Hi,
it seems the following README files are not included in the conf/postfix-files:
BDAT_README
MAILLOG_README
POSTSCREEN_3_5_README
SMTPUTF8_README
Is it intended?
Checked with the postfix-3.6.4.
Downstream Fedora bugzilla:
https://bugzilla.redhat.com/show_bug.cgi?id=2041056
thanks & regards
On Thu, Jul 29, 2021 at 3:15 PM Wietse Venema wrote:
>
> Wietse Venema:
> > Wietse Venema:
> > > Jaroslav Skarvada:
> > > > Hi,
> > > >
> > > > glibc-2.34 introduced the closefrom function which causes postfix build
> > &g
Hi,
glibc-2.34 introduced the closefrom function which causes postfix build to fail:
In file included from ./vstream.h:22,
from attr_print64.c:100:
/usr/include/unistd.h:363:13: error: conflicting types for 'closefrom'; have 'vo
363 | extern void closefrom (int __lowfd) __THROW
Reproducer:
# useradd -c " " myuser # or more whitespaces
# su - myuser
$ echo | mail -s TEST root@localhost
# journalctl -u postfix | tail
Jun 30 13:31:13 vm-postfix8 postfix/pickup[4476]: 91309115983C:
uid=1000 from=
Jun 30 13:31:13 vm-postfix8 postfix/pickup[4476]: warning:
maildrop/D19752025C
Hi,
we run coverity & clang scans time to time. I filtered all the
false positives and the following are results which seemed
suspicious to me. It's mostly for unlikely code paths which
could be triggered by error state or invalid input. I am not
familiar with the code internals, so maybe it's unl
/usr/bin/ld:
anvil_clnt.o:/builddir/build/BUILD/postfix-3.4.9/src/global/./mail_params.h:3999:
multiple definition of `var_tlsp_clnt_policy';
abounce.o:/builddir/build/BUILD/postfix-3.4.9/src/global/./mail_params.h:3999:
first defined here
/usr/bin/ld:
anvil_clnt.o:/builddir/build/BUILD/postfi
- Original Message -
> Jaroslav Skarvada:
> > Hi Wietse,
> >
> > thanks for info. Are you going to introduce this patch upstream?
>
> The process is as follows:
>
> 1) I sadistically keep the fix for myself and let everyone else suffer.
>
&
- Original Message -
> Jaroslav Skarvada:
> > Hi,
> >
> > it seems the resolver was rewritten (probably in postfix 3.1.x), but
> > it seems it now uses res_query instead of the res_search which results
> > in RES_DEFNAMES and RES_DNSRCH not su
Hi,
it seems the resolver was rewritten (probably in postfix 3.1.x), but
it seems it now uses res_query instead of the res_search which results
in RES_DEFNAMES and RES_DNSRCH not supported with smtp_host_lookup=dns,
example:
# postconf -e "relayhost = [smtp]"
# postconf -e smtp_host_lookup=dns
# p
Hi,
I got to Coverity static analysis scan report for postfix. Most of the
errors were false positives, but two minor errors seems suspicious, that
it could cause leaks, so sharing them for review. Unfortunately, it's
result for postfix-3.2.5, but the same code is in the latest postfix -
just the
Hi,
I have noticed that in postfix-3.2.0 option -keep-new-mtime was renamed
to -keep-build-mtime, but it seems the old option is still mentioned in:
postfix-3.2.0/html/INSTALL.html
postfix-3.2.0/proto/INSTALL.html
e.g.:
$ grep -r --exclude=HISTORY keep-new-mtime postfix-3.2.0
postfix-3.2.0/proto/
Hi,
biff notification can leak descriptor (reproduced with mailman).
Attached one of possible fixes (also applies to latest postfix
sources)
regards
Jaroslav
diff -up postfix-2.3.3/src/local/biff_notify.c.orig2 postfix-2.3.3/src/local/biff_notify.c
--- postfix-2.3.3/src/local/biff_notify.c.orig2
Hi,
the postconf(5) documentation for reject_invalid_helo_hostname reads:
> Reject the request when the HELO or EHLO hostname syntax is invalid.
This seems ambiguous as it can be interpreted different ways:
1) 'HELO hostname' / 'EHLO hostname' syntax, i.e. HELO / EHLO syntax
2) hostname syntax fo
13 matches
Mail list logo