The PR claims the port is broken and this patch unbreaks it, but
portsmon is clean. In any case, the port builds cleanly in poudriere
with the patch so I defer to the maintainer.
PR: ports/187827
submitted by: maintainer (Yasuhiro KIMURA)
The maintainer actually provided a PR to address the breakage long
before it was marked broken. itetcu asked him a valid question: Why
is sqlbuddy defaulting to sqlite when it is primarily a mysql tool?
The question was never answered and the PR got thrown back on the heap.
The SQLite component is what is missing. Let's remove sqlite as an
option to restore the build and stage while here. It's bumped because
of the change in default options.
PR: ports/181708
submitted by: maintainer (Chenguang Li)
modified by: marino
MFH: 2014Q2
- Split dependency-owned check its own and make into a non-fatal warning.
This is not black-and-white. Python/Rubygem ports cleanup the lib dir
for every installed package, and for python/ruby-gems too. This is fine,
the extensions should not clean them up then; the error is valid here.
However, there's other frameworks which don't cleanup in the master package
and yield unexpected results. An example is that php does not cleanup
etc/php itself. Only extensions do. So if an extension depends on another
extension it will complain that the dependent extension owns etc/php,
so we remove from bsd.php.mk, now it's a leftover. The proper place is in the
php port itself, but it's unknown how many other cases there are like this.
So for now make this non-fatal and into a warning.
This is all mostly moot once pkg auto handles directories anyhow.
With hat: portmgr
Discussed with: antoine
I estimated it would take me 15 minutes to incorporate the PR, but boy
was I wrong. After several HOURS I finally got this port staged,
building on FreeBSD 10+, DragonFly too, and fix accomplished the
intent of the PR (don't clobber conf files, less noisy).
Staging revealed the original port was installing undocumented files.
Clang was having issues with the "inline" directives and I could only
get it to build on clang by removing them (this doesn't mean the
functions won't be inlined, nor does the directive mean that it will.)
Anyway -- a gift. I wouldn't have done this had I know how much work
it was going to be. I also should have overwritten the install
target instead of patch the vendor Makefile, that would have been a
lot easier and faster.
PR: ports/176676
Submitted by: Lawrence Chen
Approved by: maintainer (Raffaele DeLorenzo)
Other work by: marino
MFH: 2014Q2
- use OPTIONS_SUB instead PLIST_SUB
- remove back ported patch
- use POSIX instead SysV SHM
- remove LDFLAGS from CONFIGURE_ENV (unbreak auto. testing)
Changes for APR 1.5.1
*) apr_os_proc_mutex_get() on Unix: Avoid segfault for cross-
process pthread mutexes. [Yann Ylavic <ylavic.dev gmail.com>]
*) When using shmget-based shared memory, the ID used for ftok is
now an APR hash of the filename instead of the constant '1'.
We do this to help avoid collisions. PR 53996 [Jim Jagielski]
*) apr_socket_atreadeof(): Fix breakage on OS X. [Jim Jagielski]
*) Fix POSIX shared memory (shm_open) use for named shared memory.
Includes adding '--enable-posix-shm' to force POSIX shm if
available, and OS X compatibility. PR 55928.
[Jozef Hatala <jh-asf skrt org>, Jim Jagielski]
*) Fix race condition when calling apr_dir_make_recursive from
multiple threads on Windows.
[Bert Huijben]
*) Fix apr_escape.c compilation errors on EBCDIC platforms.
[Eric Covener]
*) FreeBSD 10: Correct a regression in 1.5.0 which affected non-
blocking sockets in some applications, including httpd. [Jeff
Trawick]
*) apr_skiplist: Add compatibility with C++ applications.
[Jeff Trawick]
*) Correct a regression in 1.5.0 which affected out-of-tree
builds on Unix. [Rainer Jung]
*) Improve platform detection for bundled expat by updating
config.guess and config.sub. [Rainer Jung]
with hat apache@
The main problem with building this port with clang was the use of
shared_ptr without the boost namespace which clashes with libc++. This
occurs in numerous places but luckly sed can fix most of them save for
a few files that inconsistently used boost names.
It was also missing an explicit link to libboost_system [1]
PR: ports/176172
Submitted by: Oliver Hartmann
Approved by: Maintainer timeout
Stage support: blanket approval
MFH: 2014Q2
- Convert to staging [1]
- Use new LIB_DEPENDS format [1]
- Remove BSD.local.dist dirs from plist
PR: ports/186885 [1]
Submitted by: Matthew Rezny <matthew@reztek.cz> [1]
