Is there somewhere a succinct guide to just how much extra crud we need
to install on RHEL to make use of EPEL packages?
[root at troll-1 noarch]# rpm -ivh R2spec-4.1.0-1.el6.noarch.rpm
error: Failed dependencies:
fedora-packager is needed by R2spec-4.1.0-1.el6.noarch
[root at troll-1 noarch]# wget
http://dl.fedoraproject.org/pub/epel/6/x86_64/fedora-packager-0.5.9.4-1.el6.noarch.rpm
[...]
[root at troll-1 noarch]# rpm -ivh fedora-packager-0.5.9.4-1.el6.noarch.rpm
error: Failed dependencies:
bodhi-client is needed by fedora-packager-0.5.9.4-1.el6.noarch
fedora-cert = 0.5.9.4-1.el6 is needed by
fedora-packager-0.5.9.4-1.el6.noarch
fedpkg >= 1.0 is needed by fedora-packager-0.5.9.4-1.el6.noarch
koji is needed by fedora-packager-0.5.9.4-1.el6.noarch
mock is needed by fedora-packager-0.5.9.4-1.el6.noarch
redhat-rpm-config is needed by fedora-packager-0.5.9.4-1.el6.noarch
rpmdevtools is needed by fedora-packager-0.5.9.4-1.el6.noarch
rpmlint is needed by fedora-packager-0.5.9.4-1.el6.noarch
ykpers is needed by fedora-packager-0.5.9.4-1.el6.noarch
Oy.
- Allen S. Rout
Trying to use current R2spec on RHEL.
16 messages · Marc Schwartz, Allen S. Rout, Torbjørn Lindahl +2 more
On Mar 12, 2012, at 3:10 PM, Allen S. Rout wrote:
Is there somewhere a succinct guide to just how much extra crud we need to install on RHEL to make use of EPEL packages?
[root at troll-1 noarch]# rpm -ivh R2spec-4.1.0-1.el6.noarch.rpm
error: Failed dependencies:
fedora-packager is needed by R2spec-4.1.0-1.el6.noarch
[root at troll-1 noarch]# wget http://dl.fedoraproject.org/pub/epel/6/x86_64/fedora-packager-0.5.9.4-1.el6.noarch.rpm
[...]
[root at troll-1 noarch]# rpm -ivh fedora-packager-0.5.9.4-1.el6.noarch.rpm
error: Failed dependencies:
bodhi-client is needed by fedora-packager-0.5.9.4-1.el6.noarch
fedora-cert = 0.5.9.4-1.el6 is needed by fedora-packager-0.5.9.4-1.el6.noarch
fedpkg >= 1.0 is needed by fedora-packager-0.5.9.4-1.el6.noarch
koji is needed by fedora-packager-0.5.9.4-1.el6.noarch
mock is needed by fedora-packager-0.5.9.4-1.el6.noarch
redhat-rpm-config is needed by fedora-packager-0.5.9.4-1.el6.noarch
rpmdevtools is needed by fedora-packager-0.5.9.4-1.el6.noarch
rpmlint is needed by fedora-packager-0.5.9.4-1.el6.noarch
ykpers is needed by fedora-packager-0.5.9.4-1.el6.noarch
Oy.
Otherwise known as "RPM Hell", which is why years ago, Fedora and RH moved to 'yum' as the default package manager, since yum handles all of the dependencies... :-) Visit: http://fedoraproject.org/wiki/EPEL for information on getting your system configured to use the EPEL. Once you do that: yum install R2spec as root, should be all that you need...no downloading RPM files locally and dealing with circular dependencies via the CLI. Cheers, Marc Schwartz
On Mon, 2012-03-12 at 16:10 -0400, Allen S. Rout wrote:
Is there somewhere a succinct guide to just how much extra crud we need to install on RHEL to make use of EPEL packages?
All the information are provided there: http://fedoraproject.org/wiki/EPEL And basically, all you have to do is install the epel-release rpm: http://download.fedoraproject.org/pub/epel/6/i386/epel-release-6-5.noarch.rpm Regards, Pierre
On 03/12/2012 04:32 PM, Marc Schwartz wrote:
Otherwise known as "RPM Hell", which is why years ago, Fedora and RH moved to 'yum' as the default package manager, since yum handles all of the dependencies... :-)
Yes; but <grumble> I don't want to install EPEL wholesale; we try to
manage which packages we pull into our locally maintained repos. What
I want to do is make a new set of CRAN-based packages. </grumble>
So, having pulled R2spec 4.1.0-1.el5 into my rhel5.5 system,
-bash-3.2$ R2spec --help
Traceback (most recent call last):
File "/usr/bin/R2spec", line 26, in ?
