[ET Trac] [Einstein Toolkit] #241: Thorns should be able to specify implementation versions, and other thorns should be able to depend on those

[Einstein Toolkit]

#241: Thorns should be able to specify implementation versions, and other thorns
should be able to depend on those
--------------------------+-------------------------------------------------
  Reporter:  knarf        |       Owner:            
      Type:  enhancement  |      Status:  review    
  Priority:  minor        |   Milestone:  ET_2016_05
 Component:  Cactus       |     Version:            
Resolution:               |    Keywords:            
--------------------------+-------------------------------------------------

Comment (by knarf):

 Replying to [comment:17 rhaas]:
 > I am still somewhat unhappy with the non-explicit loop termination that
 always exists the loop through an return in the loop body. I added a
 version that (I personally, others may well differ) find more explicit and
 intuitive (at least the loop termination condition, if one want to have
 just shorter code then one can get rid of four more lines by assigning the
 result of the regex to a list).

 Your versions looks fine to me, too. Nicer, in fact.

 > As far as whether I prefer that version numbers should be required to
 start with a numerical or a non-numerical prefix, I'd say "neither should
 be required" ie. both OpenMPI's {{{v2.x-dev-411-gb0ff2a1}}} as well as
 {{{1.8.15-patch1}}} should be supported (and I think CompareVersion
 supports all of these).

 We could allow any start, but (using your example) an explicit 'v' in
 front of the version
 only begs for misunderstandings (is the version 2.x, or v2.x?).

 > Minor issue: I would rename the CheckVersionStrings procedure to just
 CheckVersions to avoid confusion whether the procedure checks versions of
 capabilities or verifies that a version string is valid.

 What about CheckVersionFulfilled?

 > Documentation and comments need to be made consistent with what the code
 actually does before this is applied but otherwise I have no objection to
 this. Is there also support for CONFLICTS?

 Sure.

 About the Perl-idioms: Some people like them, some don't. I find they
 interrupt my code-reading whenever I encounter them. But whatever.

-- 
Ticket URL: <https://trac.einsteintoolkit.org/ticket/241#comment:20>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit