Age | Commit message (Collapse) | Author |
|
excluding the UDK headers of course
Change-Id: Iac7ab83d60265f7d362c860776f1de9d5e444ec0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/93268
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Using SVX_DLLPUBLIC for both Library_svxcore and Library_svx had started to
cause failures with clang-cl on Windows now, presumably due to devirtualization:
> linectrl.o : error LNK2001: unresolved external symbol "protected: virtual void __cdecl SvxMetricField::DataChanged(class DataChangedEvent const &)" (?DataChanged@SvxMetricField@@MEAAXAEBVDataChangedEvent@@@Z)
> linectrl.o : error LNK2001: unresolved external symbol "protected: virtual bool __cdecl SvxMetricField::PreNotify(class NotifyEvent &)" (?PreNotify@SvxMetricField@@MEAA_NAEAVNotifyEvent@@@Z)
> linectrl.o : error LNK2001: unresolved external symbol "protected: virtual bool __cdecl SvxMetricField::EventNotify(class NotifyEvent &)" (?EventNotify@SvxMetricField@@MEAA_NAEAVNotifyEvent@@@Z)
> linectrl.o : error LNK2001: unresolved external symbol "protected: virtual void __cdecl SvxMetricField::Modify(void)" (?Modify@SvxMetricField@@MEAAXXZ)
> linectrl.o : error LNK2001: unresolved external symbol "private: virtual bool __cdecl SvxFillAttrBox::PreNotify(class NotifyEvent &)" (?PreNotify@SvxFillAttrBox@@EEAA_NAEAVNotifyEvent@@@Z)
> linectrl.o : error LNK2001: unresolved external symbol "private: virtual bool __cdecl SvxFillAttrBox::EventNotify(class NotifyEvent &)" (?EventNotify@SvxFillAttrBox@@EEAA_NAEAVNotifyEvent@@@Z)
> C:\lo-clang\core\instdir\program\svxcorelo.dll : fatal error LNK1120: 6 unresolved externals
Replacing certain uses of SVX_DLLPUBLIC with the newly introduced
SVXCORE_DLLPUBLIC (include/svx/svxdllapi.h) has been done on Linux as follows:
> git grep -w --line-number -e SVX_DLLPUBLIC --and --not -e '#define SVX_DLLPUBLIC' >LINES
to produce a file LINES containing all 640 uses. (Conveniently, all uses
happen to be on different lines.) Manually create a file TOKENS
with 640 corresponding lines, each containing the (class or function) name that
is made SVX_DLLPUBLIC by in the corresponding line in LINES. Then
> nm -D --def instdir/program/libsvxcorelo.so | grep -ivw '[vw]' | c++filt >SVXCORESYMS
> nm -D --def instdir/program/libsvxlo.so | grep -ivw '[vw]' | c++filt >SVXSYMS
> n=$(cat TOKENS | wc -l)
> for ((i=1;i<="$n";++i)); do
> tok=$(head -n "$i" TOKENS | tail -1)
> printf @
> grep -Fw "$tok" SVXCORESYMS >/dev/null && printf svxcore
> printf @
> grep -Fw "$tok" SVXSYMS >/dev/null && printf svx
> printf '@ '
> head -n "$i" LINES | tail -1
> done
to generate 640 output lines detailing for each SVX_DLLPUBLIC name occurrene
whether it is mentioned in exports from neither (@@@), only from svx (@@svx@),
only from svxcore (@svxcore@@), or from both libraries (@svxcore@svx@). The
numbers that gives is
10 @@@
180 @@svx@
424 @svxcore@@
26 @svxcore@svx@
The 10 @@@ ask for follow-up clean up, but most of them are just left as
SVX_DLLPUBLIC for now. The exceptions are sxv::ITextProvider
(include/svx/itextprovider.hxx) and SdrCustomShapeGeometryItem::PropertyPairHash
(include/svx/sdasitm.hxx, where PropertyPairHash is a member struct of
SVXCORE_DLLPUBLIC SdrCustomShapeGeometryItem). Keeping them as SVX_DLLPUBLIC
would cause "unresolved externals" errors when linking Library_svxcore on
Windows.
The 180 @@svx@ are fine to keep as-is, and the 424 @svxcore@@ need rewriting.
The 26 @svxcore@svx@ needed manual inspection to decide (in some cases, the
chosen name in TOKENS was a too generic function name like Fill, in other cases
it was the name of a class exported from one library but also mentioned in the
arguments of a function exported from the other).
