| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Redo some of the patches from the previous change set (PR pkg/56645) so
that they reflect the upstream project's method for including its
version of portable_utime(), rather than what was done previously,
which broke non-BSD builds, e.g., on Solaris derivates:
ld: fatal: symbol 'portable_utime' is multiply-defined:
(file .libs/osdep.o type=FUNC; file .libs/utime.o type=FUNC);
ld: fatal: file processing errors. No output written to .libs/libc-client.so.5.0.0
Compile tested on NetBSD 9.2_STABLE, NetBSD 9.99.93, OmniOS r151038,
and Fedora Linux 34.
|
|
|
|
by IMAP-UW and the
c-client library were incorrect. Unfortunately, on 32-bit systems, the the bug went unnoticed
because all the bits lined up correctly. On 64-bit systems, this happenstance doesn't work,
meaning the mtime and atime time stamps this software modifies are completely wrong, usually
defaulting to the beginning of the epoch, but only because the random memory locations pointed
to by the incorrect pointers are 0.
Fortunately, the software package includes a portable version of the utime library call that
converts the internal usage of this function to one that's compatible with the one resident in
libc on NetBSD, FreeBSD and other BSD systems. This patch simply enables that portable code
when compiling this package on BSD systems while using the pkgsrc system.
These patches have been compiled and are running in production on a 64-bit NetBSD/amd64
system running NetBSD-9.2.
Fixes pkg/pr-56645
|
|
|
|
OS_VERSION does not support arithmetic expressions.
|
|
All checksums have been double-checked against existing RMD160 and
SHA512 hashes
The following distfiles were unfetchable (possibly fetched
conditionally?):
./mail/qmail/distinfo netqmail-1.05-TAI-leapsecs.patch
|
|
its buildlink3.mk now includes openssl's buildlink3.mk
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Untested, except that it builds.
Bump PKGREVISION.
|
|
when EXTRAAUTHENTICATORS is passed as MAKE_FLAGS, it ends up being
doubled, mkauths then generates auths.c with doubled auth_gss.c and
auth_mit.c twice, triggering duplicate definition errors with clang
9.0.0; pass via MAKE_ENV instead
bump PKGREVISION
|
|
The actual fix as been done by "pkglint -F */*/buildlink3.mk", and was
reviewed manually.
There are some .include lines that still are indented with zero spaces
although the surrounding .if is indented. This is existing practice.
|
|
This hash_lookup() is internal function of imap but conflict with other
package (converters/php-recode) and cause php binary to crash.
Bump PKGREVISION.
|
|
|
|
|
|
Problems found locating distfiles:
Package mutt: missing distfile patch-1.5.24.rr.compressed.gz
Package p5-Email-Valid: missing distfile Email-Valid-1.198.tar.gz
Package pine: missing distfile fancy.patch.gz
Package postgrey: missing distfile targrey-0.31-postgrey-1.34.patch
Package qmail: missing distfile badrcptto.patch
Package qmail: missing distfile outgoingip.patch
Package qmail: missing distfile qmail-1.03-realrcptto-2006.12.10.patch
Package qmail: missing distfile qmail-smtpd-viruscan-1.3.patch
Package thunderbird24: missing distfile enigmail-1.7.2.tar.gz
Package thunderbird31: missing distfile enigmail-1.7.2.tar.gz
Otherwise, existing SHA1 digests verified and found to be the same on
the machine holding the existing distfiles (morden). All existing
SHA1 digests retained for now as an audit trail.
|
|
settings of INSTALL_UNSTRIPPED=yes for Darwin in individual packages.
|
|
only the latter is supported by cwrappers. Change them all to "opt" rules for
consistency and to gain compatibility with cwrappers.
|
|
Functionality is not tested.
|
|
|
|
|
|
|
|
|
|
Set LICENSE while here.
imap-2007f fixes a couple bugs.
Fix for RFC 4959 Initial Client Response auth failures noted first by
MacOSX Lion Mail users.
