Red Hat Bugzilla – Bug 188482
Review Request: scipy-0.4.8
Last modified: 2007-11-30 17:11:30 EST
Spec Name or Url: http://nbecker.dyndns.org:8080/scipy-0.4.8.spec
SRPM Name or Url: http://nbecker.dyndns.org:8080/scipy-0.4.8.src.rpm
Description: scientific python. This version is for numpy-0.9.6
Hi Neal, the two URLs return "404 Not Found" for me. Is this submission
These should work:
Hi Neal, this isn't a full review, just a few quick observations:
- Is this your first package? If so, we'll add FE-NEEDSPONSOR per:
- incorrect BuildRoot:
- please either make Source0 a full URL or indicate (with some
comments in the spec file) how one can generate the tar-ball
from, for instance, the upstream CVS [this is so we can more
easily verify that source matches upstream]
- please remove Prefix and Vendor
- please add a changelog
- please consider changing %defattr(-,root,root) to
You might want to take a look at the review guidelines:
and build your rpm in mock ("yum install mock") and run your generated
RPM through rpmlint ("yum install rpmlint") to see if you might find any
other items -- it could make the job easier easier for reviewers!
OK, I have uploaded scipy-0.4.8-2. I have tried to address the comments.
I did not use mach, because the setup looks a bit involved. I did try to
supply requires and buildrequires manually.
I did run rpmlint. There are some complaints, but I think they are ignorable.
I removed d1mach patch. If you have the latest FC5 gcc I think the bug has
been fixed. Is it safe to assume nobody will try to use older gcc?
Hi Neal, heres a few more comments (not a thorough review):
+ license is OK: its BSD-w/o-advertise-clause
+ source matches upstream
- missing license file (please add "%doc LICENSE.txt")
- missing "BuildRequires: numpy"
- please use "%setup -q"
Also, *please* build your package locally in mock. The commands are:
1) yum install mock
2) add yourself to the mock group (vi /etc/group)
3) log out/in for group addition to take full effect
4) run "mock -r fedora-5-i386-core scipy-0.4.8-2.src.rpm"
and it will help you identify problems. Its a very good idea since
mock is what the buildsystem itself uses.
I believe I have done all that was requested. Updated srpm is here
Sorry make that:
Mass-block FE-NEEDSPONSOR for the six review requestsÂ¹ of Neal Becker. Neal,
when you get sponsorship, you will have to unblock it for all your requests.
Â¹)Â Actually the four that do not block yet FE-NEEDSPONSOR.
Hi Neal, this is still not a complete review but we have at this point
covered a number of the review items.
+ builds in mock on FC5 i386
+ dir ownership appears to be OK
+ permissions look OK
- rpmlint returns 168 errors/warnings of which 167 are:
126 are devel-file-in-non-devel-package
39 are non-executable-script
2 are script-without-shellbang
and these can probably be safely ignored. The one remaining
W: scipy summary-ended-with-dot Scipy: array processing
for numbers, strings, records, and objects.
and you can easily fix it.
1) There are a number of files in directories of the form
Perhaps they should go in sub-packages such as scipy-test and
scipy-docs or similar? Whats your opinion of putting them in
2) What is your sponsorship status? Are any of the other packages
you submitted closer to passing review? I'm just a little leery
about sponsoring you since I'm not familiar with any of the other
packages and just don't have enough free time to learn about them.