incubator-bloodhound-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gary Martin <gary.mar...@wandisco.com>
Subject Re: First attempt to setup CI job for Bloodhound installer
Date Thu, 12 Jul 2012 09:08:39 GMT
Actually, it sort of looks like the commands might have been running in 
what is effectively a virtualenv anyway. I am not sure I can comment 
properly on what is going on there.

I suspect that the source command fails because there is no binary 
associated with it. You might find that replacing

    system('source %(activate)s' % ENVIRONMENT)

with

    system('. %(activate)s' % ENVIRONMENT)

might work better, although I would not be certain that it has an effect 
that lasts into the commands that follow.

Given that things appear to pip install despite the source command 
failing, it looks like the ImportError exception should be taken at face 
value.. That trac is not actually installed. And closer at the output of 
the pip install command, it fails with:

    Cannot find command 'svn'

Before it can actually starts to install the packages. At the moment 
TracAccountManager, TracThemeEngine and TracPermRedirect are all 
specified to be pulled from the trac-hacks svn server. We would have to 
check if we can get appropriate versions elsewhere if we want to change 
this.

Cheers,
     Gary

On 07/12/2012 09:00 AM, Gary Martin wrote:
> Hi,
>
> I think that to run through virtualenv you would have to resort to the kind of trickery
that I was using in the old installer.py but I am not convinced that is the best plan.
>
> If you have a throw-away VM as your slave then you could ignore the virtualenv part and
pip install to the system instead.
>
> Cheers,
>      Gary
>
> Olemis Lang <olemis@gmail.com> wrote:
>
>> Hi !
>>
>> I've been playing with Jenkins for a while . I created a job [1]_ .
>> Its execution is restricted to a slave running Ubuntu (i.e.
>> kumar-ubuntu node) , but that may change later e.g. to test it in a
>> windows slave . Build job has a single step executing a python script
>> [2]_ shown below
>>
>> {{{
>> #!python
>>
> >from os import system
> >from os.path import join, abspath
>> import sys
>>
>> scripts_folder = 'Scripts' if sys.platform == 'win32' else 'bin'
>> vcs_root = '.'
>>
>> folder_ws = abspath('.')
>> folder_installer = join(folder_ws, vcs_root, 'installer')
>>
>> venv = join(folder_installer, 'bloodhound')
>> bin = join(venv, scripts_folder)
>> activate = join(bin, 'activate')
>> easy_install = join(bin, 'easy_install')
>> pip = join(bin, 'pip')
>> python = join(bin, 'python')
>>
>> ENVIRONMENT = {
>>         'folder_ws' : folder_ws,
>>         'folder_installer' : folder_installer,
>>         'venv' : venv,
>>         'bin' : bin,
>>         'activate' : activate,
>>         'easy_install' : easy_install,
>>         'pip' : pip,
>>         'python' : python,
>>     }
>>
>> system('wget http://peak.telecommunity.com/dist/virtual-python.py')
>> system('wget http://peak.telecommunity.com/dist/ez_setup.py')
>>
>> #system('python2.6 ./virtual-python.py  --no-site-packages
>> --install-dir=%(venv)s' % ENVIRONMENT)
>> system('virtualenv --no-site-packages %(venv)s' % ENVIRONMENT)
>>
>> system('%(python)s ./ez_setup.py' % ENVIRONMENT)
>> system('%(easy_install)s pip' % ENVIRONMENT)
>>
>> system('cd %(folder_installer)s' % ENVIRONMENT)
>> if sys.platform != 'win32':
>>     system('source %(activate)s' % ENVIRONMENT)
>> system('%(pip)s install -r requirements-dev.txt' % ENVIRONMENT)
>> system('''%(python)s bloodhound_setup.py
>>             -d sqlite --database-string="sqlite:db/trac.db"
>>             --admin-password="testrun" --admin-user="bhadmin"
>>             --repository-type=svn --repository-path=
>>             ''' % ENVIRONMENT)
>>
>> }}}
>>
>> I've been trying to make it check out code from svn and execute the
>> installer in order to run tests (... or coverage , or ...) later, in
>> theory, using makefile . Nonetheless this is what I notice in console
>> output [3]_ regarding the installation process .
>>
>>   - The command `source ./bloodhound/bin/activate` , mentioned in
>>     installation instructions , fails with message
>>     `sh: source: not found` ... so I guess we should have
>>     a more generic way to do this ? ... or fix the script ;)
>>   - Maybe that's the reason why I get this message immediately after ?
>>
>> {{{
>> #!python
>>
>> Traceback (most recent call last):
>>   File "bloodhound_setup.py", line 34, in <module>
>>     from trac.admin.console import TracAdmin
>> ImportError: No module named trac.admin.console
>> Post install setup requires that Bloodhound is properly installed
>> Traceback for error follows:
>> }}}
>>
>> I look forward to your comments .
>>
>> .. [1] Run Trac test suite in Apache(TM) Blodhound repository
>>         (http://hudson.testrun.org/job/bh-trac-test/)
>>
>> .. [2] Configure bh-trac-test
>>         (http://hudson.testrun.org/job/bh-trac-test/configure)
>>
>> .. [3] Build results (#11)
>>         (http://hudson.testrun.org/job/bh-trac-test/11/console)
>>
>> -- 
>> Regards,
>>
>> Olemis.
>>
>> Blog ES: http://simelo-es.blogspot.com/
>> Blog EN: http://simelo-en.blogspot.com/
>>
>> Featured article:



Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message