On 9/18/06, Ronald Oussoren <[EMAIL PROTECTED]> wrote: > > On Sep 18, 2006, at 11:48 PM, Bob Ippolito wrote: > > > On 9/18/06, Ronald Oussoren <[EMAIL PROTECTED]> wrote: > >> > >> On Sep 18, 2006, at 10:11 PM, Gerben Wierda wrote: > >> > >> > On Sep 18, 2006, at 21:42 , Bob Ippolito wrote: > >> > > >> >> On 9/18/06, Christopher Barker <[EMAIL PROTECTED]> wrote: > >> >>> Gerben Wierda wrote: > >> >>>>> what shell are you using? > >> >>>> > >> >>>> bash (default shell). But my env has not been changed by the > >> >>> Universal pkg. > >> >>>> > >> >>>> hermione-a:~ gerben$ which python > >> >>>> /usr/bin/python > >> >>>> > >> >>>> PATH=/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin:/usr/local/ > >> >>> teTeX/bin/i386-apple-darwin-current > >> >>>> > >> >>>> > >> >>>> I also would wonder if that would be something that the > >> >>> Universal .pkg > >> >>>> should do. It would mean that I would get a new python when I > >> call > >> >>>> system utitlities that use Python from the command line > >> (because it > >> >>>> would have to be inserted at the start of my PATH and that is > >> >>> generally > >> >>>> taken as not the right thing to do) > >> >>> > >> >>> A) system scripts should have the path to Python hard coded: > >> >>> > >> >>> #!/usr/bin/python > >> >>> > >> >>> B) The installer is only supposed to alter your login shell > >> PATH, so > >> >>> anything called by the system won't be effected. > >> >>> > >> >>>> But anyway, Universal pkg did not set up my PATH for my command > >> >>> line shell. > >> >>> > >> >>> Curious: > >> >>> > >> >>> Ronald and/or Bob: isn't it supposed to have? I know it did mine: > >> >>> > >> >> > >> >> There's three different login scripts that bash can use.. The > >> 2.4.3 > >> >> postflight script knows how to patch .profile and .bash_profile... > >> >> but > >> >> if a .bash_login exists and .bash_profile does not exist then > >> it will > >> >> create a .profile that is never used. Perhaps that is what > >> happened? > >> > > >> > That is exactly what has happened. > > > > That's easy enough to fix for Python 2.5's installer. > > Except for the fact that the release process for 2.5 final started > about an hour ago. I won't be making changes to a postinstall script > literaly hours before 2.5 final is released to the public, there's no > way we'll be able to get enough testing in. >
That's fine, it's not really a common issue. It should be considered a bugfix for 2.5.1 and a 2.4.4 universal build (if that happens). Someone ought to file an issue about it so it's not forgotten. -bob _______________________________________________ Pythonmac-SIG maillist - Pythonmac-SIG@python.org http://mail.python.org/mailman/listinfo/pythonmac-sig
