Age | Commit message (Collapse) | Author | Files | Lines |
|
* In lcl_ScAddress_Parse_OOo() don't assume non-existing sheets would be
external documents if there was no sheet separator. This lead to an
external reference being created that is only an error if
dereferenced, not if fed to ISERROR() and the like.
* In ScInterpreter::ScIndirect() push errNoRef instead of
errInvalidArgument to produce a #REF! error.
Signed-off-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
During XML import formulas of defined names and conditional formatting
were remembered as regular svString tokens that then later were
retrieved without quotes for compilation. This didn't go along with the
new ScRangeData::CompileUnresolvedXML() that recreates the formula
string from an already tokenized form of the formula.
Introduced FormulaToken::AddStringXML() with ocStringXML to sort those
out and removed the IsImportingXML() hack from
FormulaCompiler::AppendString(), the ocStringXML case is handled in
FormulaCompiler::CreateStringFromToken().
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
Dependencies of defined names must not depend on the order in which they
are inserted during file load. In a second step compile defined names
that had unresolved names during load, and only those.
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
backported 34315e7ec4062f9521cd19951b5f7f6ad9ce0d2e
Signed-off-by: Michael Stahl <mstahl@redhat.com>
|
|
Same problem in ScHelperFunctions::GetBorderLine.
Signed-off-by: Eike Rathke <erack@redhat.com>
|
|
Only write "style:vertical-justify" and "css3t:text-justify" in ODF
extended mode.
(regression from 7da57d17b6179e71c8b6d7549ad89eaf3a4a28c6)
Signed-off-by: Noel Power <noel.power@novell.com>
|
|
|
|
Don't access a css::table::BorderLine struct as if it was
a css::table::BorderLine2 struct, which it isn't. A proper implementation
would need to pass BorderLine2 structs around for which it would need
a css::table::TableBorder2 struct that holds BorderLine2 instead of
BorderLine, and adapt various places for API compatibility to support both.
For now use the default table::BorderLineStyle::SOLID to set the line style at
::editeng::SvxBorderLine
This also fixes fdo#39117.
(cherry picked from commit 97c66b3b13528cc50f271217663718b3d6681bd9)
Signed-off-by: Michael Stahl <mstahl@redhat.com>
|
|
Replaces commit e2b11f4fd79dce4116badb0ecf6477546ca5d0d4
Prevent excessive references to single token instance during fill.
Because if not, we may run out of the 16-bit integer space to
store reference count.
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
* Interpreter tried to access a nonexistent matrix for external reference.
* In ScInterpreter::PopExternalDoubleRef(ScMatrixRef& rMat) set error for
unexpected conditions. In caller or descendants check for error or presence
of ScMatrix.
|
|
Fixed a misconception in transition from List to to ptr_vector of commit
db599f64e63a3141159b7f0e59783197c39b3e02, List::Next() returned NULL when at
end of list, mimic that.
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
+ Checks out-of-bounds accesses in
ScFormatRangeStyles::GetStyleNameIndex() and
ScRowFormatRanges::AddRange() and prevents crashes.
- The real cause seems to be some style row/repeat miscalculation
elsewhere, further investigation would be necessary.
Signed-off-by: Norbert Thiebaud <nthiebaud@gmail.com>
|
|
The breakage was due to the drawing layer re-work. We do need to store
charts (or all OLE shapes with associated listening ranges) for later
registration regardless of anchor types.
Signed-off-by: Eike Rathke <erack@redhat.com>
|
|
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
ScMatrixImpl lacked initialization of pErrorInterpreter=NULL.
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
Another unfortunate logic change introduced via DECLARE_LIST removal.
Reverting the logic to the original one.
Signed-off-by: Eike Rathke <erack@redhat.com>
|
|
If you do, things go very very wrong. These *limit_cast try to be
clever and check ranges of various types before casting, but they
fail to do the right thing when casting a signed integer into unsigned.
Signed-off-by: Lionel Elie Mamane <lionel@mamane.lu>
|
|
This one is pretty bad, especially for those who like to select
ranges using keyboard.
Signed-off-by: Eike Rathke <erack@redhat.com>
|
|
We don't need to auto-detect the formats of caption values; they are
already formatted by the time they arrive in to ScDPOutput.
Signed-off-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
This is a simple silly mistake; the matrix representation of the
external range should've preserved the original range; not the data
range which can be smaller than the originally requested range.
Signed-off-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
|
|
Signed-off-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
ScFormulaCell::GetMatrixEdge() when recalculating its dimension obtained
the matrix origin of itself instead of a neighboring matrix formula if
encountered such, effectively adding that other matrix to its own cells
occupying dimension.
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
This ensures that the bottom text "Drag the fields from ..." is
visible on all platforms. Previously this text overlapped with
the data and row field windows on Windows XP.
Signed-off-by: Tor Lillqvist <tlillqvist@suse.com>
|
|
Signed-off-by: Eike Rathke <erack@erack.de>
|
|
Signed-off-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
ScRangeName::maIndexToData
* The newly introduced ScRangeName::maIndexToData needs to take into
account that boost::ptr_set clones and deletes elements, hence each
collection needs its own maIndexToData setup in copy-ctor.
* Also fixed ScRangeName::erase() where a wrong index element was set to NULL.
Signed-off-by: Eike Rathke <erack@erack.de>
|
|
This should speed up formula calculations considerably during xls
import since shared formulas are also stored in ScRangeName and
they are looked up by index. (bnc#715104)
Signed-off-by: Eike Rathke <erack@erack.de>
|
|
Signed-off-by: Caolán McNamara <caolanm@redhat.com>
|
|
Old intrusive pointer in operator->() simply returned a NULL pointer if
so, boost::intrusive_ptr asserts on that. If explicitly used to obtain
the pointer use get() instead.
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
better performance on loadin calc files (esp. big ones)
|
|
|
|
Modified from the original patch submitted by Olivier Hallot.
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
This was "fixed" in i#107275, but the fix only worked in locales
where the range/argument separator was ';' (e.g. German locale).
This is the correct fix that works in all locales regardless of
the separator.
Signed-off-by: Noel Power <noel.power@novell.com>
|
|
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
If we don't invalidate the table stream when we change the local range
name, the new range name will only be written if something else has been
changed in the sheet
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
Signed-off-by: Eike Rathke <ooo@erack.de>
|
|
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
Signed-off-by: Kohei Yoshida <kohei.yoshida@suse.com>
|
|
Signed-off-by: Kohei Yoshida <kyoshida@novell.com>
|
|
|
|
Signed-off-by: Kohei Yoshida <kyoshida@novell.com>
|
|
Signed-off-by: Kohei Yoshida <kyoshida@novell.com>
|
|
Signed-off-by: Kohei Yoshida <kyoshida@novell.com>
|
|
And also fdo#37499.
It seems that this commit somehow disappeared during merge.
Signed-off-by: Noel Power <noel.power@novell.com>
|
|
Signed-off-by: Michael Meeks <michael.meeks@novell.com>
|