------------------------------
1.21 (July 2016)
Yann E. MORIN (yann.morin.1998@free.fr): Buildroot [1] is a build-system
targetting embedded devices. It is able to build toolchains for various
architectures, based on the traditional GNU compiler collection gcc, the GNU
binutils and various C libraries, such as GNU libc (glibc), uClibc [2] [3] or
musl [4].
glibc is a feature-full C library, which has had support for Yellow Pages
(and the rest of the SUN RPC stuff) for a long time. However, circa the 2.14
release, glibc stopped installing the RPC headers [5], on the principle that
providing the RPC implementation would be better served by a third-party
package, namely ti-rpc [6]. That did not happen in time, so installing the
RPC headers from glibc was re-instated circa 2.16.
uClibc (no longer maintained, replaced by a fork, uClibc-ng) is a fully
configurable C library for embedded systems. It entirely lacks a SUN RPC
implementation altogether.
musl is standards-conforming, lightweight C library. It also entirely lacks a
SUN RPC implementation.
TI-RPC is a stand-alone package. It may or may not be available (because the
user may not need SUN RPC stuff on an embedded device, for example). Besides,
it is not yet as feature-full as the implementation in glibc, though for all
practical means, it is totally useable.
So, there are cases where a SUN RPC implementation is not available.
This patch makes it so that the use of Yellow Pages is conditional on the
presence of a SUN RPC implementation, by checking at ./configure whether the
required header is present, and disabling the corresponding code if the
header is missing.
At the same time, the legacy, long unsupported AM_C_PROTOTYPES directive is
also removed, as it makes recent-ish autoconf fail.
MASTER_SITES= site1 \
site2
style continuation lines to be simple repeated
MASTER_SITES+= site1
MASTER_SITES+= site2
lines. As previewed on tech-pkg. With thanks to rillig for fixing pkglint
accordingly.
Existing SHA1 digests verified, all found to be the same on the
machine holding the existing distfiles (morden). Existing SHA1
digests retained for now as an audit trail.
this time from using gettimeofday() to using cloc_gettime(CLOCK_MONOTONIC),
to avoid the effects of UTC leap second insertion which IIUC steps
the system clock back one second, and will cause the single assert()
in dbeacon.cpp to fire. Unfortunately, my core dump was incomplete
due to a full root file system...
PKGREVISION bumped.
gettimeofday() directly. With a 32-bit clock_t and a 100Hz tick,
times() will wrap once every 497 days if my math isn't too far off,
and this will land us in an assert() in dbeacon and with a core dump.
So instead use gettimeofday() and save 2 x getrusage() per earlier
times() call. Retain some protection against gettimeofday() failing.
Bump PKGREVISION.
1.20 (October 2013)
Nick Birch (Nick.Birch@s-and-t.com): rtpsend -l (loop) flag does
not work. Modified to rewind input file on feof if loop was
set. Fails if length of line is greater than 2048
characters. Increased to 4096 (sufficient for typical
rtpdump: Put space after ext_data=<hex extension data> so that it
can be parsed by rtpsend. parse_header doesn't comprehend the
possible extension header. As a consequence any extension header
data is also dumped twice. Once in the ext_data report and then a
second time in the data report. Modify parse_header to return hlen
that comprehends the extension header if present.
The format for reporting contributing sources does not match the
format accepted by rtpsend (i.e., reports "csrc[n] = xxxx" rather
than "csrc[n]=0xxxxx"). Fixed. Note: the documentation is less
than clear on the csrc syntax accepted by rtpsend.
1.19 (August 2010)
Krzysztof Burghardt (krzysztof@burghardt.pl): "Wireshark
implemented "rtpdump" file format in a bit different way, so
cooperation beetwen those program is a bit problematic.
Moreover rtptools compiled on i386 and amd64 have different size
of file header (size of long in timeval is different)."
retry a limited number of times, and log the events. Export log() from
the main program for use in dbeacon_posix.cpp.
Fix the matrix.pl script since qw() no longer acts as parenteses, so add
some of our own where called for.
Bump PKGREVISION.
Do it for all packages that
* mention perl, or
* have a directory name starting with p5-*, or
* depend on a package starting with p5-
like last time, for 5.18, where this didn't lead to complaints.
Let me know if you have any this time.
to address issues with NetBSD-6(and earlier)'s fontconfig not being
new enough for pango.
While doing that, also bump freetype2 dependency to current pkgsrc
version.
Suggested by tron in PR 47882
a) refer 'perl' in their Makefile, or
b) have a directory name of p5-*, or
c) have any dependency on any p5-* package
Like last time, where this caused no complaints.
libXext/buildlink3.mk, now that it is included there.
Leave the places where its API version is set or variables from it
are used directly (about 3 packages).
