Python versions (was Re: os.execl())

Mats Wichmann mats at laplaza.org
Tue Aug 28 00:27:40 CEST 2001


On Mon, 27 Aug 2001 17:54:08 GMT, Sheila King <sheila at spamcop.net>
wrote:

:On 27 Aug 2001 10:29:44 -0700, aahz at panix.com (Aahz Maruch) wrote in
:comp.lang.python in article <9me028$ecq$1 at panix2.panix.com>:
:
::In article <slrn9okse0.c6a.grante at grante.comtrol.com>,
::Grant Edwards <grante at visi.com> wrote:
::>
::>I didn't realize we were discussing 1.5.2.  I assume people are
::>using a fairly recent version unless they specify otherwise.
::>Under 2.1, spawnv is available for Unix. I don't remember if it
::>was in 2.0.
::
::Until Red Hat et al upgrade the default installation, I assume there's a
::good chance questions refer to 1.5.2.
:
:I will second that. This is one of the primary reasons that my webhost
:will not entertain upgrading the system Python to a more recent version.
:He doesn't want to risk "breaking" something, and says that Python is
:"built-in" to the system, somehow.

Sigh.  I have this problem with my ISP, and affecting only mailman.
Mailman now depends on 2.x Python and since the ISP won't upgrade
Python they can't upgrade Mailman.  We're sinking into a hole here...

:Plus, there is the whole thing, with some people not even wanting to
:touch the RH 7.x. (I don't understand all the issues around this, but
:the fact is that even if RH does update their Python in their install
:routines, that some people don't even want to use the latest RH, for
:some reason...)

Youv'e got it wrong, Sheila.  The problem is that people have learned
from bitter experience not to trust Red Hat x.0 releases, the .1's are
marginally useful, the .2's often actually work - but by that time
they're almost ready for (x+1).0.  Sigh.

Unfortunately Red Hat have recently decided that they won't change
major versions of key packages on minor Red Hat releases... so we're
stuck with Python 1.5.2 as the default until, oh, about RedHat 8.2.

Folks need to be told that Linux is wonderful, but Red Hat doesn't
necessarily prove that in their choice of what to bundle...


Mats Wichmann




More information about the Python-list mailing list