[Orbit-python-list] [Jochen Voss <jvoss2@web.de>] Bug#123848: python-orbit: mapping for union types is incompatible with the specification

[Roland M. <99....@ai...>]

  Hello,

  I got the following bug report from a Debian user (bug #123848,
available from <URI:http://bugs.debian.org/123848>).

,----
| Subject: Bug#123848: python-orbit: mapping for union types is incompatible with the specification
| From: Jochen Voss <jv...@we...>
| 
| Package: python-orbit
| Version: 0.3.1-3
| Severity: normal
| 
| Hello,
| 
| recently I tried to use CORBA´s discriminated union types
| with python-orbit and had several problems with this.
| The most serious one is the following.
| 
| The "Python Language Mapping Specification" (OMG document 01-02-66)
| says on page 1-6:
| 
|     Union types are mapped to classes with two attributes.
|     The first is the discriminant _d, the second the associated value _v.
|     For each branch, there is an additional attribute, which
|     can only be accessed if the branch has been set. [...]
| 
| Python-orbit does not follow this specification.
| It uses "d" and "v" (without the underscore) as the names for
| discriminant and value.  And the "additional attribute" is
| simple not there.  Consequently all examples from page 1-7
| of the specification (and my own program, too) do not work
| with python-orbit.
| 
| I see that it may be some work to implement the
| additional-attribute-stuff, but the the names for discriminant
| and value really should be corrected.
| 
| Jochen
`----

  For what it's worth: the version mentioned there (0.3.1-3) is in
fact not the vanilla 0.3.1 release, but 0.3.1 + the patch to fix the
marshalling of booleans (which used to lead to a segmentation fault).

  I might start hacking on this one on and off, but if someone wants
to be faster than I, please do.

Roland.
-- 
Roland Mas

Qu'est-ce qui est jaune, qui pèse deux cents kilos et qui chante ?
Un sumotori dans sa salle de bains.