mailpost and crossposted mailing lists
okada at opaopa.org
Tue Nov 13 07:15:40 UTC 2001
In the message <20011113002406.A1624 at us.ibm.com>
Scott Russell <lnxgeek at us.ibm.com> wrote:
> On Tue, Nov 13, 2001 at 10:01:04AM +0900, Katsuhiro Kondou wrote:
> > Or try mailpost in CURRENT. It's capable of handling multiple
> > newsgroups.
> I'm lookin at the perl code for mailpost from CURRENT-20011116 and I want to
> as about the -c option which is used for cross post support.
> It looks like if I pass -c N to mailpost that the mailpost process will
> sleep for N seconds then check it's database to see if it can find another
> dup. If a dup is found it's newsgroup is added to the Newsgroups: header and
> the message is sent off.
If a dup is found (former mailpost), its newsgroup is added to
the Newsgroups: header on the database, but the message is not sent off.
mailpost gateway of testlist at ibm.com -> newsgroup.lists.testlist
mailpost gateway of othertest at ibm.com -> newsgroup.lists.othertest
1. mailpost(A) receives testlist at ibm.com.
The mailpost(A) adds its newsgroups to the database, and sleeps.
Now the Newsgroups: header on the database is `newsgroup.lists.testlist'.
2. mailpost(B) receives othertest at ibm.com.
Then the Newsgroups: header on the database is `newsgroup.lists.testlist'.
The mailpost(B) adds its newsgroups to the database, and sleeps.
Now the Newsgroups: header on the database is
3. The mailpost(A) awakes and exits, because the Newsgroups: header was
changed on the database.
4. The mailpost(B) awakes. The Newsgroups: header was not changed
on the database, so the mailpost(B) posts the message actually.
> Did I miss something? In all this seems like a bad way to handle this
> because it appears to me that you could have two mailpost processes racing
> against each other. Please tell me if I'm wrong and show me?
mailto:okada at opaopa.org
More information about the inn-workers