Bug 251074 - mail/nullmailer: update to 2.2
Summary: mail/nullmailer: update to 2.2
Status: Closed FIXED
Alias: None
Product: Ports & Packages
Classification: Unclassified
Component: Individual Port(s) (show other bugs)
Version: Latest
Hardware: Any Any
: --- Affects Only Me
Assignee: Fernando Apesteguía
URL: http://untroubled.org/nullmailer/Chan...
Keywords:
Depends on:
Blocks:
 
Reported: 2020-11-12 13:15 UTC by Uffe Jakobsen
Modified: 2020-11-13 07:15 UTC (History)
3 users (show)

See Also:


Attachments
mail/nullmailer ports update patch (6.84 KB, patch)
2020-11-12 13:15 UTC, Uffe Jakobsen
no flags Details | Diff
mail/nullmailer ports update patch (8.31 KB, patch)
2020-11-12 22:16 UTC, Uffe Jakobsen
no flags Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Uffe Jakobsen 2020-11-12 13:15:23 UTC
Created attachment 219598 [details]
mail/nullmailer ports update patch

Hello,

The attached patch updates port mail/nullmailer from version 1.13 to latest version 2.2

The update is long due as nullmailer 1.13 was released in 2013 and I've been missing lots of the new features for a long time (TLS improvements etc)

Here is a list of the improvements from 1.13 up to version 2.2:

-------------------------------------------------------------------------------
Changes in version 2.2

- nullmailer-send no longer generates bounces for rejected bounces.
  Thanks Fejes József

- Fixed compile error in sendmail on GCC older than 4.9.

- Fixed treating authentication failure as message rejection.
  Thanks Fejes József

- nullmailer-inject now sets the full name of the sender to the user
  name as a fallback. This helps distinguish system sent messages when
  the MTA rewrites the address (as does GMail, for example).

- Fixed compatibility issue with gnutls 3.6 (and possibly others).
-------------------------------------------------------------------------------
Changes in version 2.1

- Added support for TLS anonymous authentication.
  Thanks Uffe Jakobsen.

- Fixed sendmail wrapper handling of empty sender on command line.
  Thanks Sebastian Wiedenroth.

- Fixed handling of quoted strings in the "remotes" file.
  Thanks Mihai Moldovan.

- Fixed nullmailer-inject handling of leading "From " lines.

- Some build fixes.

- Fixed bogus temporary gethostbyname error message when the protocol
  source address was incorrect.

- Fixed potential race condition in tests.
  Thanks Felix Lechner.

- Fixed handling of time values on 32-bit big-endian systems.
  Thanks Felix Lechner.

-------------------------------------------------------------------------------
Changes in version 2.0

- Added support to nullmailer-send to move permanently failing messages
  out of the queue, and to generate bounce messages.

- Added support for IPv6.

- Added program to generate bounce/delay messages.

- Added an "allmailfrom" control file to nullmailer-queue, causing all
  messages to share a hard-coded envelope sender.

- Added logging the message sender/recipient in nullmailer-send.

- Improved handling of system errors when reading config files.

- Secured handling of password options for protocol modules.

- Support standard shell quoting for options in the "remotes" file.

- Added protocol option to set a separate TLS client private key file.

- Added protocol option to bind the source address on connections.

- Fixed nullmailer-inject to report errors to stderr.

- Fixed gnutls cast to pointer from integer of different size warning.

- Fixed nullmailer-inject and -queue to handle the null (empty) sender
  address. Needed for RFC 3798 (Message Disposition Notification).

- Moved spool directory to /var/spool/nullmailer like other MTAs.

-------------------------------------------------------------------------------

/Uffe
Comment 1 Fernando Apesteguía freebsd_committer freebsd_triage 2020-11-12 18:43:22 UTC
^Triage: [tags] in issue Titles are deprecated.

^Triage: Simplifying title

^Triage: If there is a changelog or release notes URL available for this version, please add it to the URL field 

Q/A: Patches do not conform to proper format:

 /tmp/251074/mail/nullmailer/files/patch-lib_fdbuf_tlsibuf.cc: patch was not generated using ``make makepatch''.  It is recommended to use ``make makepatch'' when you need to [re-]generate a patch to ensure proper patch format.
 /tmp/251074/mail/nullmailer/files/patch-lib_fdbuf_tlsobuf.cc: patch was not generated using ``make makepatch''.  It is recommended to use ``make makepatch'' when you need to [re-]generate a patch to ensure proper patch format.
 /tmp/251074/mail/nullmailer/files/patch-Makefile.in: patch was not generated using ``make makepatch''.  It is recommended to use ``make makepatch'' when you need to [re-]generate a patch to ensure proper patch format.
 /tmp/251074/mail/nullmailer/files/patch-src-Makefile.in: patch was not generated using ``make makepatch''.  It is recommended to use ``make makepatch'' when you need to [re-]generate a patch to ensure proper patch format.

whould you give it a try and regenerate them?

Q/A: PORTREVISION should be removed


Thanks!
Comment 2 Uffe Jakobsen 2020-11-12 19:34:59 UTC
Quote: whould you give it a try and regenerate them?

