FQDN for Leafnode - Mandriva

This is a discussion on FQDN for Leafnode - Mandriva ; How do I get rid of the must have FQDN problem. /etc/cron.daily/texpire causes # cat /var/log/news.all Jan 1 23:15:24 wb texpire[994]: config: debugmode is 3 Jan 1 23:15:24 wb texpire[994]: config: found username for news.verizon.net Jan 1 23:15:24 wb texpire[994]: ...

+ Reply to Thread
Page 1 of 3 1 2 3 LastLast
Results 1 to 20 of 47

Thread: FQDN for Leafnode

  1. FQDN for Leafnode

    How do I get rid of the must have FQDN problem.
    /etc/cron.daily/texpire causes

    # cat /var/log/news.all
    Jan 1 23:15:24 wb texpire[994]: config: debugmode is 3
    Jan 1 23:15:24 wb texpire[994]: config: found username for news.verizon.net
    Jan 1 23:15:24 wb texpire[994]: config: found password for news.verizon.net
    Jan 1 23:15:24 wb texpire[994]: config: maxage is 2
    Jan 1 23:15:24 wb texpire[994]: config: filterfile is /etc/leafnode/filters
    Jan 1 23:15:24 wb texpire[994]: Core file size: 67108864
    Leafnode must have a fully-qualified and globally unique domain name,
    not just "wb.home.invalid".
    Edit your /etc/hosts file to add a unique, fully qualified domain name.
    "localhost.localdomain" or thereabouts will not work;
    it's qualified, but not unique.
    Please see the README-FQDN file for details.

    Been there, done that, plus google.
    I have got to be missing something simple.

    Setting hostname = wb.home.invalid in /etc/leafnode/config makes no difference.



    $ cat /etc/release
    Mandriva Linux release 2008.0 (Official) for i586

    leafnode-1.11.6-1mdv2008.0.i586.rpm 345814 Oct 05 00:01

    $ head -5 /etc/hosts
    127.0.0.1 localhost.localdomain localhost
    192.168.1.11 fw.home.invalid fw
    192.168.1.12 wb1.home.invalid wb1
    192.168.1.130 wb.home.invalid wb
    192.168.1.131 beta.home.invalid beta


    $ grep -i wb.h /etc/hosts
    192.168.1.130 wb.home.invalid wb

    $ grep 130 /etc/hosts
    192.168.1.130 wb.home.invalid wb


    FYI: /etc/hosts -> ../var/lib/named/etc/hosts


    $ hostname
    wb.home.invalid

    $ hostname --ip-address
    192.168.1.130

    $ host wb
    wb.home.invalid has address 192.168.1.130

    $ host wb.home.invalid
    wb.home.invalid has address 192.168.1.130

    $ nslookup $(hostname)
    Server: 192.168.1.130
    Address: 192.168.1.130#53

    Name: wb.home.invalid
    Address: 192.168.1.130


    $ cat /etc/host.conf
    order hosts,bind
    multi on
    nospoof on
    spoofalert on

    $ grep hosts: /etc/nsswitch.conf
    hosts: files dns nis

    $ ifconfig eth0 | head -3
    eth0 Link encap:Ethernet HWaddr 00:A0:CC:21:B8:91
    inet addr:192.168.1.130 Bcast:192.168.1.255 Mask:255.255.255.0
    inet6 addr: fe80::2a0:ccff:fe21:b891/64 Scope:Link

  2. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 01:01:05 -0500, Bit Twister wrote:

    > How do I get rid of the must have FQDN problem.


    Have you set the HOSTNAME= in /etc/sysconfig/network ?

    Regards, Dave Hodgins

    --
    Change nomail.afraid.org to ody.ca to reply by email.
    (nomail.afraid.org has been set up specifically for
    use in usenet. Feel free to use it yourself.)

  3. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 01:08:20 -0500, David W. Hodgins wrote:
    > On Wed, 02 Jan 2008 01:01:05 -0500, Bit Twister wrote:
    >
    >> How do I get rid of the must have FQDN problem.

    >
    > Have you set the HOSTNAME= in


    I would be impressed if leafnode's tools look in there.

