that states that the software is based in part of [sic] the
work of the FreeType Team, in the distribution documentation.
We also encourage you to put an URL to the FreeType web
page in your documentation, though this isn't mandatory.
-- <URL:http://cvs.freetype.org/cgi-bin/viewcvs.cgi/~checkout~/freetype2/docs/
FTL.txt?rev=1.7>
by programs such as mv (used in installation); remove librpm404 and
rpm404-python which are unused. Make rpm an EXTRACT_DEPENDS rather
than BUILD_DEPENDS, based on Phil Pennock's suggestion for the
sgifonts port.
Submitted by: Frank Reppin (adding libattr and libacl)
Installation of linux_base-7.1_2 fails when linprocfs is
mounted. Unmounting linprocfs before installing works
around the issue.
At Lars' suggestion, this script also mounts the linprocfs after
installation. I assume that if the user has a linprocfs entry in
/etc/fstab, the user wants it to be mounted. It also seemed like
a good idea to un-mount it before de-installation.
PR: 46172
The function gzprintf within zlib, when called with a string
longer than Z_PRINTF_BUFZISE (= 4096 bytes), can overflow
without giving a warning. zlib-1.1.4 and earlier exhibit
this behavior. There are no known exploits of the gzprintf
overrun, and only a few programs, including rpm2html and
gimp-print, are known to use the gzprintf function. The
problem has been fixed by checking the length of the output
string within gzprintf.
--<URL:http://rhn.redhat.com/errata/RHSA-2003-079.html>
package. libstdc++-2.96 is a core package that is referenced by many
linux_base applications such as linux-eagle. For instance, linux-eagle
doesn't work with linux_base-7.1_3.
Despite the importance, libstdc++-2.96 has been forgotten to be extracted
since 6 months. (Actually, the RPM package is listed in the Makefile,
but was badly defined so that no extraction was performed.)
Since linux_base is an important package, I hesitated to commit this
correction (for a long time :). If there is any incovenience in the
way of package definition, please feel free to correct my change.
remotely exploitable RPC bug (Red Hat advisory RHSA-2003-089 and
CERT advisory CA-2003-10).
According to <URL:http://www.redhat.com/apps/support/errata/>, "all
errata support for Alpha products, Sparc products, and Powertools
products have reached end of life."
I didn't do proper run-time testing of this, but "make package
deinstall" leaves a clean /compat/linux/ behind.
different MD5 hash from the corresponding package from 7.1/i386.
The contained etc/redhat-release files have the same contents but
different time stamps. By having both hashes in both distinfo
files, installation can proceed for users who share the same distfile
cache for both architectures.
list by bsd.port.mk insert anti foot-shooting device, which prevents
infinite fork loop when the user defines corresponding USE_XXX in global
make.conf, command line or environment.
Similar devices should probably be inserted into ports that might be inserted
into dependency list by others bsd.foo.mk files (bsd.ruby.mk, bsd.python.mk
and so on.)
warning about unknown owners. Add a reminder in the Makefile, because this
fix should be removed when a new release of the RPMS fixes the problem.
Specifically, the only affected file in the LINUXBASE hierarchy is
usr/man/man8/setserial.8 (-> root:wheel)
Approved by: marcel
settings and fails to install if the language is not C.
Override any locale settings when running rpm.
PR: 16621
Submitted by: Mikhail Teterin <mi@video-collage.com>
by the port had actual definitions with bogus values. Comment
out the definitions.
PR: 21554
Originator: Youki Kadobayashi <youki-k@is.aist-nara.ac.jp>
by the globally defined LINUXBASE. Setting DESTDIR on the
command line now does what one expects.
Also, don't download and install glib-1.2.5 when we're also
downloading and installing glib-1.2.6. It's now possible to
reinstall this port over itself.
independent packages that, despite the identical names, have
different contents.
This "feature" is currently implemented in a way that works
around a limitation in bsd.port.mk and can be considered a
hack.
versions are not the same as the i386 version, even though they
have the same name.
This fix does not have effect until the architecture specific md5
files are used.
