With Dannon's help, I was able to straighten out our local installation by 
doing the following:

Update mercurial to new version (2.4.2) for hg to work as expected.

Commit local changes:
  hg diff -p > local_changes.patch  (save local changes as a patch, in case 
needed)
  hg ci -m "Local changes to cleanup_datasets and various tools"  (commit 
changes)

hg update   (no files unresolved)

hg tip  (shows latest update)

hg heads  (shows 2 heads so need to merge)

hg merge

hg ci -m "Merged changes"  (commit merged changes)

hg head  (only 1 head now)

Restarted server, ran the needed database migration script and option tool 
migration script.

The instance is now successfully updated, and this fixed the original problem 
of not being able to install a tool from the tool shed.



Mike Waldron
Systems Specialist
ITS - Research Computing Center
University of North Carolina at Chapel Hill


________________________________________
From: Dannon Baker [dannonba...@me.com]
Sent: Tuesday, January 29, 2013 2:04 PM
To: Waldron, Michael H
Cc: Paul Boddie; James Taylor; Greg Von Kuster; galaxy-...@bx.psu.edu Dev
Subject: Re: [galaxy-dev] Error trying to add tools from tool shed respository

Ok, so that confirms that you definitely have local customizations we'll want 
to make sure to keep.  What I'd do before we go any further is this:

`hg diff -p > local_changes.patch`

This will save all of your current changes to a patch, should we have to 
recreate them (though I don't expect this).

`hg ci -m "Local changes to cleanup_datasets and various tools"`

Which will commit your local changes to your local repository.  At this point, 
hopefully, you can just 'hg update', possibly needing to merge as well.

For IRC:  Just join the "Freenode" network shown in XChat, and then Join 
Channel should pop up a box you can put #galaxyproject in



On Jan 29, 2013, at 1:52 PM, "Waldron, Michael H" <mwald...@email.unc.edu> 
wrote:

> Here is hg st output:
>
> $ hg st
> M scripts/cleanup_datasets/delete_datasets.sh
> M scripts/cleanup_datasets/delete_datasets_main.sh
> M scripts/cleanup_datasets/delete_userless_histories.sh
> M scripts/cleanup_datasets/delete_userless_histories_main.sh
> M scripts/cleanup_datasets/purge_datasets.sh
> M scripts/cleanup_datasets/purge_datasets_main.sh
> M scripts/cleanup_datasets/purge_folders.sh
> M scripts/cleanup_datasets/purge_folders_main.sh
> M scripts/cleanup_datasets/purge_histories.sh
> M scripts/cleanup_datasets/purge_histories_main.sh
> M scripts/cleanup_datasets/purge_libraries.sh
> M scripts/cleanup_datasets/purge_libraries_main.sh
> M scripts/cleanup_datasets/update_metadata.sh
> M tools/peak_calling/macs_wrapper.xml
> M tools/plotting/boxplot.xml
> M tools/sr_mapping/bfast_wrapper.xml
> M tools/sr_mapping/bowtie2_wrapper.xml
> M tools/sr_mapping/bowtie_color_wrapper.xml
> M tools/sr_mapping/bowtie_wrapper.xml
> M tools/sr_mapping/bwa_color_wrapper.xml
> M tools/sr_mapping/bwa_wrapper.xml
> ? contrib/galaxy-init
> ? galaxy_cleanup.log
> ? galaxy_cleanup.sh
> ? migrated_tools_conf.xml.110712
> ? paster.pid
> ? scripts/cleanup_datasets/delete_userless_histories.log
> ? scripts/cleanup_datasets/purge_datasets.log
> ? scripts/cleanup_datasets/purge_folders.log
> ? scripts/cleanup_datasets/purge_histories.log
> ? scripts/cleanup_datasets/purge_libraries.log
> ? static/UNC_galaxy_mirror_housekeeping_notes.pdf
> ? tool-data/bwa_index.loc.sample
> ? tool-data/bwa_index_color.loc.sample
> ? tool-data/tool_data_table_conf.xml.sample
> ? tool_conf.xml.110212
> ? tool_conf.xml.new
> ? tool_data_table_conf.xml.030112
> ? tools/plotting/boxplot.xml.012512
> ? typescript
>
> The hg diff output is quite long.
>
> It turns out I didn't connect to the IRC channel. I have XChat IRC client, 
> but not familiar with using it. How to I connect to the irc.freenode.net 
> server to access the #galaxyproject channel?
>
> Mike Waldron
> Systems Specialist
> ITS - Research Computing Center
> University of North Carolina at Chapel Hill
>
>
> ________________________________________
> From: Dannon Baker [dannonba...@me.com]
> Sent: Tuesday, January 29, 2013 1:48 PM
> To: Waldron, Michael H
> Cc: Paul Boddie; James Taylor; Greg Von Kuster; galaxy-...@bx.psu.edu Dev
> Subject: Re: [galaxy-dev] Error trying to add tools from tool shed respository
>
> I don't know why I thought we'd already checked this, but --
>
> 'hg st' will tell you which files have been modified in your local 
> repository, and
> 'hg diff' will tell you exactly what changes have taken place
>
> I'm not sure of the exact state of your repository at the moment, so this 
> *might* display all changes from the pull as well, which wouldn't be 
> particularly useful.  Worth trying, though.
>
> Don't see you in #galaxyproject yet, sure it's working?
>
>
> On Jan 29, 2013, at 1:42 PM, "Waldron, Michael H" <mwald...@email.unc.edu> 
> wrote:
>
>> The hg update --check command reports the following:
>>
>> $ hg update --check
>> abort: uncommitted local changes
>>
>> I just connected to the #galaxyproject IRC channel.
>>
>>
>> Mike Waldron
>> Systems Specialist
>> ITS - Research Computing Center
>> University of North Carolina at Chapel Hill
>>
>>
>> ________________________________________
>> From: Dannon Baker [dannonba...@me.com]
>> Sent: Tuesday, January 29, 2013 1:37 PM
>> To: Waldron, Michael H
>> Cc: Paul Boddie; James Taylor; Greg Von Kuster; galaxy-...@bx.psu.edu
>> Subject: Re: [galaxy-dev] Error trying to add tools from tool shed 
>> respository
>>
>> So that's interesting.  My best guess was that this was caused by swapping 
>> from galaxy-dist to galaxy-central in the situation where galaxy-dist has 
>> had changesets applied directly and then ported to galaxy-central out of 
>> order, combined with your oder version of mercurial.  That parent, however, 
>> would have been before that happened.
>>
>> Anyway, at this point, does `hg update --check` actually work?
>>
>> And, lastly, If you want -- hop on IRC (irc.freenode.net #galaxyproject, 
>> http://wiki.galaxyproject.org/Get%20Involved#IRC_Channel) and it might be 
>> faster to iterate through what we need to do to get your repository back in 
>> shape.
>>
>>
>> On Jan 29, 2013, at 1:32 PM, "Waldron, Michael H" <mwald...@email.unc.edu> 
>> wrote:
>>
>>> Forgot to also include the hg parents output.
>>>
>>> $ hg parents
>>> changeset:   7986:12fcd068b12e
>>> user:        Daniel Blankenberg <d...@bx.psu.edu>
>>> date:        Thu Oct 18 11:22:12 2012 -0400
>>> summary:     Do not hide failed datasets with HideDatasetAction post job 
>>> action.
>>>
>>>
>>> Mike Waldron
>>> Systems Specialist
>>> ITS - Research Computing Center
>>> University of North Carolina at Chapel Hill
>>>
>>>
>>> ________________________________________
>>> From: Waldron, Michael H
>>> Sent: Tuesday, January 29, 2013 1:28 PM
>>> To: Paul Boddie; James Taylor
>>> Cc: Greg Von Kuster; galaxy-...@bx.psu.edu
>>> Subject: RE: [galaxy-dev] Error trying to add tools from tool shed 
>>> respository
>>>
>>> As it turns out, it was fairly simple to install a newer hg version (2.4.2).
>>>
>>> So now running hg merger --preview and hg heads returns:
>>>
>>> $ hg merge --preview
>>> abort: branch 'default' has one head - please merge with an explicit rev
>>> (run 'hg heads' to see all heads)
>>> [galaxy@galaxy galaxy-dist]$ hg heads
>>> changeset:   8677:22788c1262a2
>>> tag:         tip
>>> user:        jeremy goecks <jeremy.goe...@emory.edu>
>>> date:        Mon Jan 28 16:36:02 2013 -0500
>>> summary:     Language improvements for tophat wrappers.
>>>
>>> Mike Waldron
>>> Systems Specialist
>>> ITS - Research Computing Center
>>> University of North Carolina at Chapel Hill
>>>
>>>
>>> ________________________________________
>>> From: Paul Boddie [paul.bod...@biotek.uio.no]
>>> Sent: Tuesday, January 29, 2013 12:01 PM
>>> To: James Taylor
>>> Cc: Waldron, Michael H; Greg Von Kuster; galaxy-...@bx.psu.edu
>>> Subject: Re: [galaxy-dev] Error trying to add tools from tool shed 
>>> respository
>>>
>>> On 29/01/13 17:33, James Taylor wrote:
>>>> Wasn't sure of the order, those are definitely after running the
>>>> update. Dannon's advice makes a good point, upgrading to the latest
>>>> mercurial should help.
>>>
>>> If upgrading Mercurial will involve introducing another detour, you
>>> could potentially check which version of the Galaxy code you're using
>>> now as defined by the "parent" revisions:
>>>
>>> hg parents
>>>
>>> This (or these in certain cases) might not be a repository "head" and
>>> thus won't appear in "hg heads". If you have the graphlog extension
>>> enabled, running the following might be informative as it should show
>>> you where you are (with a "@" symbol) in the code history:
>>>
>>> hg glog | more
>>>
>>> If you don't want to risk your working directory, you could first clone
>>> it to make something that you can then experiment with:
>>>
>>> hg clone . ../galaxy-dist-test
>>>
>>> Then, you'd move into this clone:
>>>
>>> cd ../galaxy-dist-test
>>>
>>> You might need to update to the version previously printed by "hg
>>> parents" above:
>>>
>>> hg update <revision>
>>>
>>> One might then assume that any desirable update would involve getting to
>>> the "tip", and so you could try a merge with that:
>>>
>>> hg merge tip
>>>
>>> If all of this works reasonably, you should be able to do the same merge
>>> on your real working directory (and can thus delete the test directory).
>>> The result of any merge will need to be committed:
>>>
>>> hg commit
>>>
>>> Anyway, these are just some thoughts. The crucial point is that you can
>>> clone the repository and then work with the clone, discarding it if your
>>> experiments go wrong. Doing this might be quicker than having to get a
>>> more recent Mercurial version installed.
>>>
>>> Paul
>>>
>>> ___________________________________________________________
>>> Please keep all replies on the list by using "reply all"
>>> in your mail client.  To manage your subscriptions to this
>>> and other Galaxy lists, please use the interface at:
>>>
>>> http://lists.bx.psu.edu/
>>
>


___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:

  http://lists.bx.psu.edu/

Reply via email to