Well, in the past I was port maintainer for around ~40 ports

And gave them all up due to lack of resources - both time but also new requirements for ports - running poudriere etc etc... 

My work is done on a small 10 years old laptop ("netbook" class) - so both disk and memory resources are also low...

That is the reason for the diff not being a subversion diff - I have to use portsnap to conserve diskspace - otherwise I have no space for poudriere jails...

Also many of the patch files that you point to are not touched by my update - they are left as they were before my work began.

As you can see I'm not really opting in as maintainer - because I think the process for porters have become too "heavy" to lift in my spare-time...

Just delivering a subversion diff will require that I delete poudriere...

I'll have a look at what can be done with the hardware I have available :-S
Comment 3 Chris Hutchinson 2020-11-12 20:13:15 UTC
(In reply to Uffe Jakobsen from comment #2)
I feel your pain. I'm @ ~160 ports.
Here's a quickie that'll produce a usable svn diff
for all intents and purposes:
( the following presumes you do NOT already have a DEV
folder in your home directory)

$ cd ~
$ svn co --depth empty svn://svn.freebsd.org/ports/head DEV
$ svn up --set-depth empty DEV/mail
$ svn up DEV/mail/nullmailer

# apply your patch. Then...
$ cd DEV/mail
$ svn diff --ignore-properties > mail_nullmailer.diff

Then simply upload (attach) mail_nullmailer.diff to this
pr, and depreciating the one here currently.

All told, svn will only be pulling in a few k. So space
need not be a concern.

Hope this helps! :-)

--Chris
Comment 4 Chris Hutchinson 2020-11-12 20:18:50 UTC
(In reply to Chris Hutchinson from comment #3)
OH. Forgot to mention. In case it hasn't already occurred
to you. You'll want to delete any backup files created
by patch after you run your patch, and before you svn diff...

--Chris
Comment 5 Fernando Apesteguía freebsd_committer freebsd_triage 2020-11-12 21:19:09 UTC
(In reply to Uffe Jakobsen from comment #2)
Ok, no problem.

I can take care of that.
Comment 6 Uffe Jakobsen 2020-11-12 22:03:29 UTC
Hello Chris and Fernando,

Thanks to both of you for accepting and suggesting workarounds to my currenty limited hardware situation...

I think that I've managed to get along with the help of the suggestions from Chris...

I'll post my second try in a few minuttes :-)

Kind regards Uffe
Comment 7 Uffe Jakobsen 2020-11-12 22:16:09 UTC
Created attachment 219613 [details]
mail/nullmailer ports update patch

Hello,

Ok, here is my second attempt to produce this patch with my currently limited hardware resources :-)

kind regards Uffe :-)
Comment 8 Chris Hutchinson 2020-11-12 22:58:50 UTC
(In reply to Uffe Jakobsen from comment #7)
Well done Uffe! Done like a master. :-)
Glad it worked for you.

--Chris
Comment 9 Uffe Jakobsen 2020-11-13 01:18:07 UTC
@Chris: Thanks :-)

FYI: Based on your suggestion I found a way to improve my disk usage situation a bit:

By doing "sparse" subversion checkout of the following files and dirs in ports (in addition to the mail/nullmailer directory that you suggested): 

Mk/*
Makefile
mail/Makefile

With these extra files/dirs in a "sparse" subversion checkout - I found out that you can actually build and test the mail/nullmailer port within the "sparse" subversion extract.

That enabled me to get rid of a full portsnap extract - freeing up some extra much needed disk space.

/Uffe :-)
Comment 10 Chris Hutchinson 2020-11-13 01:25:59 UTC
(In reply to Uffe Jakobsen from comment #9)
Cool, huh? :-)
Don't tell anybody; but I am sometimes guilty of
building/testing ports simply checking out the port,
making necessary patches, and then building them in
the same way I suggested you could check out your
target port. You can perform the entire operation,
including build from the "sparsely" checked out
ports repo. But remember, I never suggested any of
this. Because I would never do that. ;-)

--Chris
Comment 11 commit-hook freebsd_committer freebsd_triage 2020-11-13 07:13:04 UTC
A commit references this bug:

Author: fernape
Date: Fri Nov 13 07:12:20 UTC 2020
New revision: 555007
URL: https://svnweb.freebsd.org/changeset/ports/555007

Log:
  mail/nullmailer: update to 2.2

  ChangeLog: http://untroubled.org/nullmailer/ChangeLog

  PR:	251074
  Submitted by:	uffe@uffe.org

Changes:
  head/mail/nullmailer/Makefile
  head/mail/nullmailer/distinfo
  head/mail/nullmailer/files/patch-Makefile.in
  head/mail/nullmailer/files/patch-fdobuf.h
  head/mail/nullmailer/files/patch-lib_fdbuf_tlsibuf.cc
  head/mail/nullmailer/files/patch-lib_fdbuf_tlsobuf.cc
  head/mail/nullmailer/files/patch-src-Makefile.in
  head/mail/nullmailer/pkg-plist
Comment 12 Fernando Apesteguía freebsd_committer freebsd_triage 2020-11-13 07:15:58 UTC
Committed,

Thank you all for your work.