zopectl's -p options
...is currently useless, as far as I can see. This stops you specifying a different location for runzope, which is handy if you're doing buildout-based stuff that uses a deployment. I think the attached patch fixes this, but I'm wondering: - are these are tests for this area - why does self.program need to be a one element list? cheers, Chris =================================================================== --- src/Zope2/Startup/zopectl.py (revision 115104) +++ src/Zope2/Startup/zopectl.py (working copy) @@ -114,11 +114,14 @@ positional_args_allowed = 1 - program = "zopectl" schemadir = os.path.dirname(Zope2.Startup.__file__) schemafile = "zopeschema.xml" uid = gid = None + # this indicates that no explict program has been provided. + # the command line option can set this. + program = None + # XXX Suppress using Zope's <eventlog> section to avoid using the # same logging for zdctl as for the Zope appserver. There still # needs to be a way to set a logfile for zdctl. @@ -169,7 +172,9 @@ config = self.configroot self.directory = config.instancehome self.clienthome = config.clienthome - if config.runner and config.runner.program: + if self.program: + self.program = [self.program] + elif config.runner and config.runner.program: self.program = config.runner.program else: self.program = [os.path.join(self.directory, "bin", "runzope")] -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk
On Mon, Jul 26, 2010 at 7:36 PM, Chris Withers <chris@simplistix.co.uk> wrote:
...is currently useless, as far as I can see.
I thought the -p option was an artifact of zdaemon, which zopectl just didn't support. I took it as a case of the lower level API's leaking through.
I think the attached patch fixes this, but I'm wondering:
- are these are tests for this area
I'm not aware of any.
- why does self.program need to be a one element list?
No idea. I find most of the zdaemon and Zope2 startup code rather obscure. Hanno
Hanno Schlichting wrote:
On Mon, Jul 26, 2010 at 7:36 PM, Chris Withers <chris@simplistix.co.uk> wrote:
...is currently useless, as far as I can see.
I thought the -p option was an artifact of zdaemon, which zopectl just didn't support. I took it as a case of the lower level API's leaking through.
I think it's an extremely useful feature ;-) (likely broken by someone who didn't understand it, some time ago...)
I think the attached patch fixes this, but I'm wondering:
- are these are tests for this area
I'm not aware of any.
OK, I guess that makes life easier in some ways ;-)
- why does self.program need to be a one element list?
No idea. I find most of the zdaemon and Zope2 startup code rather obscure.
Indeed. Well, unless anyone objects, I'd like to commit the patch I proposed... cheers, Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk
Chris Withers wrote:
Hanno Schlichting wrote:
On Mon, Jul 26, 2010 at 7:36 PM, Chris Withers <chris@simplistix.co.uk> wrote:
...is currently useless, as far as I can see. I thought the -p option was an artifact of zdaemon, which zopectl just didn't support. I took it as a case of the lower level API's leaking through.
I think it's an extremely useful feature ;-) (likely broken by someone who didn't understand it, some time ago...)
I think the attached patch fixes this, but I'm wondering:
- are these are tests for this area I'm not aware of any.
OK, I guess that makes life easier in some ways ;-)
- why does self.program need to be a one element list? No idea. I find most of the zdaemon and Zope2 startup code rather obscure.
Indeed. Well, unless anyone objects, I'd like to commit the patch I proposed...
OK, if there are no comments, I'll commit some time soon, likely on Monday. cheers, Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk
participants (2)
-
Chris Withers -
Hanno Schlichting