a warning was printed.
A mounted procfs is a runtime dependency of valgrind, but since the
implementation of automatic suppression file generation the valgrind
binary is run in the post-install stage.
- An exception needs to be made for the ports cluster: The machine compiling
the list of IGNOREd ports does not have a mounted /proc, so the check (and
the generation of supression files is disabled there).
Users who want to suppress false positives in valgrinds output should build
the port instead of installing the package (so the rules can be customized to
their specific installation).
Reported by: Paul Ledbetter III <aeacides@gmail.com>
Approved by: portmgr (kris)
- Save arm-elf-gcc295 from being deprecated
PR: ports/84347
Submitted by: James E. Flemer <jflemer@alum.rpi.edu>
Approved by: Espen Skoglund <esk@ira.uka.de> (maintainer)
Approved by: portmgr (erwin, clement)
post to -ports on this topic
2. Fix build on amd64 by not trying to inline asm code designed for i386
3. Add a patch to fix window positioning on amd64, from
http://article.gmane.org/gmane.compw.window-managers.windowmaker.devel/749
2 and 3 provided/reminded by Dejan Lesjak <dejan.lesjak@ijs.si>
Fixes for 2 and 3 tested by Dejan, and A Dude <adudek16@gmail.com>.
Approved by: portmgr (krion)
on sparc64 and -fpic elsewhere. While here, make the following
improvements:
. ignore the vendor's fdlibm and use our own -lm. fdlibm is
derived from the same msun as ours, but spidermonkey was
misteriously linking with _both_. All mozilla-ports seem
to have the same problem right now;
. use our -lreadline instead of compiling vendor's own
libeditline;
. fix all warnings (clean build with -Wall -Werror);
. link the installed executable (js) against the shared
library libjs.so instead of against the invididual objects;
. unless WITHOUT_TEST is set, download and run vendor's own
tests in post-build (this triggers USE_PERL_BUILD). Some
tests had to be patched from Mozilla's CVS, because the
released tarball of them was not updated since 2002.
Bump PORTREVISION.
Approved by: portmgr (marcus)
Approved by: maintainer timeout
it started using the FT_GlyphSlot_Embolden function from freetype2
if it's available at configure type. libXft depends on freetype2
in ports tree so use this symbol. OOo has its own freetype2 and
in this version, it doesn't have FT_GlyphSlot_Embolden.
2. python fix for 6 and 7.
PR: 84464 [2]
Submitted by: "Stephane E. Potvin" <sepotvin@videotron.ca> [1]
"Carlos F. A. Paniago" <pan@panix.homeunix.org> [1]
Ron MacNeil[2], kris [1], [2]
Peter Czanik <pczanik@fang.fa.gau.hu> [2]
Approved by: portmgr (marcus)
- Also mark NO_PACKAGE because DEFAULT_PHP_VER=5 doesn't propagate to
dependencies, causing conflicts between php4 and php5 (like
pear-PHPUnit2 and a number of others).
PR: ports/84527
Submitted by: leeym
Approved by: portmgr (marcus), antonio@php.net (maintainer)
- Add another master site
Submitted by: Mikolaj Rydzewski <miki@ceti.pl>
Approved by: portmgr (marcus),
Krzysztof Kowalewski, Jr. <pyzmen@kam.pl> (maintainer)
* Redo SUBDIRS passing to work with new make(1)
* Patch out TIOCREMOTE ioctl calls, it was removed from FreeBSD
* Some gcc34 fixes
- Fix plist for xorg users
Approved by: portmgr (marcus)
