Commit graph

1369 commits

Author SHA1 Message Date
Michael Stahl
bb35d3b241 WNT_INTEL_MSC.mk: use tabs for consistency 2012-03-16 15:49:41 +01:00
Luboš Luňák
cc9a101857 clang doesn't have a problem with -Wnon-virtual-dtor
Even if some older version possibly does, I doubt anybody would
realistically use it.
2012-03-15 11:48:58 +01:00
Lubos Lunak
370b3b6232 disable msvc warning 4265 (non-virtual dtor)
msvc2008 now prints out a lot of warnings, probably a result
of the recent -Wnon-virtual-dtor changes where msvc does not
understand the protected dtor actually makes the code to be ok
2012-03-15 11:48:28 +01:00
Stephan Bergmann
6e67c03dc0 Enable -Wnon-virtual-dtor for GCC 4.6
...which has the necessary features to support it.

Change a lot of classes to either contain a protected non-virtual dtor
(which is backwards compatible, so even works for cppumaker-generated
UNO headers) or a public virtual one.

cppuhelper/propertysetmixin.hxx still needs to disable the warning, as
the relevant class has a non-virtual dtor but friends, which would still
cause GCC to warn.

Includes a patch for libcmis, intended to be upstreamed.
2012-03-14 13:32:02 +01:00
Luboš Luňák
3d9f5c856a make OString work even with old gcc that has SFINAE broken
Since OString already has a ctor accepting const char*, I cannot
find out a way to distinguish string literals other than using
a template, otherwise const char* somehow takes precedence
(all of gcc, clang, msvc). But the template requires
the Substitution Is Not A Failure Idiom to actually create
only wanted instances. And the compiler can try evaluate
the OString ctor as a possibility when comparing an int to
an anonymous enum, and anonymous enum as a type without linkage
cannot be a template argument before C++11. SFINAE should still
work, but not with gcc older than 4.0.2 (which we right now
use only on macs). So for that case disable the string literal
ctors, which means macs will have one extra strlen call,
and also that embedded \0's in string literals will be
inconsistent. The tiny performance problem shouldn't matter that
much and will eventually go away, the \0 problem should not
matter, since before string literal ctors were introduced
\0's had not been included anyway unless RTL_CONSTASCII_STRINGPARAM
was used. So we should be safe and when removing the CONSTASCII
macros \0 cases should be handled by explicitly mentioning
the length.
2012-03-13 16:21:04 +01:00
Bjoern Michaelsen
3b6773c9d4 add bug-id directly to the error message 2012-03-13 14:53:03 +01:00
Bjoern Michaelsen
58f5a26e2d fdo#47246: abort on duplicate linked objects 2012-03-13 14:53:02 +01:00
Korrawit Pruegsanusak
380f0af0a8 gtar actually expects --strip-components, with trailing 's' 2012-03-13 18:38:37 +07:00
Stephan Bergmann
760e0d2d73 Document reason for -Wno-non-virtual-dtor 2012-03-13 10:39:52 +01:00
Matúš Kukan
794f4e2329 missing $: announce only if something is happening 2012-03-13 02:03:11 +01:00
Matúš Kukan
c289ee218f sdext: fix unit test for Windows 2012-03-10 21:23:23 +01:00
Matúš Kukan
6c1c06f1c2 gbuild: remove some more realpaths
Where necessary, replace with wildcard, what does not change path.
2012-03-10 21:23:21 +01:00
Matúš Kukan
84b3f7b345 gbuild: get rid of realpath in gb_Foo_set_include 2012-03-10 21:23:20 +01:00
Luboš Luňák
91fc0f37c1 clang does not know -ggdb2 2012-03-10 16:39:22 +01:00
David Tardon
ac488d581a quick hack to fix building sdext with l10n 2012-03-08 19:02:30 +01:00
Noel Power
c3d806be7d fix uno bootstrapping for .NET ( and perhaps c++ ) fdo#46832 2012-03-08 16:01:36 +00:00
François Tigeot
9ce48e3c88 Use the same CFLAGS on DragonFly and NetBSD. 2012-03-08 16:09:41 +01:00
François Tigeot
cf7e892932 Remove unused files.
They were identical to NETBSD_INTEL_GCC.mk anyway
2012-03-08 16:05:04 +01:00
David Tardon
74a6d7a135 fix mis-merge 2012-03-08 15:40:05 +01:00
Michael Stahl
294b86e3db gbuild: fix mkdir optimization with make 3.81
The problem is that make 3.81 does not match pattern rules whose target
ends with '/' against directories.
(regression from 408822b547,
90491a073c)
So use a fake file ".dir" in the respective directory, and try not to
spawn unnecessary mkdir processes by checking for existence via realpath.
This is all quite ugly and should be reverted once support for make 3.81
is dropped.