And for sdr::table::SdrTableObj the class itself is defined in svxcore while the
static member functions ExportAsRTF and ImportAsRTF are defined in svx. But
MSVC does not allow to mark the class as SVXCORE_DLLPUBLIC and the two static
member functions as SVX_DLLPLUBIC, so move the two functions out of the class.
(There appears to be no real necessity that they were static member functions in
the first place; they don't even need to be friends of the class. Nevertheless,
this mixture of functionality from svxcore and svx in include/svx/svdotable.hxx
may ask for follow-up clean up, one way or another.)
All the output lines that need rewriting (all the @svxcore@@ ones, and the
manually picked subset of @@@ and @svxcore@svx@ ones) are copied into a new file
CHANGE (containing 451 lines). Then
> sed -E -e 's|^@.*@.*@ ([^:]+):([0-9]+):.*$|sed -i -e "\2 s/SVX_DLLPUBLIC/SVXCORE_DLLPUBLIC/" \1|' <CHANGE >COMMANDS
> . COMMANDS
to do the changes.
Change-Id: If9b6dd1c9e9ba2eb883dbdac4385d28c6fc8a203
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87794
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
...instead of boost::optional<const OutputDevice&> (in preparation for replacing
boost::optional with std::optional, which doesn't support references types as
template arguments)
Change-Id: I02d2b617123cfa9ca0d6436c8a4104c5987f4eb5
Reviewed-on: https://gerrit.libreoffice.org/84121
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Also
- rename SetOfByte to SdrLayerIDSet
- add asserts in SdrLayerAdmin::GetUniqueLayerID so that we don't
allocate overlapping SdrLayerID values
- add a new constant SDRLAYERPOS_NOTFOUND to be returned from
SdrLayerAdmin::GetLayerPos
Change-Id: I3bb3489f9338e3d02c4040bcbd811744699941c8
Reviewed-on: https://gerrit.libreoffice.org/37467
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
The issue of 362d4f0cd4e50111edfae9d30c90602c37ed65a2 "Explicitly mark
overriding destructors as 'virtual'" appears to no longer be a problem with
MSVC 2013.
(The little change in the rewriting code of compilerplugins/clang/override.cxx
was necessary to prevent an endless loop when adding "override" to
OOO_DLLPUBLIC_CHARTTOOLS virtual ~CloseableLifeTimeManager();
in chart2/source/inc/LifeTime.hxx, getting stuck in the leading
OOO_DLLPUBLIC_CHARTTOOLS macro. Can't remember what that
isAtEndOfImmediateMacroExpansion thing was originally necessary for, anyway.)
Change-Id: I534c634504d7216b9bb632c2775c04eaf27e927e
|
|
Change-Id: I2ea407acd763ef2d7dae2d3b8f32525523ac8274
|
|
Bin pointless '// predeclarations' or '// forward declarations' comments. If
a programmer doesn't know what a forward declaration is, well, why read C++
source code?
Also bin some pointless '// end of namespace foo' comments after namespace
blocks just a handful of lines long, and some superfluous vertical whitespace.
Change-Id: I2a31c5d73d9b4de8825fb8eb9e7559dbd7303ef9
|
|
Change-Id: I911de1c7b47768233bdd9580361b8cf430b3684e
|
|
...mostly done with a rewriting Clang plugin, with just some manual tweaking
necessary to fix poor macro usage.
Change-Id: Ie656f9d653fc716f72ac175925272696d509038f
|
|
Change-Id: Idef06778251d7437cfce2151c6fb9654db04d3a4
|
|
Conflicts:
include/framework/preventduplicateinteraction.hxx
include/sfx2/sfxbasecontroller.hxx
include/sfx2/sfxbasemodel.hxx
include/toolkit/awt/vclxtabpagemodel.hxx
include/vcl/field.hxx
include/vcl/settings.hxx
Change-Id: Ibccf9f88c68267a3d7e656012b51eaf644c418c2
Reviewed-on: https://gerrit.libreoffice.org/8272
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: Ia5f104bfd707bcf4e159c78ca2764c861fb0b6d9
|
|
Change-Id: I2c280be12f36c1538e922286745aabc62482423d
|
|
see https://gerrit.libreoffice.org/#/c/3367/
and Change-Id: I00c96fa77d04b33a6f8c8cd3490dfcd9bdc9e84a for details
Change-Id: I199a75bc4042af20817265d5ef85b1134a96ff5a
|