Bug 591545

When possible, we should avoid of using Epoch, it may interfere some other packages which depends on particular version of apache-commons-net.


In this package, we can avoid of using epoch in the following two ways:

Release:        3%{?dist}
Provides:       jakarta-%{short_name} = %{version}-%{release}
Obsoletes:      jakarta-%{short_name} < 2.0-3





See 
http://fedoraproject.org/wiki/Package_Renaming_Process

http://fedoraproject.org/wiki/Packaging:NamingGuidelines#Renaming.2Freplacing_existing_packages

I can take this one.

Please hold the review until I upload new spec/srpm. Chen is right and there will be bigger modifications to the spec because of this change.

Changes done:
Spec URL: http://sochotni.fedorapeople.org/apache-commons-net.spec
SRPM UR: http://sochotni.fedorapeople.org/apache-commons-net-2.0-3.fc12.src.rpm

I note that this is a re-review due to a package rename.

When you say Chen is right, does that mean you intended to drop the use of epoch in your provides/obsoletes? Is there any point in specifying an epoch of zero?

Apart from that, there is really nothing else wrong with the package. The rpmlint report only has false positives, the rest of the package is to the guidelines. You've clearly been taking notes from the other apache-commons-* reviews, I like easy reviews. :-)

On the condition that you just clarify your position on the use of epoch as mentioned above, this package is:

APPROVED!



On a side note for future reference, did you know that your two calls to install on lines 71/72 and again on lines 83/84 can be combined into a single call? For example, these two lines:

install -d -m 755 foo_dir
install -p -m 644 bar_file foo_dir/bar_file

Are equivalent to this one line:

install -pD -T -m 644 bar_file foo_dir/bar_file

(In reply to comment #5)
> When you say Chen is right, does that mean you intended to drop the use of
> epoch in your provides/obsoletes? Is there any point in specifying an epoch of
> zero?

I meant that packaging guidelines state that we should avoid using Epoch if possible. This was possible with this rename because it was enough to specify epoch in provides/obosletes, no need to actually specify one for this package. It was there from way back, probably when no guidelines on Epoch use existed.

Current guidelines state (2nd link from Chen):
If the provided package had an Epoch set, it must be preserved in both the Provides and Obsoletes. It may and should be removed from the actual new package. 


> On a side note for future reference, did you know that your two calls to
> install on lines 71/72 and again on lines 83/84 can be combined into a single
> call? For example, these two lines:
> 
> install -d -m 755 foo_dir
> install -p -m 644 bar_file foo_dir/bar_file
> 
> Are equivalent to this one line:
> 
> install -pD -T -m 644 bar_file foo_dir/bar_file    

Thanks, I'll keep it in mind :-) I suppose 2 commands were originally mkdir && cp and I just converted them to install.


New Package CVS Request
=======================
Package Name: apache-commons-net
Short Description: Internet protocol suite Java library
Owners: sochotni
Branches: 
InitialCC:

CVS Done

Koji build:
http://koji.fedoraproject.org/koji/buildinfo?buildID=174135