- Convert USE_EMACS to USES=emacs
- Remove editors/emacs-nox11 (refer to nox flavors of editors/emacs and
editors/emacs-devel)
- Permit default Emacs flavor to be specified in make.conf
- Rename japanese/migemo-emacs23 to japanese/migemo-emacs
- Update and simplify audio/emms and fix build on FreeBSD 10 [1]
- Update databases/bbdd and fix build on FreeBSD 10 [1]
- Update editors/emacs-devel
- Ensure Makefile shell commands that change directory are executed in a
subshell
- Silence some portlint warnings
[1] By not depending on base texinfo
PR: 225404
Reviewed by: antoine
Approved by: portmgr (mat) ashish (maintainer)
Differential Revision: https://reviews.freebsd.org/D13506
that were in files/patch-security.
Follow upstream and disable GNOMEVFS by default, using gio instead.
Don't enable crashdumps by default. This build knob causes the build
to be done with debug symbols, which greatly increase the size of the
build, but this is all in vain because the executables and libraries
are subsequently stripped.
Build with debug symbols if WITH_DEBUG is set.
Build with -fstack-protector.
Only fetch unowinreg.dll when building the SDK.
Remove tomcat from distinfo. It was leftover from an earlier
experiment.
Add three new options:
dd three new options:
CRASHDUMP - Enable crashdumps and set WITH_DEBUG.
DBGUTIL - Enable assertions, object counting, and other
non-production runtime debugging.
DEBUG - Compile with -O0 for better debugability. The ports
framework also sets WITH_DEBUG.
to be done with debug symbols, which greatly increase the size of the
build, but this is all in vain because the executables and libraries
are subsequently stripped.
Build with debug symbols if WITH_DEBUG is set.
Build with -fstack-protector.
Only fetch unowinreg.dll when building the SDK.
Add three new options:
CRASHDUMP - Enable crashdumps, requires WITH_DEBUG
DBGUTIL - Enable assertions, object counting, and other
non-production runtime debugging.
DEBUG - Compile with -O0 for better debugability. The ports
framework also sets WITH_DEBUG.
Fold a couple long lines in the Makefile.
In file included from externals/hunspell/affixmgr.cxx:12:
externals/hunspell/affentry.hxx:30:94: error: cannot initialize a parameter of type 'unsigned short' with an rvalue of type 'nullptr_t'
struct hentry * check_twosfx(const char * word, int len, char in_compound, const FLAG needflag = NULL);
^ ~~~~
externals/hunspell/affentry.hxx:30:94: note: passing argument to parameter 'needflag' here
externals/hunspell/affentry.hxx:93:103: error: cannot initialize a parameter of type 'unsigned short' with an rvalue of type 'nullptr_t'
struct hentry * check_twosfx(const char * word, int len, int optflags, PfxEntry* ppfx, const FLAG needflag = NULL);
^ ~~~~
Reported by: pkg-fallout
Obtained from: upstream (Hunspell 1.4.0)
and use the ccache code in the ports framework handle all of the
details.
Don't prepend ${WRKSRC}/solenv/bin to PATH in CONFIGURE_ENV since
it is no longer needed and blocks the framework from adding
CCACHE_WRAPPER_PATH to PATH in CONFIGURE_ENV and MAKE_ENV.
Sanitize MAKE_ENV and export it to the build, similar to what is
done by the default do-build target. This passes CCACHE_DIR and
HOME (set to WRKDIR) to the build, so the $HOME override in r459363
is no longer needed. We need to remove the CPPFLAGS setting from
MAKE_ENV before exporting to the build because it breaks the build
of the bundled icu module. CFLAGS and CXXFLAGS are also removed
because they would interfere with the carefully constructed compiler
flags set by the port. The include paths are already handled by
the port, and the port also modifies the optmization flags for some
files to work around compiler bugs.
While here, remove files/generate.pl since it has been obsolete for
a long while.
No PORTREVISION bump since the package should be unchanged.
PR: 224276
and use the ccache code in the ports framework handle all of the
details.
Don't prepend ${WRKSRC}/solenv/bin to PATH in CONFIGURE_ENV since
it is no longer needed and blocks the framework from adding
CCACHE_WRAPPER_PATH to PATH in CONFIGURE_ENV and MAKE_ENV.
Sanitize MAKE_ENV and export its to the build, similar to what is
done by the default do-build target. This passes CCACHE_DIR and
HOME (set to WRKDIR) to the build, so the $HOME override in r459316
is no longer needed. We need to remove the CPPFLAGS setting from
MAKE_ENV before exporting to the build because it breaks the build
of the bundled icu module. CFLAGS and CXXFLAGS are also removed
because they would interfere with the carefully constructed compiler
flags set by the port. The include paths are already handled by
the port, and the port also modifies the optmization flags for some
files to work around compiler bugs.
While here, remove files/ooo.sh files/generate.pl since they have
been obsolete for a long while.
No PORTREVISION bump since the package should be unchanged.
PR: 224276
-std=gnu++98 since that was the default for the compilers historically
used for the build. It turns out that -std=c++98 does not work
because the vigra headers depend on LLONG_MAX and friends, which
are not defined without the GNU extensions.
