launchpad-dev team mailing list archive

[Karl Fogel <karl.fogel@xxxxxxxxxxxxx>]

I've had a multi-day yak-shaving spree with launchpadlib.  Is this list
the right place to discuss launchpadlib?  Let's find out...

Summary: launchpadlib depends on lazr.restful, apparently just for the
tests.  However, this seems to make installs of trunk launchpadlib fail,
as per this transcript: http://paste.ubuntu.com/252186/

My conversation with William Grant about it:

  <wgrant>   kfogel: That's installing faaaar too much stuff. Like
             lazr.restful.
  
  <wgrant>   kfogel: Although if you want to run the tests properly, I
             guess you need that.
  
  <kfogel>   wgrant: I thought launchpadlib depends on lazr.restful?  (was
             told by jml or barry it does)
  
  <wgrant>   kfogel: lazr.restful is the server-side.
  
  <wgrant>   kfogel: And depends on a large portion of the ZTK.
  
  <kfogel>   wgrant: that makes sense.  So why would launchpadlib depend
             on it?
  
  <wgrant>   kfogel: For tests.
  
  * wgrant checks how to make it stop.
  
  <wgrant>   kfogel: ARGH WTF
  
  <wgrant>   lazr.restfulclient specifies a runtime dependency on
             lazr.restful
  
  <wgrant>   == insane
  
  <kfogel>   wgrant: that does indeed sound like insanity.
  
  <wgrant>   And as it's all inside stupid eggs, it's probably easier to
             make lazr.lifecycle work properly.
  
  <wgrant>   And I'm not sure what's wrong there.
  
  <kfogel>   wgrant: I'm going to post about this launchpadlib problem on
             launchpad-dev@.
  
  <wgrant>   kfogel: Good idea.

I *think* there is a bug here, but wanted more understanding before
filing.  Hence this post.  Thoughts?

-Karl