PARSER = setup_parser('R2spec')
File "/usr/lib/python2.4/site-packages/r2spec/r2spec_obj.py", line
54, in setup_parser
version='%(prog)s ' +VERSION)
File "/usr/lib/python2.4/site-packages/argparse.py", line 1264, in
add_argument
action = action_class(**kwargs)
TypeError: __init__() got an unexpected keyword argument 'version'
- Allen S. Rout
On Mar 12, 2012, at 3:49 PM, Allen S. Rout wrote:
On 03/12/2012 04:32 PM, Marc Schwartz wrote:
Otherwise known as "RPM Hell", which is why years ago, Fedora and RH moved to 'yum' as the default package manager, since yum handles all of the dependencies... :-)
Yes; but <grumble> I don't want to install EPEL wholesale; we try to manage which packages we pull into our locally maintained repos. What I want to do is make a new set of CRAN-based packages. </grumble>
So, having pulled R2spec 4.1.0-1.el5 into my rhel5.5 system,
-bash-3.2$ R2spec --help
Traceback (most recent call last):
File "/usr/bin/R2spec", line 26, in ?
PARSER = setup_parser('R2spec')
File "/usr/lib/python2.4/site-packages/r2spec/r2spec_obj.py", line 54, in setup_parser
version='%(prog)s ' +VERSION)
File "/usr/lib/python2.4/site-packages/argparse.py", line 1264, in add_argument
action = action_class(**kwargs)
TypeError: __init__() got an unexpected keyword argument 'version'
- Allen S. Rout
Allen, You seem to have a special need which is perhaps not satisfied by the default RPM build of the package, given the additional dependencies. You might review: https://fedorahosted.org/r2spec/ for additional installation scenarios. If that is not helpful, I would suggest that you directly contact the R2spec package maintainer for specific assistance. That appears to be Pierre-Yves (pingou at pingoured.fr). Marc
On Mon, 2012-03-12 at 16:49 -0400, Allen S. Rout wrote:
So, having pulled R2spec 4.1.0-1.el5 into my rhel5.5 system,
-bash-3.2$ R2spec --help
Traceback (most recent call last):
File "/usr/bin/R2spec", line 26, in ?
PARSER = setup_parser('R2spec')
File "/usr/lib/python2.4/site-packages/r2spec/r2spec_obj.py", line
54, in setup_parser
version='%(prog)s ' +VERSION)
File "/usr/lib/python2.4/site-packages/argparse.py", line 1264, in
add_argument
action = action_class(**kwargs)
TypeError: __init__() got an unexpected keyword argument 'version'
hm, I am thinking of an older version of python-argparse. Do you have 1.2.1 ? (which is available on EPEL btw). Pierre
On 03/13/2012 02:14 AM, Pierre-Yves Chibon wrote:
hm, I am thinking of an older version of python-argparse. Do you have 1.2.1 ? (which is available on EPEL btw).
Was at 1.0.1; updated to 1.2.1, and got past that error. We'll see what the next hurdle looks like. Would you like a patch for the dependency? - Allen S. Rout
An embedded and charset-unspecified text was scrubbed... Name: not available URL: <https://stat.ethz.ch/pipermail/r-sig-fedora/attachments/20120313/d60508ff/attachment.pl>
On 03/12/2012 07:28 PM, Marc Schwartz wrote:
[...]
So I've updated everything Python from EPEL on my RHEL 5 box. Versions
included below.
# R2spec -p RColorBrewer
Traceback (most recent call last):
File "/usr/bin/R2spec", line 29, in ?
R2spec().main(ARG)
File "/usr/lib/python2.4/site-packages/r2spec/r2spec_obj.py", line
235, in main
pack.search_package_in_repo()
File "/usr/lib/python2.4/site-packages/r2spec/rpackage.py", line 237,
in search_package_in_repo
self.source = '{name}_{version}.tar.gz'.format(name=self.name,
AttributeError: 'str' object has no attribute 'format'
This feels to me like another library drift issue. Aaand stackoverflow
says yeah, that's it.
Pierre-Yves, would you entertain a patch to fix this?
- Allen S. Rout
On Tue, 2012-03-13 at 09:23 -0400, Allen S. Rout wrote:
On 03/13/2012 02:14 AM, Pierre-Yves Chibon wrote:
hm, I am thinking of an older version of python-argparse. Do you
have
1.2.1 ? (which is available on EPEL btw).
Was at 1.0.1; updated to 1.2.1, and got past that error. We'll see what the next hurdle looks like. Would you like a patch for the dependency?
Sure :) Pierre
On Tue, 2012-03-13 at 09:50 -0400, Allen S. Rout wrote:
On 03/12/2012 07:28 PM, Marc Schwartz wrote:
[...]
So I've updated everything Python from EPEL on my RHEL 5 box.
Versions included below.
# R2spec -p RColorBrewer
Traceback (most recent call last):
File "/usr/bin/R2spec", line 29, in ?
R2spec().main(ARG)
File "/usr/lib/python2.4/site-packages/r2spec/r2spec_obj.py", line
235, in main
pack.search_package_in_repo()
File "/usr/lib/python2.4/site-packages/r2spec/rpackage.py", line
237,
in search_package_in_repo
self.source = '{name}_{version}.tar.gz'.format(name=self.name,
AttributeError: 'str' object has no attribute 'format'
This feels to me like another library drift issue. Aaand
stackoverflow says yeah, that's it.
