Opened 15 years ago

Closed 15 years ago

Last modified 15 years ago

#17475 closed defect (fixed)

swig python variant is based on python26

Reported by: mf2k (Frank Schima) Owned by: kevin@…
Priority: Normal Milestone:
Component: ports Version: 1.7.0
Keywords: Cc: jyrkiwahlstedt, skymoo (Adam Mercer), MarcusCalhoun-Lopez (Marcus Calhoun-Lopez)
Port: swig

Description (last modified by mf2k (Frank Schima))

The swig python variant now seems to be based on python26. There needs to be a python25 variant. The python26 port isn't even functional currently. py25-scipy depends on swig. So someone wanting that would suddenly also have python26 installed since +python is a default variant.

See also #15854.

Change History (14)

comment:1 Changed 15 years ago by mf2k (Frank Schima)

Description: modified (diff)

comment:2 Changed 15 years ago by jmroot (Joshua Root)

Cc: mcalhoun@… added

comment:3 Changed 15 years ago by MarcusCalhoun-Lopez (Marcus Calhoun-Lopez)

Cc: mcalhoun@… removed

Cc Me!

comment:4 in reply to:  description ; Changed 15 years ago by MarcusCalhoun-Lopez (Marcus Calhoun-Lopez)

Replying to macsforever2000@…:

The swig python variant now seems to be based on python26. There needs to be a python25 variant.

This was my doing in r42765.
I thought (perhaps mistakenly) that the latest available python should be used.
Adding a new variant would be easy enough

set bindings {
    python25      port:python25          python=${prefix}/bin/python2.5
    python26      port:python26          python=${prefix}/bin/python2.6

but making them conflict would not be pleasant.
If indeed #15854 was correct and python_select poses no problem, then that seems to me to be to be a better solution
if installing python26 is unacceptable.

The python26 port isn't even functional currently.

If this is the case, please let the maintainers know.
It is thought to be functional (although still missing many external packages).

comment:5 in reply to:  4 Changed 15 years ago by mf2k (Frank Schima)

Replying to mcalhoun@…:

if installing python26 is unacceptable.

It is if one is installing py25-scipy! ;)

Replying to mcalhoun@…:

If this is the case, please let the maintainers know.
It is thought to be functional (although still missing many external packages).

I mean ready for use in production like python 2.5. So yes I mean that since all the nice py26-* packages are not available yet since the python26 portgroup does not yet exist. :)

comment:6 Changed 15 years ago by blb@…

The problem now isn't so much that python26 isn't functional, it's that there's no python26 port group and hence, no py26-* ports that work (there are three right now, two are unneeded with the latest python26, and the third won't work).

So until there are enough modules to make python26 useful enough, I say keep python25, or until a more generic, non-version, method can be used.

comment:7 Changed 15 years ago by skymoo (Adam Mercer)

Cc: ram@… added

I original filled #15854 and the contents of the swig python variant is independent of the python specified. I have been using swig with this patch applied for several months now with no ill effects, py25-scipy builds without issue. I have tested building py25-scipy against the swig python variant depending on python_select with specifying a python version using python_select and it builds without issue. Therefore I cannot see why python_select can't be used as the dependency for the python variant.

However a post-activate hook should probably be added to python_select to recommend the installation of a python port, probably python25 at the moment (well at least until python26 is more mature).

comment:8 in reply to:  7 Changed 15 years ago by skymoo (Adam Mercer)

Sorry I had a typo in that last comment

...depending on python_select with specifying a...

should have said:

...depending on python_select without specifying a...

comment:9 Changed 15 years ago by MarcusCalhoun-Lopez (Marcus Calhoun-Lopez)

The original change to swig was committed after maintainer timeout of #17391.
Unless I am mistaken, I do not think that either of the maintainers have commented on this ticket either.

Unless someone else has any objection, I vote for changing the swig dependency to python_select.

comment:10 Changed 15 years ago by MarcusCalhoun-Lopez (Marcus Calhoun-Lopez)

Cc: mcalhoun@… added

Cc Me!

comment:11 in reply to:  9 Changed 15 years ago by skymoo (Adam Mercer)

Replying to mcalhoun@…:

Unless someone else has any objection, I vote for changing the swig dependency to python_select.

Seconded, obviously. Just doubled checked my last statement, as the last time I checked that was several months ago, and py25-scipy builds (and runs!) without issue (ignoring all the nose deprecation warnings which are resolved in 0.7.0.svn which I've also checked) against swig built against python_select pointing to python25-apple on Intel Leopard.

comment:12 in reply to:  9 Changed 15 years ago by mf2k (Frank Schima)

Replying to mcalhoun@…:

Unless someone else has any objection, I vote for changing the swig dependency to python_select.

This makes complete sense to me too.

comment:13 Changed 15 years ago by MarcusCalhoun-Lopez (Marcus Calhoun-Lopez)

Resolution: fixed
Status: newclosed

Fixed in r43198 (maintainer timeout).

comment:14 Changed 15 years ago by (none)

Milestone: Port Bugs

Milestone Port Bugs deleted

Note: See TracTickets for help on using tickets.