    $ grep -i home /etc/sysconfig/network
    HOSTNAME=wb.home.invalid

    And Oh, by the way

    $ cat /etc/resolv.conf
    # Dynamic resolv.conf(5) file for glibc resolver(3) generated by resolvconf(8)
    # DO NOT EDIT THIS FILE BY HAND -- YOUR CHANGES WILL BE OVERWRITTEN
    nameserver 192.168.1.130
    domain home.invalid


    $ dig $(hostname)

    ; <<>> DiG 9.4.1-P1 <<>> wb.home.invalid
    ;; global options: printcmd
    ;; Got answer:
    ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 38404
    ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 2, ADDITIONAL: 1

    ;; QUESTION SECTION:
    ;wb.home.invalid. IN A

    ;; ANSWER SECTION:
    wb.home.invalid. 86400 IN A 192.168.1.130

    ;; AUTHORITY SECTION:
    home.invalid. 86400 IN NS ns1.home.invalid.
    home.invalid. 86400 IN NS ns2.home.invalid.

    ;; ADDITIONAL SECTION:
    ns1.home.invalid. 86400 IN A 192.168.1.130

    ;; Query time: 3 msec
    ;; SERVER: 192.168.1.130#53(192.168.1.130)
    ;; WHEN: Wed Jan 2 01:54:27 2008
    ;; MSG SIZE rcvd: 101

  4. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 02:56:40 -0500, Bit Twister wrote:

    >>> How do I get rid of the must have FQDN problem.


    Have you confirmed that "host 192.168.1.130" is returning wb.home.invalid,
    and that the named service is starting prior to leafnode (xinetd service)?

    Regards, Dave Hodgins

    --
    Change nomail.afraid.org to ody.ca to reply by email.
    (nomail.afraid.org has been set up specifically for
    use in usenet. Feel free to use it yourself.)

  5. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 10:03:32 -0500, David W. Hodgins wrote:

    > Have you confirmed that "host 192.168.1.130" is returning wb.home.invalid,


    $ host 192.168.1.130
    130.1.168.192.in-addr.arpa domain name pointer wb.home.invalid.

    > and that the named service is starting prior to leafnode (xinetd service)?


    named is started on boot and leafnode is started on demand.

    # chkconfig --list named
    named 0ff 1ff 2ff 3n 4n 5n 6ff

    xinetd based services:
    leafnode: on


  6. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 11:34:21 -0500, Bit Twister wrote:

    > named is started on boot and leafnode is started on demand.


    leafnode is executed on demand, but the config file is only loaded
    when xinetd starts, or you run service leafnode start.

    Is the error message being generated when you run service leafnode start,
    or only when xinetd is initially loading?

    Regards, Dave Hodgins


    --
    Change nomail.afraid.org to ody.ca to reply by email.
    (nomail.afraid.org has been set up specifically for
    use in usenet. Feel free to use it yourself.)

  7. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 15:27:39 -0500, David W. Hodgins wrote:
    > On Wed, 02 Jan 2008 11:34:21 -0500,
    > Bit Twister wrote:
    >
    >> named is started on boot and leafnode is started on demand.

    >
    > leafnode is executed on demand, but the config file is only loaded
    > when xinetd starts, or you run service leafnode start.
    >
    > Is the error message being generated when you run service leafnode start,
    > or only when xinetd is initially loading?


    Message is from /etc/cron.daily/texpire, See:

    Leafnode must have a fully-qualified and globally unique domain name,
    not just "wb.home.invalid".
    Edit your /etc/hosts file to add a unique, fully qualified domain name.
    "localhost.localdomain" or thereabouts will not work;
    it's qualified, but not unique.
    Please see the README-FQDN file for details.

    ----- and here is results of cat /etc/cron.daily/texpire ------------
    #!/bin/sh
    #
    # Leafnode texpire cron job. To be run daily.

