Age | Commit message (Collapse) | Author |
|
Change-Id: I294a8ff5070af20403ec103ad4420a6e12a97561
|
|
Change-Id: Idde31b2b07e27a1d9c91145516f8e78c638ce3b4
|
|
Change-Id: I82474c48924aca36d9fde468081fb2457a153ac6
|
|
Change-Id: I8ad33bd796953fb8c64b0220e53b54f7ac5be170
|
|
Change-Id: I8ba765a4d89618f301572c5fd9931f86d87af10d
|
|
Change-Id: I4825d94c58d8cfb245f110b7f46397c2724436a3
|
|
Change-Id: I06db8f7466f5baddc55ea9cb117bb8a222de0c01
|
|
Change-Id: Iee5554f1b51db0f6dc22ccbed743ae431bb895f6
|
|
Change-Id: Id6a444bd01873e30ea1522aaf3b951f4d5adc261
|
|
Change-Id: I99678f9b0595b8d5adfe939a1bfae4f4f2e8454e
|
|
Change-Id: I90e12d6e2999522ef0d252e1592b8649e8e4560c
|
|
To make this experience slightly less confusing.
Change-Id: I844dd2cdbc0349433d445955cdcf5e9723673286
|
|
Change-Id: Ie6443fd9b1ac6332b4937c4b6b5d565cca1fe612
|
|
Change-Id: I83abe55a758a6900942de4ba337f9fb5444cd866
|
|
Change-Id: I9ccd976bc384e6bee078caf693db737f9b0ef0e1
|
|
Change-Id: I667f93be72bd5d6cc68725b7cb5d97e30d8df521
|
|
Change-Id: I73acb3a150b01114d32274a6842f6db9654b3e63
|
|
Convert code like:
rTxt.match( "---", rTxt.getLength()-3 )
to:
rTxt.endsWith( "---" )
Change-Id: Iada74c5e714f7234f25b326526843a36255d5599
|
|
And this already has revealed one bug in the xlsx import filter code!
Change-Id: Ib845a5492ec7b62b8528b31eac8cd1b285500184
|
|
We need these for Excel interop.
Change-Id: I91450c1d205f28636edfb4392aa6ae5091b1d7b7
|
|
Change-Id: I43d699b1e89ec327ec27566a98160a424216c78c
|
|
... so if it says leftmost bit 0 then test for high bit.
Change-Id: I260ff09fff66918da887742b5d177715e59548c2
|
|
Change-Id: Id38a7629ea5ed4bbb1a7d696926335ce0bdec6a6
|
|
Change-Id: I9fe0558c013641a902ce5a95c02d37be9729fe38
|
|
Change-Id: I375a72c34e46778385a9fdc24c9ebd0f0a3f3c9b
|
|
Change-Id: I87c5a180566e9da185c2992844e6522e82c17747
|
|
Change-Id: I733cd63e321bdc775739228c269fa66ff7780780
|
|
1) Use ods format and store the expected formula results as numeric cells,
to avoid having to load the same file twice (which is slower).
2) Use rtl::math::approxEqual to evaluate two numbers that may be subject
to numerical instability especially when the expected values are very
small.
Change-Id: Ie08e91f797065f4593f996df3541da84019cc53c
|
|
Change-Id: I0a5fbfc0b33b16e0eba7410957ab83ffbff4e008
|
|
I could not reproduce the crash of that bug (probably having more memory
available), but the backtrace had nTabSpan = -2 implicitly casted to
size_t leading to allocation of a huge amount of memory with
vector::reserve(), which ScRange::Justify() exactly prevents.