I guess that means replacing all .format to use %
Pierre-Yves, would you entertain a patch to fix this?
That would be nice. I have to say that I do not have a python 2.4 around on which I can run the unit-tests. Pierre
From: Allen S. Rout <asr at ufl.edu> On 03/12/2012 07:28 PM, Marc Schwartz wrote:
? [...]
So I've updated everything Python from EPEL on my RHEL 5 box.?? Versions
included below.
#? R2spec -p RColorBrewer
Traceback (most recent call last):
? File "/usr/bin/R2spec", line 29, in ?
??? R2spec().main(ARG)
? File "/usr/lib/python2.4/site-packages/r2spec/r2spec_obj.py", line 235, in
? main
??? pack.search_package_in_repo()
? File "/usr/lib/python2.4/site-packages/r2spec/rpackage.py", line 237, in
? search_package_in_repo
??? self.source = '{name}_{version}.tar.gz'.format(name=self.name,
AttributeError: 'str' object has no attribute 'format'
This feels to me like another library drift issue.? Aaand stackoverflow says
yeah, that's it.
Pierre-Yves, would you entertain a patch to fix this?
I haven't been following this thread, but I'm pretty sure that the Python string method "format" did not appear until Python 2.6.? It appears that you're using Python 2.4. -- Mike
On 03/13/2012 11:25 AM, Pierre-Yves Chibon wrote:
> On Tue, 2012-03-13 at 09:23 -0400, Allen S. Rout wrote:
>> >> [... patches ...] > > Sure :) OK, here are four of them. One fixes the dependencies I happened to run into. One reverts the .format invocations to '%' Those are what I've already spoken about. Two more are submitted for your consideration. In one, I've added some comments to the specfile template, to make it clear, as you're editing the generated spec, what the R view of the dependencies is, to help you mess around with them after generation time. In the other, I've made the '-p' package name willing to accept both "packagename" and "R-packagename". This is to help the recursive build-dependencies problem, one step closer to (Good lord, I hope) automating this process. - Allen S. Rout -------------- next part -------------- A non-text attachment was scrubbed... Name: 0001-Fix-some-python-and-python-package-deps.patch Type: text/x-patch Size: 896 bytes Desc: not available URL: <https://stat.ethz.ch/pipermail/r-sig-fedora/attachments/20120313/894ce8c0/attachment-0003.bin> -------------- next part -------------- A non-text attachment was scrubbed... Name: 0001-Switch-.format-invocations-to-backward-compatible.patch Type: text/x-patch Size: 1767 bytes Desc: not available URL: <https://stat.ethz.ch/pipermail/r-sig-fedora/attachments/20120313/894ce8c0/attachment-0004.bin> -------------- next part -------------- A non-text attachment was scrubbed... Name: 0001-Document-the-R-perspective-and-our-gaps-in-it-for-th.patch Type: text/x-patch Size: 953 bytes Desc: not available URL: <https://stat.ethz.ch/pipermail/r-sig-fedora/attachments/20120313/894ce8c0/attachment-0005.bin>
On Tue, 2012-03-13 at 17:02 -0400, Allen S. Rout wrote:
One fixes the dependencies I happened to run into.
Applied and pushed
One reverts the .format invocations to '%'
Applied and pushed
Those are what I've already spoken about. Two more are submitted for your consideration. In one, I've added some comments to the specfile template, to make it clear, as you're editing the generated spec, what the R view of the dependencies is, to help you mess around with them after generation time.
Applied and pushed
In the other, I've made the '-p' package name willing to accept both "packagename" and "R-packagename". This is to help the recursive build-dependencies problem, one step closer to (Good lord, I hope) automating this process.
I think this one was missing from the email :) Pierre
On 03/13/2012 05:21 PM, Pierre-Yves Chibon wrote:
On Tue, 2012-03-13 at 17:02 -0400, Allen S. Rout wrote:
In the other, I've made the '-p' package name willing to accept both "packagename" and "R-packagename". This is to help the recursive build-dependencies problem, one step closer to (Good lord, I hope) automating this process.
I think this one was missing from the email :)
Duuh. Trying again. - Allen S. Rout -------------- next part -------------- A non-text attachment was scrubbed... Name: 0001-Accept-package-names-with-leading-R.patch Type: text/x-patch Size: 1269 bytes Desc: not available URL: <https://stat.ethz.ch/pipermail/r-sig-fedora/attachments/20120314/9d0aa652/attachment.bin>
On Wed, 2012-03-14 at 09:54 -0400, Allen S. Rout wrote:
On 03/13/2012 05:21 PM, Pierre-Yves Chibon wrote:
On Tue, 2012-03-13 at 17:02 -0400, Allen S. Rout wrote:
In the other, I've made the '-p' package name willing to accept
both
"packagename" and "R-packagename". This is to help the
recursive
build-dependencies problem, one step closer to (Good lord, I hope) automating this process.
I think this one was missing from the email :)
Duuh. Trying again.
Applied and pushed Thanks for the patches :) Pierre