Jason Tibbits response to my post pointed me in the right direction. Here's my reply to him, which I neglected to copy to the list. The problem was not in the coding of spec file but rather in how fedora-review determines which spec file to use.
Sanford
-------- Forwarded Message -------- Subject: Re: [Fedora-packaging] directory ownership problem Date: Wed, 12 Jul 2017 15:46:39 -0400 From: Sanford Rockowitz rockowitz@minsoft.com Organization: Minaret Software To: Jason L Tibbitts III tibbs@math.uh.edu
Jason,
Thanks for the swift reply. Your request for the a link to the SRPM pointed me in the right direction. Even though the -n option to fedora-review specifies the name of the spec file in the current directory, fedora-review was using a slate spec file in the input srpm. So it seems to be a workflow issue, not a spec file syntax issue.
Sanford
On 07/12/2017 01:58 PM, Jason L Tibbitts III wrote:
"SR" == Sanford Rockowitz rockowitz@minsoft.com writes:
SR> I hope this list is appropriate for asking a question regarding rpm SR> file ownership. As a relative rpm newbie, I suspect I'm missing some SR> piece of "secret sauce" that's just obvious to anyone with rpm SR> experience. If there's a more appropriate place to post the SR> question, I'd appreciate a pointer. Thanks in advance.
It's always a good idea to provide a link to your SRPM. Without it, all I can say is that what you're doing _should_ work but that I'd have to see the whole thing and build it myself to see exactly what's happening.
- J<
Sanford Rockowitz rockowitz@minsoft.com wrote:
Thanks for the swift reply. Your request for the a link to the SRPM pointed me in the right direction. Even though the -n option to fedora-review specifies the name of the spec file in the current directory, fedora-review was using a slate spec file in the input srpm. So it seems to be a workflow issue, not a spec file syntax issue.
Didn't Fedora-review warn you that the spec files differed? It really should wave a big red flag in such cases.
Björn Persson
Bjorn,
fedora-review did warn that the 2 spec files differed, but more in the form of an observation than waving a big red flag.
The old workflow used the "--name <package name>" method of specifying the spec file and SRPM, on the (mis-)understanding that the explicit spec file would override the spec file in the SRPM.
Here's how the --name (-n) and related --rpm-spec (-r) options are documented in the man file:
$ fedora-review -n <package name>
This alternative usage expects <package name>.spec and source rpm in current directory. ...
$ fedora-review --rpm-spec -n <srpm path>
This form accepts a single path to a source rpm. It uses the specfile bundled in the srpm package.
I read this as implying that without the --rpm-spec option, fedora-review uses the explicitly specified specfile.
At startup, fedora-review reported:
INFO: Processing local files: ddcutil INFO: Getting .spec and .srpm Urls from : Local files in /shared/playproj/i2c/package/fedora/fedrev_temp INFO: --> SRPM url: file:///shared/playproj/i2c/package/fedora/fedrev_temp/ddcutil-0.8.3-1.fc25.src.rpm INFO: --> Spec url: file:///shared/playproj/i2c/package/fedora/fedrev_temp/ddcutil.spec
Generated file review.txt contains the following lines:
[!]: Spec file according to URL is the same as in SRPM. Note: Spec file as given by url is not the same as in SRPM (see attached diff). See: (this test has no URL)
Later there's a diff of the 2 files. That's it. No discussion of how it might affect the report. I took it as something to be cleaned up later, and kept trying to figure out what was wrong with the spec file.
Sanford
On 07/13/2017 05:03 AM, Björn Persson wrote:
Sanford Rockowitz rockowitz@minsoft.com wrote:
Thanks for the swift reply. Your request for the a link to the SRPM pointed me in the right direction. Even though the -n option to fedora-review specifies the name of the spec file in the current directory, fedora-review was using a slate spec file in the input srpm. So it seems to be a workflow issue, not a spec file syntax issue.
Didn't Fedora-review warn you that the spec files differed? It really should wave a big red flag in such cases.
Björn Persson
packaging mailing list -- packaging@lists.fedoraproject.org To unsubscribe send an email to packaging-leave@lists.fedoraproject.org
packaging@lists.fedoraproject.org