Oops forgot to cc the list. On Nov 18, 2014 7:26 AM, "John Chilton" jmchilton@gmail.com wrote:
I will admit to not actually understanding Galaxy's dependency management but I think virtualenv is exactly the advice people who do understand it give http://dev.list.galaxyproject.org/Local-installation-problem-td4662627.html. It is a widely used tool precisely designed to solve such problems - I think it is the best way to go. I don't know why it would not be appropriate for existing installations - I think it is in fact somethimg of a "best practice" for existing installations.
Certainly that error message should be more helpful but I am not sure we should do anything to address this beyond that - do you have a particular idea in mind? On Nov 18, 2014 5:49 AM, "Peter Cock" p.j.a.cock@googlemail.com wrote:
On Mon, Nov 17, 2014 at 7:23 PM, John Chilton jmchilton@gmail.com wrote:
Ummm... I think it is that the VM started shipping with an incompatible paramkio.
[...]
Anyway - planemo's TravisCI integration tests Galaxy in a virtualenv and it works fine [...]
That's a useful workaround, but a virtualenv is not viable for existing Galaxy installations - which may also come to have this problem if their system copy of paramiko is updated?
I'm not familiar with paramiko or markupsafe, but if there is a conflict it would be good to have a direct fix.
Peter
On Nov 18, 2014 7:26 AM, "John Chilton" jmchilton@gmail.com wrote:
I will admit to not actually understanding Galaxy's dependency management but I think virtualenv is exactly the advice people who do understand it give http://dev.list.galaxyproject.org/Local-installation-problem-td4662627.html. It is a widely used tool precisely designed to solve such problems - I think it is the best way to go. I don't know why it would not be appropriate for existing installations - I think it is in fact somethimg of a "best practice" for existing installations.
Our existing installation is not using a virtual env, and I fear switching to that could be disruptive.
Certainly that error message should be more helpful but I am not sure we should do anything to address this beyond that - do you have a particular idea in mind?
Not show the IndexError exception? :P
Here the user should be told something about a conflict between MarkupSafe and paramiko (assuming this is the real problem).
Peter
On Tue, Nov 18, 2014 at 1:24 PM, Peter Cock p.j.a.cock@googlemail.com wrote:
On Nov 18, 2014 7:26 AM, "John Chilton" jmchilton@gmail.com wrote:
I will admit to not actually understanding Galaxy's dependency management but I think virtualenv is exactly the advice people who do understand it give http://dev.list.galaxyproject.org/Local-installation-problem-td4662627.html. It is a widely used tool precisely designed to solve such problems - I think it is the best way to go. I don't know why it would not be appropriate for existing installations - I think it is in fact somethimg of a "best practice" for existing installations.
Our existing installation is not using a virtual env, and I fear switching to that could be disruptive.
Getting back to TravisCI, using a virtual env wasn't too painful: https://github.com/peterjc/pico_galaxy/commit/26489a65a9cd60f9d055488d003346...
I can now get back to tweaking the tests I was working on :)
Thanks,
Peter
Peter,
Unless you've made modifications to Galaxy that depend on external libraries, switching to a virtualenv for the server itself should be pretty safe. Tools themselves can still run without using the/any virtualenv, if desired.
--nate
On Tue, Nov 18, 2014 at 8:24 AM, Peter Cock p.j.a.cock@googlemail.com wrote:
On Nov 18, 2014 7:26 AM, "John Chilton" jmchilton@gmail.com wrote:
I will admit to not actually understanding Galaxy's dependency management but I think virtualenv is exactly the advice people who do understand it give http://dev.list.galaxyproject.org/Local-installation-problem-td4662627.html. It is a widely used tool precisely designed to solve such problems - I think it is the best way to go. I don't know why it would not be appropriate for existing installations - I think it is in fact somethimg of a "best practice" for existing installations.
Our existing installation is not using a virtual env, and I fear switching to that could be disruptive.
Certainly that error message should be more helpful but I am not sure we should do anything to address this beyond that - do you have a particular idea in mind?
Not show the IndexError exception? :P
Here the user should be told something about a conflict between MarkupSafe and paramiko (assuming this is the real problem).
Peter ___________________________________________________________ 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/
To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
On Tue, Nov 18, 2014 at 5:51 PM, Nate Coraor nate@bx.psu.edu wrote:
Peter,
Unless you've made modifications to Galaxy that depend on external libraries, switching to a virtualenv for the server itself should be pretty safe. Tools themselves can still run without using the/any virtualenv, if desired.
--nate
OK - that sounds more straightforward than I had feared - but I will cross that bridge as needed ;)
Thanks,
Peter
galaxy-dev@lists.galaxyproject.org
-
John Chilton -
Nate Coraor -
Peter Cock