    # Expire old news.
    su - news -c /usr/sbin/texpire

    # Get new news.
    #su - news -c /usr/sbin/fetch
    -------------------- end of cat /etc/cron.daily/texpire ----------------------

    # file /usr/sbin/texpire
    /usr/sbin/texpire: ELF 32-bit LSB executable, Intel 80386, version 1
    (SYSV), for GNU/Linux 2.6.9, dynamically linked (uses shared libs), stripped


  8. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 06:01:05 +0000, Bit Twister wrote:

    > How do I get rid of the must have FQDN problem.
    > /etc/cron.daily/texpire causes
    >
    > # cat /var/log/news.all
    > Jan 1 23:15:24 wb texpire[994]: config: debugmode is 3
    > Jan 1 23:15:24 wb texpire[994]: config: found username for news.verizon.net
    > Jan 1 23:15:24 wb texpire[994]: config: found password for news.verizon.net
    > Jan 1 23:15:24 wb texpire[994]: config: maxage is 2
    > Jan 1 23:15:24 wb texpire[994]: config: filterfile is /etc/leafnode/filters
    > Jan 1 23:15:24 wb texpire[994]: Core file size: 67108864
    > Leafnode must have a fully-qualified and globally unique domain name,
    > not just "wb.home.invalid".
    > Edit your /etc/hosts file to add a unique, fully qualified domain name.
    > "localhost.localdomain" or thereabouts will not work;
    > it's qualified, but not unique.
    > Please see the README-FQDN file for details.
    >
    > Been there, done that, plus google.
    > I have got to be missing something simple.

    -snip-
    Hi Bit,
    Yes it is something simple. Leafnode doesn't like the ".invalid" part.
    I've been using leafnode for a couple years now and use ".inet" which it
    seems to like. I know you like to tell the newbies (like me) to set their
    hostname to xxx.xxx.invalid but when I first did that, fetchnews stopped
    working for me. Changing it back to .inet fixed it.
    Oh, and I haven't set up any DNS server for my nntp server and simply use
    the static local IP address I assigned to read the news on my intranet.
    Hope this helps.
    Jeff


    --
    Posted via a free Usenet account from http://www.teranews.com


  9. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 16:23:09 -0800, Jeff@gammalinux.inet wrote:

    > Yes it is something simple. Leafnode doesn't like the ".invalid" part.


    Well, frap.

    And double frap, it does not like ".test" either. :-(
    ..test should work in my opinion.


  10. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 16:25:17 -0500, Bit Twister wrote:

    > Message is from /etc/cron.daily/texpire, See:


    May want to setup a dynamic dns host name (to ensure no one else is using it),
    so you can use a real domain name, safely.

    Checking the source code from the leafnode package ...

    int is_validfqdn(const char *f) {
    /* do not let us fool by trailing dots */
    char *fqdn = strdup(f);
    if (!fqdn) return 0;
    while (fqdn[strlen(fqdn)-1] == '.')
    fqdn[strlen(fqdn)-1] = '\0';

    if (/* reject unqualified names */
    !(strchr(fqdn, '.'))
    /* Red Hat list the FQDN on the same line as localhost, thus,
    * the qualification returns two "localhost*" aliases */
    || 0 == strncasecmp(fqdn, "localhost", 9)
    /* protect against broken hosts or DNS */
    || 0 == strncmp(fqdn, "127.", 4)
    /* SuSE default hostname on some installs is linux.local */
    || 0 == strcasecmp(fqdn, "linux.local")
    /* kill RFC 2606 second- and top-level domains */
    || 0 == strcasecmpsuffix(fqdn, "example.org")
    || 0 == strcasecmpsuffix(fqdn, "example.com")
    || 0 == strcasecmpsuffix(fqdn, "example.net")
    || 0 == strcasecmpsuffix(fqdn, ".example")
    || 0 == strcasecmpsuffix(fqdn, ".invalid")
    || 0 == strcasecmpsuffix(fqdn, ".local")
    || 0 == strcasecmpsuffix(fqdn, ".localdomain")
    || 0 == strcasecmpsuffix(fqdn, ".localhost")
    || 0 == strcasecmpsuffix(fqdn, ".test")
    || 0 == strcasecmpsuffix(fqdn, ".site")
    )
    {
    free(fqdn);
    return 0;
    }
    free(fqdn);
    return 1;
    }

    The readme states that the above restrictions are in place, to try and guarantee
    the generation of unique message ids.

    Regards, Dave Hodgins

    --
    Change nomail.afraid.org to ody.ca to reply by email.
    (nomail.afraid.org has been set up specifically for
    use in usenet. Feel free to use it yourself.)

