1 |
On Tue, Jun 16, 2020 at 1:55 PM Zac Medico <zmedico@g.o> wrote: |
2 |
> |
3 |
> On 6/16/20 10:46 AM, Mike Gilbert wrote: |
4 |
> > On Tue, Jun 16, 2020 at 1:45 PM Mike Gilbert <floppym@g.o> wrote: |
5 |
> >> |
6 |
> >> On Mon, Jun 15, 2020 at 9:39 AM Sid Spry <sid@××××.us> wrote: |
7 |
> >>> |
8 |
> >>> On Mon, Jun 15, 2020, at 2:36 AM, Ulrich Mueller wrote: |
9 |
> >>>> But we know that it is in /usr/bin, so why add yet another indirection? |
10 |
> >>>> |
11 |
> >>>> Attachments: |
12 |
> >>>> * signature.asc |
13 |
> >>> |
14 |
> >>> Ah, sorry -- I forgot to note this here. If you wish to support prefix it is possible it may not be in /usr/bin. Granted I am not sure if the prefix stage3 I was using is old enough to be broken in some way, but adding this would prevent future breakage. |
15 |
> >> |
16 |
> >> The portage ebuild and the python distutils module already take care |
17 |
> >> of updating shebangs at install time. |
18 |
> > |
19 |
> > I suppose your patch might be useful if you are trying to run portage |
20 |
> > from a git checkout on a prefix system. |
21 |
> > |
22 |
> |
23 |
> So, given that the ebuild updates shebangs automatically, should't we |
24 |
> optimize the default shebangs to be as flexible as possible? |
25 |
|
26 |
Yes, that makes sense. |
27 |
|
28 |
However, we should test to make sure that distutils is smart enough to |
29 |
parse that "/usr/bin/env -S python" string and replace it with |
30 |
version-specific python shebang. |