittests fail in opencl mode without this. Change-Id: I08ea7dd0a0222408abf899d2da649be53f314a4c Reviewed-on: https://gerrit.libreoffice.org/64227 Tested-by: Jenkins Reviewed-by:
-
