Re: [Zope-dev] dependency cleanup progress report

2009-02-05 Thread Tres Seaver
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Tres Seaver wrote:
 Martijn Faassen wrote:
 Hi there,
 
 After a lot of work we have progress to report on the dependency 
 reduction front:
 
 http://faassen.n--tree.net/blog/view/weblog/2009/01/29/0
 
 It's been a lot of work to get this far and there's a huge amount of 
 work to be done still, but there is progress!
 
 The second dependency graph is against the trunks of all the packages, 
 as we haven't done the releases yet to make this real. We will look into 
 this tomorrow. We also need to document the various procedures we have 
 been using to do this work so that others can jump in and help us.
 
 Excellent work!
 
 Looking at the zope.container graph:
 
 - The zope.publisher dependency is purely there for the 'traverser.py'
   module, which probably doesn't belong in zope.container at all:  it
   only imports interfaces to set up __adapts__, which suggests that it
   should be in zope.publisher, or some other package which depends on
   both (losing the __adapts__ wouldn't be terrible, for instance).
 
 - The dependency module pulls in zope.app.dependable:  I think it would
   make more sense to move the event subscriber *into* that package.
   which would remove the other, indirect dependency on zope.traversing.
 
 - The 'testing' module should lose all the zope-specific module-scope
   imports, even  if that means losing some convenience.

I have implemented this one by creating a new 'zope.broken' package, and
making 'zope.container' and 'zope.app.broken' depend on it:

 - The 'contained' module has a bunch of dependencies, including a
   suspicous one on zope.app.broken:  if the IBroken check is really
   needed, then that interface should be moved to another package.

and this one:

 - The zope.exceptions dependency, again from contained, is due to
   the use of a couple of dubious zope-specific exceptions:
 
   o UserError should probably be replaced with the standard ValueError.
 
   o DuplicationError should probably be replaced with KeyError.

now released in zope.container 3.7.1 and zope.app.container 3.7.1.

I also updated the 'zope.app.zcmlfiles' package (now version 3.5.3),
which was not including the 'zope.app.broken' package in its
dependencies, although it pulls in its ZCML.



Tres.
- --
===
Tres Seaver  +1 540-429-0999  tsea...@palladion.com
Palladion Software   Excellence by Designhttp://palladion.com
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFJiy01+gerLs4ltQ4RAhfHAJ9bhVfaH+Jw/4qEE/c1Yg3ycCl/SQCfbykt
InDS4RvBO5qlJ6JPioIhE8I=
=pQ0n
-END PGP SIGNATURE-
___
Zope-Dev maillist  -  Zope-Dev@zope.org
http://mail.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://mail.zope.org/mailman/listinfo/zope-announce
 http://mail.zope.org/mailman/listinfo/zope )


Re: [Zope-dev] dependency cleanup progress report

2009-01-30 Thread Martijn Faassen
Hey,

Thanks for this feedback. It's a good sign that such feedback is now 
possible due to us having a saner dependency graph!

We won't be taking action on these issues during this sprint as we're 
trying to wrap it up and make releases of our changes. At least the 
dependency graph for zope.container has no more cycles (they're still 
abundant elsewhere but we have a better grip on them now at least).

hat shouldn't stop anyone from trying out Tres's suggestions, which all 
look like good ideas.

When you try these things it's a good idea to test the trunks of various 
packages against each other, so do consider using z3c.recipe.compattest 
with use_trunk mode to make sure your changes don't break anything.

Regards,

Martijn

___
Zope-Dev maillist  -  Zope-Dev@zope.org
http://mail.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://mail.zope.org/mailman/listinfo/zope-announce
 http://mail.zope.org/mailman/listinfo/zope )


Re: [Zope-dev] dependency cleanup progress report

2009-01-30 Thread Jan-Wijbrand Kolman
On Fri, Jan 30, 2009 at 3:47 PM, Martijn Faassen faas...@startifact.com wrote:
(snip)
 When you try these things it's a good idea to test the trunks of various
 packages against each other, so do consider using z3c.recipe.compattest
 with use_trunk mode to make sure your changes don't break anything.

That should be the use_svn = true option.

regards,
jw
-- 
Jan-Wijbrand Kolman
___
Zope-Dev maillist  -  Zope-Dev@zope.org
http://mail.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://mail.zope.org/mailman/listinfo/zope-announce
 http://mail.zope.org/mailman/listinfo/zope )


Re: [Zope-dev] dependency cleanup progress report

2009-01-29 Thread Tres Seaver
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Martijn Faassen wrote:
 Hi there,
 
 After a lot of work we have progress to report on the dependency 
 reduction front:
 
 http://faassen.n--tree.net/blog/view/weblog/2009/01/29/0
 
 It's been a lot of work to get this far and there's a huge amount of 
 work to be done still, but there is progress!
 
 The second dependency graph is against the trunks of all the packages, 
 as we haven't done the releases yet to make this real. We will look into 
 this tomorrow. We also need to document the various procedures we have 
 been using to do this work so that others can jump in and help us.

Excellent work!

Looking at the zope.container graph:

- - The zope.publisher dependency is purely there for the 'traverser.py'
  module, which probably doesn't belong in zope.container at all:  it
  only imports interfaces to set up __adapts__, which suggests that it
  should be in zope.publisher, or some other package which depends on
  both (losing the __adapts__ wouldn't be terrible, for instance).

- - The dependency module pulls in zope.app.dependable:  I think it would
  make more sense to move the event subscriber *into* that package.
  which would remove the other, indirect dependency on zope.traversing.

- - The 'testing' module should lose all the zope-specific module-scope
  imports, even  if that means losing some convenience.

- - The 'contained' module has a bunch of dependencies, including a
  suspicous one on zope.app.broken:  if the IBroken check is really
  needed, then that interface should be moved to another package.

- - The zope.exceptions dependency, again from contained, is due to
  the use of a couple of dubious zope-specific exceptions:

  o UserError should probably be replaced with the standard ValueError.

  o DuplicationError should probably be replaced with KeyError.



Tres.
- --
===
Tres Seaver  +1 540-429-0999  tsea...@palladion.com
Palladion Software   Excellence by Designhttp://palladion.com
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFJgjDu+gerLs4ltQ4RAtv4AJ902n6ZnbpoA3patMcW8H/93cAYeACeNoeR
3XSe5cO+WIcCHz2o/TtDZkc=
=6kKl
-END PGP SIGNATURE-

___
Zope-Dev maillist  -  Zope-Dev@zope.org
http://mail.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://mail.zope.org/mailman/listinfo/zope-announce
 http://mail.zope.org/mailman/listinfo/zope )