1 |
On Thu, 1 Nov 2012 14:48:55 -0400 |
2 |
Mike Gilbert <floppym@g.o> wrote: |
3 |
|
4 |
> On Thu, Nov 1, 2012 at 11:44 AM, Mike Gilbert <floppymaster@×××××.com> wrote: |
5 |
> > On Thu, Nov 1, 2012 at 7:54 AM, Michał Górny <mgorny@g.o> wrote: |
6 |
> >> The idea is very simple: /etc/env.d/python/python[23] with a one-line |
7 |
> >> value similar to the main interpreter /config file. |
8 |
> >> |
9 |
> >> That should be simpler & more reliable than reading a symlink. And at |
10 |
> >> some point we can replace the symlink with an $EPYTHON-aware wrapper |
11 |
> >> as well. |
12 |
> > |
13 |
> > I don't understand the point of this. Do we have some need to enable |
14 |
> > EPYTHON usages for scripts that have a python2 or python3 shebang? |
15 |
|
16 |
I'm mostly thinking of a future potential. Or maybe just I dislike that |
17 |
partially we're using wrappers, and partially symlinks. |
18 |
|
19 |
> > I also don't understand how a text file is more reliable than a |
20 |
> > symlink; they are basically the same thing, but the symlink has a |
21 |
> > different file mode. |
22 |
|
23 |
They have a different goal, I believe. The goal of a symlink is to tie |
24 |
one path onto another; the goal of a file is to store data. I'd say |
25 |
that symlink is valid as long as it points to the correct path, one way |
26 |
or the other. |
27 |
|
28 |
But well, it's more like 'we are keeping python in one place, but |
29 |
python2/3 in another'. python-exec would also become simpler if one |
30 |
function would be needed to get both kinds of implementations rather |
31 |
than two different approaches. |
32 |
|
33 |
Last but probably least, POSIX doesn't list 'readlink' utility. Not |
34 |
sure how various Prefixes (interix?) handle that but I really dislike |
35 |
the idea of storing data in symlinks. |
36 |
|
37 |
-- |
38 |
Best regards, |
39 |
Michał Górny |