[poky] Zypper question

[Mark Hatle]

On 5/4/11 6:21 AM, Gary Thomas wrote:
> On 05/03/2011 10:49 PM, Mark Hatle wrote:
>> On 5/3/11 8:45 PM, Gary Thomas wrote:
>>> On 05/03/2011 05:36 PM, Mark Hatle wrote:
>>>> On 5/3/11 6:09 PM, Gary Thomas wrote:
>>>>> I'm trying to set up zypper on a qemu based system.  I've built
>>>>> the image&   additional packages in 192.168.1.125:/local/qemu_test/tmp/deploy/rpm/
>>>>> This system is running lighttpd and has the path http://192.168.1.125/qemu-repo
>>>>> set to that path.  I verified it by fetching http://192.168.1.125/qemu-repo/solvedb.conf
>>>>> which shows up as
>>>>>      /local/qemu_test/tmp/deploy/rpm/qemux86/solvedb
>>>>>      /local/qemu_test/tmp/deploy/rpm/i586/solvedb
>>>>>      /local/qemu_test/tmp/deploy/rpm/all/solvedb
>>>>>
>>>>> On the QEMU system, I ran
>>>>>      # zypper addrepo http://192.168.1.125/qemu-repo yocto
>>>>>      # zypper refresh
>>>>> No errors were reported
>>>>>
>>>>> When I tried to install a package, I got this:
>>>>>      # zypper install gst-plugins-base
>>>>>      Error building the cache.
>>>>>      [1] Repository type can't be determined.
>>>>>      warning: Disabling repository 'yocto' because of the above error.
>>>>>
>>>>> What am I doing wrong?
>>>>>
>>>>> Note: I tried to follow https://wiki.pokylinux.org/wiki/Zypper_Repository_Setup
>>>>> Can I help with this (edit the page, etc)?  If so, how do I get a login?
>>>>>
>>>>
>>>> This is odd.  I didn't experience these issues when I was doing the development
>>>> / testing.
>>>>
>>>>>  From the above there is an "Error building the cache."  the only caching steps
>>>> that occur during an install is a sync of the system's RPM cache.  If something
>>>> is wrong there you can get a similar error.
>>>>
>>>> Try simply running "rpm -qa" and make sure you get reasonable output.  If not,
>>>> then RPM isn't installed properly.  (By default on a minimal system, the RPM
>>>> database is removed after the rootfs is populated.)
>>>
>>> 'rpm -qa ' worked just fine.
>>>
>>> I started with core-image-sato, so it's a pretty complete system to start.
>>>
>>> Do I have the repo stuff set up correctly?  I can see from the lighttpd logs
>>> that it seems to be looking for files I don't have:
>>>     192.168.7.2 192.168.1.125 - [03/May/2011:16:57:38 -0600] "HEAD /qemu-repo/repodata/repomd.xml HTTP/1.1" 404 0 "-" "ZYpp 8.11.0 (curl 7.21.2)"
>>>     192.168.7.2 192.168.1.125 - [03/May/2011:16:57:38 -0600] "HEAD /qemu-repo/content HTTP/1.1" 404 0 "-" "ZYpp 8.11.0 (curl 7.21.2)"
>>> Neither of these files seems to be in tmp/deploy/rpm
>>>
>>> Did I miss a step?  I ran 'bitbake package-index' just in case.
>>>
>>>
>>
>> did you run createrepo on the /local/qemu_test/tmp/deploy/rpm directory?
>>
>> (step 4 in the wiki page)  I usually run createrepo<path>  -- but apparently it
>> works if you are in the same directory...
> 
> Indeed, I missed that step.
> 
> Note: I had expected this operation to be handled by 'bitbake package-index'
> Could it not be done that way so it's always the same, no matter the style
> of package management?

It likely can.  I didn't even know bitbake package-index was a target.  Please
file a bug and I'll investigate and get it resolved.

--Mark

> Thanks
>