krfkeith / metacello

Automatically exported from code.google.com/p/metacello
0 stars 0 forks source link

Interesting behavior in presence of duplicate pacakge names #140

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago
When ConfigurationOfZincHTTPComponents 1.0 (PaulDebruiker.14) is loaded into 
GLASS 1.0-beta.8.7 (DaleHenrichs.180) with Metacello 1.0-beta.29 (dkh.612). The 
package is loaded from SqueakSource instead of GemSource.

The configuration declares both repositories, so it is "correct" that the 
package could be loaded from either repository. 

However, there are two unique packages:

  GemSource:

    UUID: d3f596f3-2eae-4255-ae7c-d1952a2b35fe

  SqueakSource:

    UUID: 9ac457ad-7824-4c0c-8d5d-e7ebe36f0280

With two packages with the same name, "you get what you pay for" and for the 
record, the repositories can be disambiguated by using the following 
specification for Zinc-HTTP in the #gemstone section:

  package: 'Zinc-HTTP' with: [
    spec 
      repository: 'http://seaside.gemstone.com/ss/ZincHTTPComponents';
      requires: #('Zinc-Gemstone-Preliminary')  ]

In the end, though, the "interesting behavior" and reason for the bug report is 
that the Monticello repository reports that the package was loaded from the 
GemSource repository, when the actual source was the SqueakSource repository ...

Now, I thought that we were using GoferResolvedReference to determine which 
package to load and I thought that we set the repository for the package based 
on the GoferResolvedReference, but this result appears to contradict that ... 
need to dig into this one a little deeper ...

Original issue reported on code.google.com by henrichs...@gmail.com on 23 May 2011 at 6:46

GoogleCodeExporter commented 9 years ago

Original comment by henrichs...@gmail.com on 18 Jan 2012 at 6:26