  11. Re: FQDN for Leafnode

    On Thu, 03 Jan 2008 01:24:25 +0000, Bit Twister wrote:

    > On Wed, 02 Jan 2008 16:23:09 -0800, Jeff@gammalinux.inet wrote:
    >
    >> Yes it is something simple. Leafnode doesn't like the ".invalid" part.

    >
    > Well, frap.
    >
    > And double frap, it does not like ".test" either. :-(
    > .test should work in my opinion.


    David H. has is right. Now, can we change it? Let us know.
    Oh, and I got a "frap" from BT. Where do I get the T-shirt.
    Thanks BT for all you've done for us mere mortals.
    Jeff

    --
    Posted via a free Usenet account from http://www.teranews.com


  12. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 20:33:05 -0500, David W. Hodgins wrote:

    Thank you David for that bit of research.

    Also, thank you to Jeff, I was going to go back and verify .test was
    broke but David's post proves it a waste of time and get back to you.


    > On Wed, 02 Jan 2008 16:25:17 -0500, Bit Twister wrote:
    >
    > if (/* reject unqualified names */
    > !(strchr(fqdn, '.'))


    > || 0 == strcasecmpsuffix(fqdn, ".invalid")
    > || 0 == strcasecmpsuffix(fqdn, ".local")
    > || 0 == strcasecmpsuffix(fqdn, ".test")
    > || 0 == strcasecmpsuffix(fqdn, ".site")



    > The readme states that the above restrictions are in place,


    Source readme, maybe, not in /usr/share/doc/leafnode/README
    or http://www.leafnode.org/doc_en/README-FQDN.html

    I still think .test should be allowed. The .test domain was tagged for
    live testbed testing which helps up line filters if something
    escapes to the Internet.

    Pick the case where a company has it's own Production Usenet server.
    It uses a samecompany.test LAN which is used to regression test
    all software before moving it into production.

    The ntp.samecompany.test would be running from svr.samecompany.test,
    but would not be able to test Leafnode on a workstation test node.


    > to try and guarantee the generation of unique message ids.


    Anyone could name their LAN/node the same as mine and wind up
    generating duplicate message ids.

    Shoot, I could have a registered domain, do a clean install, post for
    a week, another clean install and be generating duplicate message ids. :-(


  13. Re: FQDN for Leafnode

    On Wed, 02 Jan 2008 18:00:45 -0800, Jeff@gammalinux.inet wrote:

    > David H. has is right. Now, can we change it? Let us know.


    I would guess complaints at leafnode.org might do it.

    At best they aught to add the reject name list to their DNS-FAQ.

    > Oh, and I got a "frap" from BT. Where do I get the T-shirt.


    Hahaha.

    > Thanks BT for all you've done for us mere mortals.


    Hey thanks to you for jumping in and giving a solution.

    I tried your .inet without a DNS server and leafnode still complained.
    Started named and it worked. Guess I need to register a domain name.

    Quick search would suggest
    http://www.cydots.com, http://www.smartdots.com, http://www.unonic.com
    to be a free registration answer.

  14. Re: FQDN for Leafnode SOLUTION

    On Wed, 2 Jan 2008 06:01:05 +0000 (UTC), Bit Twister wrote:
    > How do I get rid of the must have FQDN problem.
    > /etc/cron.daily/texpire causes



    Short answer, changed nodes' FQDN
    from wb.home.invalid
    to wb.home.bogus


    Long answer:
    You can get the message because of the following test snippets found
    in the source code. Thanks to David W. Hodgins for that research.

    Do a
    host $(hostname --fqdn)

    and if the parts of the results match any of the following
    "localhost"
    "127."
    "linux.local"
    "example.org"
    "example.com"
    "example.net"
    ".example"
    ".invalid"
    ".local"
    ".localdomain"
    ".localhost"
    ".test"
    ".site"

    you get something like
    Leafnode must have a fully-qualified and globally unique domain name,
    not just "wb.home.invalid".
    Edit your /etc/hosts file to add a unique, fully qualified domain name.
    "localhost.localdomain" or thereabouts will not work;
    it's qualified, but not unique.
    Please see the README-FQDN file for details.