Change-Id: Idb79e1be62649922ba793cab01e00011479fade9
|
|
Change-Id: I85e8823e872bf53182ce90576225e462edb1930a
|
|
Change-Id: Ib19ecab287396dc6d6cf4a77dcf7072c5c4c84f8
|
|
Change-Id: Ic624b6bab3215148ff0b9a73fd24577623b398f9
|
|
Change-Id: I9a18699554bb0831478e5ca177e87ae11a6206cf
|
|
Change-Id: I67345e7eaae16baf677e9aad721d39b4ea0c43d1
|
|
Otherwise listeners may end up listening to larger areas than their original
listening areas after the relocation.
Change-Id: I12a7f49aec1c78611db5c31ba2d18574e91a5416
|
|
It's safe this way in case we are transferring them within the same
column. Any empty broadcasters are checked and purged at the end.
Change-Id: Ib2d46e616cde4923720ad21cb101d3a97dc8c5d9
|
|
Change-Id: I4e4b788f34e01f926d60d97515de4191d7346d5e
|
|
Change-Id: I8b60969312fd262ee8cd996c861e4d69f9ea3372
|
|
Also with this commit, the signature of SvtListener::Notify() changes,
by dropping the first argument which nobody uses. This change was necessary
in order to call it directly without needing to pass any broadcaster instance.
Change-Id: I6a1e97f0fb1e070d1d8f7db614690b04c9e8024e
|
|
Change-Id: I20de3d5c3c254b2ce96678553d91492dc95e2c39
|
|
Ideally they'd even be stored as document options. Future work?
Change-Id: Ida126a441ca88c35406fb66ea3722d6335e6e8e2
|
|
http://retrace.fedoraproject.org/faf/reports/338291/
this crash refuses to be captured alive. But seeing as
we sometimes check mpViewShell against NULL and sometimes check
GetViewData against NULL its worth making those checks everytime
for consistency
see also...
rhbz#923993 rhbz#896711 rhbz#602273 rhbz#577595 lp#811057 lp#760665
Change-Id: I5553848018f500b63ade60bcfc756a736b27c64f
|
|
Change-Id: I6a6d0ba558d0aae075e2a00f76ce96990af882d3
|
|
Change-Id: I9da62b026b40b08784415e0421556946646d25e2
|
|
Speeds up sorting large ranges containing formula groups by a
computational order or two: can sort 1 million rows of =A1 in sub 10s
Change-Id: I729c801e68154986956b1726e1c35ef973cc4248
|
|
Change-Id: Ie76ee86f6ae48012c92a3f06ad1372f126f1ea29
|
|
Change-Id: I0014cc8af22e4f2284d8e72b6b79fa300ad89028
|
|
Change-Id: If7d9c032d6025a6c6bf850e338923296ba2590dd
|
|
Determines how to treat text when encountered as operand in an arithmetic
operation or as argument to a function that expects a number instead.
Selectable under Tools->Options->Calc->Formula "Detailed calculation settings"
"Custom" from "Conversion from text to number" are:
Generate #VALUE! error: =1+"1" or =1+"x" give #VALUE!
Treat as zero: =1+"1" or =1+"x" give 1
Convert only unambiguous: =1+"1" gives 2, but =1+"1.000" or =1+"x" give #VALUE!
Convert also locale dependent: =1+"1.000" may be 2 or 1001 ... =1+"x" gives #VALUE!
For "Generate #VALUE! error" and "Treat as zero" the "Treat empty string as
zero" option follows these settings, for "Convert only unambiguous" and
"Convert also locale dependent" it can be set independently.
When reading documents created by other spreadsheet applications or older
versions of LibreOffice, and to interchange documents between different locales
the "Convert only unambiguous" with "Treat empty string as zero = True" setting
is recommended, though LibreOffice so far acted as "Convert also locale
dependent" with "Treat empty string as zero = False", which is the reason that
option is kept as default.
The best setting to create new documents that can be interpreted by all
spreadsheet applications without on-the-fly string conversion is
"Generate #VALUE! error". Not having to convert strings during
calculation ist also faster, of course.
Change-Id: Ie6dc34a00a82064a2d862b2178ce715fab945f85
|