Also, fix a pre-existing problem of pattern rules with multiple targets
that used to work by accident in Package.mk.
2012-03-08 10:01:48 +01:00
David Tardon
35750371f7 fix adding files to extension's root dir 2012-03-06 10:31:30 +01:00
David Tardon
371efa17fa set localize.sdf path separately for each file
It cannot be done at the main target, because there is no guarantee that
all help (or properties) files are in one directory.
2012-03-06 10:31:30 +01:00
David Tardon
66a7815451 fail if any previous command failed 2012-03-06 10:29:05 +01:00
David Tardon
5ea8272801 hardcode manifest source path again 2012-03-06 10:29:03 +01:00
David Tardon
826c955136 add support for platform spec. data in manifest and description 2012-03-06 10:29:01 +01:00
David Tardon
2da1e35b5a simplify rule 2012-03-06 10:27:29 +01:00
David Tardon
154cc802eb add helper function to convert path to native format 2012-03-06 10:27:27 +01:00
David Tardon
6ed50e5159 allow adding deps on packages 2012-03-06 10:23:26 +01:00
David Tardon
81d7b8f821 allow using manifest in different location
This is necessary for C++ extensions (at least for those from sdext
module), because the manifest needs preprocessing (the extension's name
is platform-dependent etc.)
2012-03-06 10:23:26 +01:00
David Tardon
4abc59f063 support adding more than 1 file at once 2012-03-06 10:23:26 +01:00
Matúš Kukan
a5d28eaa69 UnoApiTarget: add missing dependency 2012-03-06 01:14:18 +01:00
Matúš Kukan
90491a073c gbuild: do not call mkdir -p in gb_Deliver__deliver
Rather create new order-only dependencies on directories where
targets should be delivered.

On cygwin this is much faster.
2012-03-05 23:02:41 +01:00
Matúš Kukan
408822b547 UnoApiTarget: avoid calling mkdir -p when not necessary 2012-03-05 23:02:40 +01:00
Tor Lillqvist
be5567d9e3 Revert "Workaround for weird gbuild looping problem with Clang for iOS"
The root cause for the problem was in solenv/bin/concat-deps.c, has
been fixed now.

This reverts commit af9618b29f.
2012-03-05 20:55:28 +02:00
Tor Lillqvist
af9618b29f Workaround for weird gbuild looping problem with Clang for iOS 2012-03-05 02:05:28 +02:00
Tor Lillqvist
fa65ff8b1d Check for Clang in configury and store and use result 2012-03-02 15:42:25 +02:00
Andras Timar
48666b7d60 Cygwin uses internal Python which does not understand Cygwin paths 2012-03-02 14:37:55 +01:00
Matúš Kukan
811180e9dd translations converted to gbuild 2012-03-02 11:22:23 +01:00
Tor Lillqvist
7a0b721462 Filter out options that Clang doesn't have 2012-03-01 01:26:20 +02:00
Tor Lillqvist
651dcc29f4 Add --disable-extensions switch for future implementation 2012-02-27 17:02:59 +02:00
Tor Lillqvist
c16bc82637 s/interpreters/scripting/ 2012-02-27 16:51:56 +02:00
Tor Lillqvist
4c737b6470 Decouple disabling of scripting (BASIC &co) from OS being iOS 2012-02-27 10:06:22 +02:00
Tor Lillqvist
9eb1c0cb7c Fix typos 2012-02-26 09:17:05 +02:00
Matúš Kukan
7e454a3e3d gbuild_simple: add gb_Python and include also Tempfile here 2012-02-25 13:10:43 +01:00
Matúš Kukan
66fdc38fb7 partial_build: make this work also for modules from clone/ 2012-02-25 13:10:42 +01:00
Matúš Kukan
b715e378aa gbuild: little more cleaning 2012-02-24 17:06:52 +01:00
Matúš Kukan
ea37d89410 gbuild: get rid of REPODIR 2012-02-24 17:06:51 +01:00
Matúš Kukan
4ee6123d05 Do not define PRUDUCT, we only use DBG_UTIL now 2012-02-24 17:06:50 +01:00
Luboš Luňák
1cf7ab61a7 use -std=gnu++0x rather than -std=c++0x
The gcc default for C++ is -std=gnu++98, and -std=c++98 is used
explicitly or with -ansi, so the C++0x "default" should be gnu++0x.
2012-02-24 14:06:19 +01:00
Tor Lillqvist
5f9657ba54 Filter out forms/util/frm component for a build without database connectivity 2012-02-23 22:59:58 +02:00