  15. Re: FQDN for Leafnode SOLUTION

    Bit Twister wrote:

    > On Wed, 2 Jan 2008 06:01:05 +0000 (UTC), Bit Twister wrote:
    >> How do I get rid of the must have FQDN problem.
    >> /etc/cron.daily/texpire causes

    >
    >
    > Short answer, changed nodes' FQDN
    > from wb.home.invalid
    > to wb.home.bogus
    >
    >
    > Long answer:
    > You can get the message because of the following test snippets found
    > in the source code. Thanks to David W. Hodgins for that research.
    >
    > Do a
    > host $(hostname --fqdn)
    >
    > and if the parts of the results match any of the following
    > "localhost"
    > "127."
    > "linux.local"
    > "example.org"
    > "example.com"
    > "example.net"
    > ".example"
    > ".invalid"
    > ".local"
    > ".localdomain"
    > ".localhost"
    > ".test"
    > ".site"
    >
    > you get something like
    > Leafnode must have a fully-qualified and globally unique domain name,
    > not just "wb.home.invalid".
    > Edit your /etc/hosts file to add a unique, fully qualified domain name.
    > "localhost.localdomain" or thereabouts will not work;
    > it's qualified, but not unique.
    > Please see the README-FQDN file for details.


    Interesting that you were caught too BT. I am now using "dougshost. domain>", since you thought that my previous "dougshost.mydomain.org" might
    overlap with an issued domain name.

    Leafnode's maintainer Andre is very sure that if you
    use "localhost.localdomain" you will lose posts, and his software will get
    the blame for it. So he insists that you have a "proper" FQDN.
    Interesting to see how far he has gone to enforce that.

    Doug.


  16. Re: FQDN for Leafnode SOLUTION

    On Sat, 05 Jan 2008 01:26:01 +1100, Doug Laidlaw wrote:
    >
    > Interesting that you were caught too BT. I am now using "dougshost. > domain>", since you thought that my previous "dougshost.mydomain.org" might
    > overlap with an issued domain name.


    Normally when I replay to a network problem, I assume a newbie and
    assume they made up a domain name. I'll check domain name with
    whois mydomain.org
    and indicate the domain may be in use. Here try
    $ whois junk-stuff.com
    $ whois junkstuff.com

    You may have to install whois urpmi --wget whois --auto


    Sorry if mydomain.org is yours.

    > Leafnode's maintainer Andre is very sure that if you
    > use "localhost.localdomain" you will lose posts,


    Yes, the message
    "localhost.localdomain" or thereabouts will not work;
    is misleading since I was not using "localhost.localdomain"

    Should be an easy fix to plug in the node name he is checking.


    > and his software will get
    > the blame for it. So he insists that you have a "proper" FQDN.
    > Interesting to see how far he has gone to enforce that.


    Yes I understand, but http://www.rfc-editor.org/rfc/rfc2606.txt
    indicates .test is for testing.


  17. Re: FQDN for Leafnode SOLUTION

    On Thu, 3 Jan 2008, in the Usenet newsgroup alt.os.linux.mandriva, in article
    , Bit Twister wrote:

    >Bit Twister wrote:
    >> How do I get rid of the must have FQDN problem.
    >> /etc/cron.daily/texpire causes


    >Short answer, changed nodes' FQDN
    >from wb.home.invalid
    > to wb.home.bogus


    which fixes the problem, but doesn't solve it.

    >Long answer:
    >You can get the message because of the following test snippets found
    >in the source code. Thanks to David W. Hodgins for that research.


    What they are trying to do is to ensure a unique message ID which is
    a "requirement" of RFC1036 (which itself is _not_ a "standard", having
    been created before that word took it's current meaning in RFCs) section
    2.1.5. Briefly, it states that

    the Message-ID must have the format:



    where full_domain_name is the full name of the host at which the
    message entered the network, including a domain that host is in,

    This "requirement" is more restrictive than those in RFC0822 and 2822
    which also apply and are actually listed as standards.

    There is a DRAFT replacement for RFC1036 (actually, there are and have
    been quite a number of proposed replacements) that can be found on the
    net "draft-ietf-usefor-usefor-12.txt" (except that this draft officially
    expired 6 months ago) and should this "work-in-progress" or it's followup
    ever get adopted, it will be an RFC standard. The current draft has the
    same wording about the Message-ID format, but defines it as a "SHOULD"
    rather than a "MUST" (see RFC2119).

