Bug 1520024 - Review Request: netatalk - Open Source Apple Filing Protocol (AFP) File Server
Summary: Review Request: netatalk - Open Source Apple Filing Protocol (AFP) File Server
Keywords:
Status: CLOSED DUPLICATE of bug 1658199
Alias: None
Product: Fedora
Classification: Fedora
Component: Package Review
Version: rawhide
Hardware: All
OS: Linux
unspecified
medium
Target Milestone: ---
Assignee: Neal Gompa
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2017-12-02 04:44 UTC by Ryan Breaker
Modified: 2018-12-11 13:54 UTC (History)
8 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2018-12-11 13:54:02 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
Updated netatalk.spec (25.89 KB, text/plain)
2018-01-13 22:46 UTC, Alan Hamilton
no flags Details

Description Ryan Breaker 2017-12-02 04:44:57 UTC
Spec URL: http://breaker.rocks/netatalk.spec
SRPM URL: http://breaker.rocks/netatalk-3.1.11-1.fc28.src.rpm
Description: Netatalk is a freely-available Open Source AFP file server. A *NIX/*BSD system running Netatalk is capable of serving many Macintosh clients simultaneously as an AppleShare file server.
Fedora Account System Username: ryanbreaker

This is my first package contribution to any Linux distribution and I also currently seek a sponsor. This package was previously in the Fedora repositories but was later orphaned and abandoned in 2016 (https://src.fedoraproject.org/rpms/netatalk). This first package of mine is an attempt to revive it to its latest version as well as maintain it.

This revision is much like the orphaned version but with Patch1 removed, which was worked into the upstream of the project itself, as well as the sources updated to the most recent version of 3.1.11 from the previous 3.1.7.

Except for arch s390x, which I am still troubleshooting and may require assistance with, here is a successful scratch build: https://koji.fedoraproject.org/koji/taskinfo?taskID=23489939

I have also tested this package and its service in my own environment successfully.

Thanks!

Comment 1 Neal Gompa 2017-12-02 18:03:27 UTC
Taking this review. I'll also be sponsoring this packager,

Comment 2 Ryan Breaker 2017-12-02 18:20:21 UTC
(In reply to Neal Gompa from comment #1)
> Taking this review. I'll also be sponsoring this packager,

Thanks!

I did notice the compiled binary RPM has some rpmlint issues which I'm still trying to figure out, and may need some help with.

For example:

* E: missing-call-to-setgroups-before-setuid /usr/sbin/afpd
* E: missing-call-to-setgroups-before-setuid /usr/sbin/cnid_dbd

Looking through the source for each of these however, the order of the function calls appear to be correct (https://wiki.sei.cmu.edu/confluence/display/c/POS36-C.+Observe+correct+revocation+order+while+relinquishing+privileges).

I'm still new to RPMs so thank you for your patience.

Comment 3 Alan Hamilton 2018-01-13 22:44:15 UTC
From looking at what rpmlint is doing, it's dumping the external symbols from the executable and assuming that the calls are in the order that they appear in the symbol table. So it's expecting setgid to come before setuid, and it's not in this build.

However, I'm not sure that's a valid assumption on rpmlint's part. As long as the code looks okay, I'd say it's safe to call it a false positive.

Some of the others are easy. The dir-or-file-in-var-lock error can be fixed by running configure with --with-lockfile=/run/netatalk to move the file to the standard run directory.The %ghost directive should be changed to /run/netatalk

I've taken a crack at fixing some of the other issues in the spec file, notably moving /var/lock/netatalk to /run/netatalk, and removing the ownership of /var/lib.

Comment 4 Alan Hamilton 2018-01-13 22:46:24 UTC
Created attachment 1380830 [details]
Updated netatalk.spec

Comment 5 HAT 2018-03-27 15:54:40 UTC
To: Ryan Breaker

your spec file:

> # Ghost lock dir.
> mkdir -p %{buildroot}/var/lock/netatalk

> %ghost %dir /var/lock/netatalk

The /var/lock/netatalk is file, not directory.


To: Alan Hamilton

your spec file:

> # Ghost lock dir.
> mkdir -p %{buildroot}/run/netatalk

> %ghost %dir /run/netatalk

The /run/netatalk is file, not directory.

Comment 6 HAT 2018-03-27 16:25:57 UTC
http://netatalk.sourceforge.net/wiki/index.php/Netatalk_3.1.11_SRPM_for_Fedora_and_CentOS
Spec URL: http://www003.upp.so-net.ne.jp/hat/files/netatalk.spec
SRPM URL: http://www003.upp.so-net.ne.jp/hat/files/netatalk-3.1.11-1.2.fc29.src.rpm

I updated Ryan Breaker's SRPM.

pam_ck_connector.so isn't always installed
https://bugzilla.redhat.com/show_bug.cgi?id=1520024

require dconf package
https://bugzilla.redhat.com/show_bug.cgi?id=1248157

require perl-IO-Socket-INET6 for asip-status.pl script 

define with_ldap
https://bugzilla.redhat.com/show_bug.cgi?id=1249403

The UAM path should be netatalk, not atalk
https://bugzilla.redhat.com/show_bug.cgi?id=1249404

fix multilib conflict of tracker-devel
https://sourceforge.net/p/netatalk/bugs/637/

Support EL6.

Support Fedora 28 and later.

However, many problems of rpmlint aren't being corrected.

Comment 7 Steven Schlansker 2018-05-06 16:27:13 UTC
I have been running the discontinued fc26 netatalk, which I uninstalled.  I then upgraded the host f27 -> f28.
HAT, I was not able to download your SRPM (404) but I did build 3.1.11-1.3.fc28 based on your specfile.  I installed it, restored the afp.conf.rpmsave so I had configuration, and am working through a Time Machine backup.  Thanks to all of you for picking this package up, looking forward to seeing it in the repos soon.

Comment 9 Andrew Bauer 2018-11-28 13:12:24 UTC
Most of the issues identified in this review appear to have been fixed by HAT, who appears to be upstream developer. Thanks, HAT.

Is the original author still interested in completing this review?
I am not in the sponsor group, but I can help finish the review if that is what is needed.

If not, can this review be closed so that I can create a new review request?

Comment 10 Andrew Bauer 2018-12-11 13:54:02 UTC
Since this review request has not seen any recent activity, I am going to close this out to allow a new request to take its place.

I could use a reviewer for the new review request:
https://bugzilla.redhat.com/show_bug.cgi?id=1658199

*** This bug has been marked as a duplicate of bug 1658199 ***


Note You need to log in before you can comment on or make changes to this bug.