Compensate for a recent poudriere change by overriding $HOME and
setting it to /tmp if $HOME is not writeable. Some of the test
code run during the build wants to use mkstemp() to create temporary
directories under $HOME, which does not work for the "nobody" user
which points $HOME at /nonexistent.
No PORTREVISION bump since this is a build fix and there is no
change to the package.
PR: 225129
Reported by: jbeich
-std=gnu++98 since that was the default for the compilers historically
used for the build. It turns out that -std=c++98 does not work
because the vigra headers depend on LLONG_MAX and friends, which
are not defined without the GNU extensions.
Compensate for a recent poudriere change by overriding $HOME and
setting it to /tmp if $HOME is not writeable. Some of the test
code run during the build wants to use mkstemp() to create temporary
directories under $HOME, which does not work for the "nobody" user
which points $HOME at /nonexistent.
No PORTREVISION bump since this is a build fix and there is no
change to the package.
PR: 225129
Reported by: jbeich
actually any lighter---it's built with exactly the same feature-set as
the main vim port. The only difference is lack of gtk20 GUI, and the
language bindings are off by default.
An UPDATING entry is included.
By popular request, this is a slave port that installs only the vim binary. It has
no dependencies, produces a 1 MB package with a 3 MB installed footprint, and
is unable to do anything except edit files. It contains no help files, no runtime
files, no syntax highlighting, no filetype-specific indenting, non-US keymaps,
macros, or spell-checking.
vim-tiny is designed for minimal installs, and is the wrong choice for most users.
xap_Dialog.cpp:270:17: error: cast from pointer to smaller type 'char' loses information
*pWindowName = (char) NULL;
^~~~~~~~~~~
ap_UnixApp.cpp:866:29: error: invalid suffix on literal; C++11 requires a space between literal and identifier [-Wreserved-user-defined-literal]
if(0 == strcmp (suffix, "."G_MODULE_SUFFIX))
^
ie_exp_XSL-FO.cpp:1454:13: error: invalid suffix on literal; C++11 requires a space between literal and identifier [-Wreserved-user-defined-literal]
buf += " "x"=\""; \
^
ie_exp_LaTeX.cpp:1332:43: error: invalid suffix on literal; C++11 requires a space between literal and identifier [-Wreserved-user-defined-literal]
#define SUB(a,who) case a: subst = "\\(\\"who"\\)"; return true;
^
RealmProtocol.cpp:16:2: error: constant expression evaluates to -1 which cannot be narrowed to type 'uint32_t' (aka 'unsigned int') [-Wc++11-narrowing]
RPV1_PACKET_VARIABLE, /* user joined */
^~~~~~~~~~~~~~~~~~~~
RealmProtocol.cpp:10:30: note: expanded from macro 'RPV1_PACKET_VARIABLE'
#define RPV1_PACKET_VARIABLE -1
^~
RealmProtocol.cpp:16:2: note: insert an explicit cast to silence this issue
RPV1_PACKET_VARIABLE, /* user joined */
^~~~~~~~~~~~~~~~~~~~
static_cast<uint32_t>( )
RealmProtocol.cpp:10:30: note: expanded from macro 'RPV1_PACKET_VARIABLE'
#define RPV1_PACKET_VARIABLE -1
^~
soa_soup.cpp:166:11: error: no viable conversion from returned value of type 'soa::GenericPtr' (aka 'shared_ptr<soa::Generic>') to function return type 'bool'
return soa::GenericPtr();
^~~~~~~~~~~~~~~~~
Reported by: antoine (via bug 224669)
Obtained from: upstream (partially)
clearing `errno' prior to calling strtol(3).
Typically this is not needed (return value must be checked before the
`errno' check), but in case of strtol(3) return value can be zero for
both successful and failed conversions.
- Remove the use of comp_DETECT.
- Remove all comp_DETECT variables.
- Remove a use of comp_DETECT outside of gnome.mk
- Remove py3gobject3.
- Have ports depending directly on devel/py-gobject3 use gnome to do it.
PR: 224618
Submitted by: mat
Exp-run by: antoine
Sponsored by: Absolight
Differential Revision: https://reviews.freebsd.org/D13550
Usage:
USES=eigen:<version>[,<type>]
version: 2 or 3 (required)
type: build (default), run
For example:
USES=eigen:2,build,run
will add a BUILD- and RUN_DEPENDS on math/eigen2, and
USES=eigen:3
will add a BUILD_DEPENDS on math/eigen3.
* Convert the existing ports to use it
- biology/iqtree: remove run time dependency (seemed not to be needed)
- graphics/movit: remove run time dependency (seemed not to be needed)
- science/avogadro: add run time dependeny (installed cmake file requires it to be present)
Reviewed by: rakuco, mat
Differential Revision: https://reviews.freebsd.org/D13702
The following bugs are fixed:
BZ 127568 Diagrams become corrupt (not retrievable) when Calc document
is saved
BZ 127580 Fix for Issue 127568 created a new bug in Base
BZ 127581 Writer crashes after copying all content