    The reason for the "unique message ID" is that this controls how news
    servers distribute messages - that field is used to identify the
    article rather than by subject, sender, date, or anything else. Modern
    news readers use the message ID for threading purposes (look at the
    References: header in this article). The graybeards who created this
    scheme felt that using a FQDN would help solve the uniqueness problem.

    Changing the FQDN from mumble.invalid to mumble.bogus merely shifts the
    problem once again. A better solution would be for the 'leafnode'
    authors to fix their Message-ID generator, so that the 'left-of-the-@'
    function is more robust. As it stands now, they are just tossing the
    problem to the users who REALLY aren't prepared to come up with a better
    solution (how many times have you seen unique names used for munging
    in Usenet posts - yeah, domain.com or mydomain.com sounds pretty good).

    Old guy

  18. Re: FQDN for Leafnode SOLUTION

    On Fri, 4 Jan 2008, in the Usenet newsgroup alt.os.linux.mandriva, in article
    , Bit Twister wrote:

    >Doug Laidlaw wrote:


    >> Interesting that you were caught too BT. I am now using "dougshost. >> Web domain>", since you thought that my previous "dougshost.mydomain.org"
    >> might overlap with an issued domain name.


    >Normally when I replay to a network problem, I assume a newbie and
    >assume they made up a domain name.


    That's pretty much the case now. People never look at RFC2606 and think
    of using 'example.(com|net|org)' or something ending in 'invalid'.

    >I'll check domain name with
    > whois mydomain.org
    >and indicate the domain may be in use.


    Marginally faster (especially for those without access to whois) would
    simply be to try to look up the domain. Usually, the domain is set such
    that a query will get a referral to a mail server record or similar.
    Everyone on the entire Internet knows that www.somedomain.someplace
    must exist, right ;-)

    >Sorry if mydomain.org is yours.


    It's was registered in Canada in 1996 - but I sometimes I wonder about
    people who own these "obvious" domains. They get their domain names
    posted eleventy-zillion times a month, and who knows how often someone
    blindly plugs the name in to an application - I guess they're counting
    on the exposure for ads or similar.

    >>So he insists that you have a "proper" FQDN.
    >> Interesting to see how far he has gone to enforce that.


    Pity he hasn't put the same amount of energy into creating a more robust
    message-ID generator.

    >Yes I understand, but http://www.rfc-editor.org/rfc/rfc2606.txt
    >indicates .test is for testing.


    But noticing the FQDN is one of his "bad names" and branching to a piece
    of code that generates a RHS part based on a randomly generated (at
    installation) string would be less of a hassle for everyone.

    Old guy

  19. Re: FQDN for Leafnode SOLUTION

    On Fri, 04 Jan 2008 14:27:37 -0600, Moe Trin wrote:

    > Pity he hasn't put the same amount of energy into creating a more robust
    > message-ID generator.


    Maybe something as simple as (postnumer)nic_mac_id@anything.user.uses
    No nic, use cpu serial number. No cpu serial, use drive UUID.

  20. Re: FQDN for Leafnode SOLUTION

    On Fri, 4 Jan 2008, in the Usenet newsgroup alt.os.linux.mandriva, in article
    , Bit Twister wrote:

    >Moe Trin wrote:


    >> Pity he hasn't put the same amount of energy into creating a more
    >> robust message-ID generator.

    >
    >Maybe something as simple as (postnumer)nic_mac_id@anything.user.uses
    >No nic, use cpu serial number. No cpu serial, use drive UUID.




    Hey - that sounds like SPYWARE!!! The guys in the Black Helicopters
    would be able to trace me!!!111!!!



    To please the propeller_beanie crowd, it would probably be desirable
    to use an MD5 hash of these numbers or similar, but there might still
    be privacy issues from those that refuse to understand hashes. While
    RFC1036 doesn't mention a size limit of the field, and the proposed
    'draft-ietf-usefor-usefor-12.txt' is talking about 250 characters, I'm
    seeing posts whining about the length of googlegroup message-IDs which
    are now ~66 characters. I suppose another alternative would be to
    use 'time_t' and 10-20 bytes of the mime-encoded output of /dev/random
    which would probably be unique enough.

    Old guy

+ Reply to Thread
Page 1 of 3 1 2 3 LastLast