...which broke all the stock /bin shells to no longer pass through any DYLD_*
environment variables, so the DYLD_LIBRARY_PATH passed into the firebird Make
would not be passed to all the places that need it (to find the external/icu
libraries that some executables link against which are built and run as part of
building firebird).
What works with two little tweaks is to build your own bash and pass it to make
with SHELL=...:
* For one, there is an empbuild executable that uses system(...) to call another
isql executable, where the latter needs DYLD_LIBRARY_PATH set, but which would
not pass through the system(...) call (which implicityl uses /bin/sh).
* For another, it is still necessary to invoke Firebird's make with an explicit
SHELL=$(SHELL), for reasons that are not entirely clear to me. (There are
some Makefile.in in Firebird's extern/ sub-tree that set "SHELL = @SHELL@" to
configure's CONFIG_SHELL, unless overriden via an explicit command line
arguemnt, but I don't think those are relevant here.)
Change-Id: I1e68faa898e758f09efb602d96fd6b35657e0480
...where firebird's replacements of global new/delete would still be called for
"new" but not for "delete". Lets hope that always adding the C++14 size_t
overloads of operator delete is harmless in all relevant pre-C++14 environments,
and only causes harmless warnings like -Wimplicit-exception-spec-mismatch in all
relevant C++14 environments.
(But why does a /library/ replace global new/delete in the first place?)
Change-Id: Ib0b0ad748c6641c07ffed6cec3d6809a1530679f
Reviewed-on: https://gerrit.libreoffice.org/19081
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
It would require NSS which I just decided to skip. If/when we patch it
to support native iOS (and OS X?) crypto APIs, re-consider. Or
something.
Now the TiledLibreOffice iOS demo app builds again.
Change-Id: I2816a49bbdbde5a8f9d3b9fcdfff420bdca05ff8
...which becomes evident with recent Clang trunk versions, and is fixed for
libjpeg-turbo master with <https://github.com/libjpeg-turbo/libjpeg-turbo/
commit/8b2c04f774d18e05e321ee67a9a38b4d7e84f168> "Fix x86-64 ABI conformance
issue in SIMD code" (but unfortunately needs a quite different patch for our
libjpeg-turbo 1.3.1 tarball).
Change-Id: Id9842cc85b4ea87b2faf73f1c8d10533621557e1
9a6cdce37e601b1406c71fef16ad9b315045c9da was trying to fix the problem
with exposing deprecated vars and functions in system's error_code.hpp
include file by patching bundled boost version. This approach would
only make sense, when upstream version is going to be fixed ASAP. Apply
another approach, and follow the same pattern as applied in external
libraries, by defining
-DBOOST_ERROR_CODE_HEADER_ONLY \
-DBOOST_SYSTEM_NO_DEPRECATED
instead of patching bundled boost version. This way, the code would
work with unpatched system boost 1.59 final as well.
Change-Id: I8684ca458ea4a5b7d7c3c3acfe7c14a6d19bc665
Reviewed-on: https://gerrit.libreoffice.org/18201
Reviewed-by: David Ostrovsky <david@ostrovsky.org>
Tested-by: David Ostrovsky <david@ostrovsky.org>
...which isn't even needed, as apr only uses APR_HAS_OS_UUID for apr_os_uuid_get
(apr_portable.h), which is neither used internally in apr nor by either of the
two clients of apr in LO, external/serf and ucb/source/ucp/webdav
Change-Id: I2e9d1f2640df0a8125ae2840f54488e77656c3ec
Fix a problem when neon send the token list for
methods modifying a WebDAV resource.
The problem showed up when working whith Sharepoint 2013.
Other WebDAV servers seem unaffected by it.
The If Request Header is currently formed in neon using the
"Tagged-List" format, while Sharepoint 2013 only accepts the
"No-Tag-List" format.
References:
section 7.5 of RFC4918:
http://tools.ietf.org/html/rfc4918#section-7.5
section 10.4 of RFC4918:
http://tools.ietf.org/html/rfc4918#section-10.4
Change-Id: I24d607fde251f1846f0e7b630b627b1500da18ea
Reviewed-on: https://gerrit.libreoffice.org/17423
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Jan Holesovsky <kendy@collabora.com>
It is only needed for Cygwin, and some patch(1) don't support it,
like NetBSD's.
Solution suggested by vmiklos on irc.
Change-Id: I65a3a82e755167828658218c7d73c4f2294e587f
use defines from static version of autoconfig.h for msvc instead of
fiddling with configure. Removes the need for adding some defines/checks
in the sources.
Change-Id: I21d4217ebb687eb13f7e7db2519a209f3b178a90