Bug 33929

Summary: Section 15.15 of the FreeBSD Porter's Handbook needs to be updated.
Product: Documentation Reporter: Yarema <yds>
Component: Books & ArticlesAssignee: freebsd-doc (Nobody) <doc>
Status: Closed FIXED    
Severity: Affects Only Me CC: dwhite
Priority: Normal    
Version: Latest   
Hardware: Any   
OS: Any   
Attachments:
Description Flags
PortersHandbook.diff.gz none

Description Yarema 2002-01-15 23:00:02 UTC 
Now that PR31473 has been committed, Section 15.15 of the FreeBSD Porter's
Handbook needs to be updated.  I chose to add user 'courier' and group
'courier' as UID and GID 62.  There's a problem with that section of the
Handbook.  Since the qmail users are not sorted by UID 'msql' and 'qmails'
ended up with the same UID which is probaly a security hole for either one
or both of those ports.  The same thing could be said about 'ifmail' sharing
the same UID with 'pgsql'.  While putting 'ifmail' in group 'uucp' might be
the Right Thing (TM) for that port.  Using the same UID as PostgreSQL is
probably a Bad Thing (TM).

The patch below updates:

doc/en_US.ISO8859-1/books/porters-handbook/book.sgml
doc/ru_RU.KOI8-R/books/porters-handbook/book.sgml
doc/ja_JP.eucJP/books/porters-handbook/book.sgml

to bring them all in sync.  Changes include adding user 'bind' UID 53 to
the top of the list, adding user 'courier' and sorting the list by UID so
that future UID duplication is easier to avoid.  User 'bind' UID 53 already
exists in /usr/src/etc/master.passwd and this section documents UIDs 50
through 999.  So user 'bind' clearly needs to be documented here.

Of interest to the mail/sendmail port is that I removed:

smmsp:*:90:90:Sendmail Queue:/nonexistent:/nonexistent

since user 'smmsp' already exists as UID 25 in /usr/src/etc/master.passwd --
patching the mail/sendmail port to use UID 25 instad of 90 makes more sense
(to me at least).

The patch is gzipped and uuencoded since it contains both ru_RU.KOI8-R and
ja_JP.eucJP encodings which are not 7bit.
Comment 1 Giorgos Keramidas freebsd_committer freebsd_triage 2002-01-24 00:47:55 UTC 
Adding to audit-trail:

    Date: Wed, 16 Jan 2002 08:52:57 +0100
    From: dirk.meyer@dinoex.sub.org (Dirk Meyer)
    Subject: Re: Section 15.15 of the FreeBSD Porter's Handbook needs to be updated.

    > to bring them all in sync.  Changes include adding user 'bind' UID 53 to
    > the top of the list, adding user 'courier' and sorting the list by UID so
    > that future UID duplication is easier to avoid.  User 'bind' UID 53 already
    > exists in /usr/src/etc/master.passwd and this section documents UIDs 50
    > through 999.  So user 'bind' clearly needs to be documented here.
    >
    > Of interest to the mail/sendmail port is that I removed:
    >
    > smmsp:*:90:90:Sendmail Queue:/nonexistent:/nonexistent
    >
    > since user 'smmsp' already exists as UID 25 in /usr/src/etc/master.passwd --
    > patching the mail/sendmail port to use UID 25 instad of 90 makes more sense
    > (to me at least).

    1) The sendmail port uses an already existiting UID/GID and does not remove it.
    2) It may violate POLA to make the sendmail port create this UID/GID on
       older FreeBSD-Systems with 25 instead of 90.

    kind regards Dirk
Comment 2 Giorgos Keramidas freebsd_committer freebsd_triage 2002-01-24 00:49:28 UTC 
Adding to audit trail:

Date: Wed, 16 Jan 2002 06:06:06 -0500
From: Yarema <yds@CoolRat.org>
Subject: Re: Section 15.15 of the FreeBSD Porter's Handbook needs to be updated.

--On Wednesday, January 16, 2002 08:52:57 +0100 Dirk Meyer
<dirk.meyer@dinoex.sub.org> wrote:

>> to bring them all in sync.  Changes include adding user 'bind' UID 53 to
>> the top of the list, adding user 'courier' and sorting the list by UID so
>> that future UID duplication is easier to avoid.  User 'bind' UID 53
>> already exists in /usr/src/etc/master.passwd and this section documents
>> UIDs 50 through 999.  So user 'bind' clearly needs to be documented here.
>>
>> Of interest to the mail/sendmail port is that I removed:
>>
>> smmsp:*:90:90:Sendmail Queue:/nonexistent:/nonexistent
>>
>> since user 'smmsp' already exists as UID 25 in
>> /usr/src/etc/master.passwd -- patching the mail/sendmail port to use UID
>> 25 instad of 90 makes more sense (to me at least).
>
> 1) The sendmail port uses an already existiting UID/GID and does not
> remove it.

I noticed that.

> 2) It may violate POLA to make the sendmail port create this
> UID/GID on older FreeBSD-Systems with 25 instead of 90.

Perhaps I over did it with having the patch remove
smmsp:*:90:90:Sendmail Queue:/nonexistent:/nonexistent

I think the rest of the patch makes good sense and fixes a number of
inconsistencies in the documentation.  Perhaps if the right people are
proded it might get applied. :)

Dirk, what about ifmail using UID 70, the same as pgsql?  Wouldn't it make
more sense for news/ifmail to use 65 which seems to be available?

-- 
Yarema
Comment 3 Giorgos Keramidas freebsd_committer freebsd_triage 2002-01-24 00:51:21 UTC 
Adding to audit-trail:

Date: Thu, 17 Jan 2002 11:38:31 +0100
From: dirk.meyer@dinoex.sub.org (Dirk Meyer)
Subject: Re: Section 15.15 of the FreeBSD Porter's Handbook needs to be updated.

> Dirk, what about ifmail using UID 70, the same as pgsql?  Wouldn't
> it make more sense for news/ifmail to use 65 which seems to be
> available?

I can changes this 70 was assigned way back in the years:
http://www.FreeBSD.org/cgi/cvsweb.cgi/ports/news/ifmail/Attic/pkg-install

While the postgresql7 change was just some 10 Month ago:
http://www.FreeBSD.org/cgi/cvsweb.cgi/ports/databases/postgresql7/pkg-install?rev=1.3&content-type=text/x-cvsweb-markup

If the new one keeps free, I will take the anger of the users, (less
then postgresql7)

kind regards Dirk

- Dirk Meyer, Im Grund 4, 34317 Habichtswald, Germany
Comment 4 Pete Fritchman freebsd_committer freebsd_triage 2002-02-05 08:36:18 UTC 
Responsible Changed
From-To: freebsd-ports->freebsd-doc

A -doc guy should take care of this.
Comment 5 Mark Linimon freebsd_committer freebsd_triage 2004-06-30 03:33:07 UTC 
State Changed
From-To: open->closed

Except for the courier change (which is no longer correct), all 
of these changes have been made long ago.  This PR seems to have 
been just forgotten.