Adjust tcp_open.c:tcp_socket_open to make it a little more useful by adding
a write file descriptor test to the select in the case that the open
timeout is set.
In osdep/unix/env_unix.c:create_path there was a printf that should have
been an sprintf. Doesn't matter on modern systems.
|
|
pam.buildlink3.mk. This is wrong and a hack, but it gets the package
building again in my test environment. Someone please revert this and
do it right.
Currently there seems to be logic for enabling PAM in the build or not
on various platforms depending on whether native PAM appears to exist.
This is higly bogus; the package should be cleaned up so it either
depends on PAM or doesn't, regardless of platform, or preferably makes
it a build option.
XXX.
|
|
|
|
|
|
|
|
to use "cc" rather than an absolute path (which is no longer correct).
This fixes the build with that compiler without breaking the build
using GCC.
|
|
|
|
build at least on setups where "cc" is not present or not GCC which is
the default under Solaris 10.
|
|
|
|
|
|
|
|
Should be fix PR 41232, pick up incorrect location as pkgsrc expected.
|
|
including <utime.h> to make sure that compiler renaming was used to
select __utime50() instead of utime() as 64-bit time_t was being used.
|
|
This changes the buildlink3.mk files to use an include guard for the
recursive include. The use of BUILDLINK_DEPTH, BUILDLINK_DEPENDS,
BUILDLINK_PACKAGES and BUILDLINK_ORDER is handled by a single new
variable BUILDLINK_TREE. Each buildlink3.mk file adds a pair of
enter/exit marker, which can be used to reconstruct the tree and
to determine first level includes. Avoiding := for large variables
(BUILDLINK_ORDER) speeds up parse time as += has linear complexity.
The include guard reduces system time by avoiding reading files over and
over again. For complex packages this reduces both %user and %sys time to
half of the former time.
|
|
This version is a maintenance release, consisting primarily of bugfixes to
problems discovered in the release that affected a small number of users
plus a security fix for users of the RFC822BUFFER routines.
Approved by Thomas Klausner.
|
|
It is included from flocksim.h, required by some platforms. PR 40075.
Bump PKGREVISION.
While here, marked as MAKE_JOBS_SAFE=no
|
|
- The shared library major version because it is no longer compatible
with old binaries.
- The package revision because of the above change.
- The minimum API and ABI versions because new binary packages are no
longer compatible with an old "imap-uw" package.
|
|
- security fix for users of tmail or dmail.
- bug fixes and reliability improvements.
- A new function, utf8_csvalidmap(), has been added for the benefit of
Alpine to use in examining UTF-8 text and determining efficiently
whether it can be downgraded to a legacy charset. If you develop an
MUA, this may be useful for you too, although you'll have to read the
source code to see how to use it. The purpose of the "not-CJK" bit is
to prevent messages being downgraded to a CJK charset if all they have
in that charset are some special punctuation.
This update address the security vulnerability reported in SA32483.
|
|
version of pine in mail/pine. I've switched to mail/alpine as MUA, and
mail/dovecot for imap service, and so should you.
This package should be updated to the current version of imap-uw. Although
we are _not_ vulnerable to the current secunia advisory for imap-uw -- we do
not install tmail or dmail with the setuid bit set, nor advise users to
configure their mta to pass untrusted mailbox names to dmail on the command
line (who would do this?), there will eventually be a vulnerability which
does affect this old version.
However, as long as we support mail/pine, upgrading is not quite
straightforward -- the last version of mail/pine links against the version
of imap-uw currently included herein, and an upgrade would require testing.
|
|
Approved by Jim Wise.
|
|
that duse the c-client library do not have their callback function symbols
stripped at install time. Fixes the following:
PR pkg/34031
and indicates that this need not be reopened:
PR pkg/35592
also bump PKGREVISION and BUILDLINK_ABI_DEPENDS
|
|
PR pkg/37490
|
