1 |
On Sat, 16 Jun 2012 17:24:22 +0200 |
2 |
Peter Stuge <peter@×××××.se> wrote: |
3 |
> Ciaran McCreesh wrote: |
4 |
> > > Could it work to make automatic signatures of imported ABI, and |
5 |
> > > simply compare signatures when a provider package is updated? |
6 |
> > |
7 |
> > No. |
8 |
> |
9 |
> Can you say why? |
10 |
|
11 |
There's no way for a program to work out what "the ABI" of something |
12 |
is (whatever that means), and there's even less of a way for it to know |
13 |
up-front. |
14 |
|
15 |
> > Also, can we stop using the term "ABI" in reference to this please? |
16 |
> > It's misleading. Let's call them sub-slots instead. |
17 |
> |
18 |
> I think ABI fits well though? The situation is that A DEPENDs on B, |
19 |
> and at some point B changes in a way that A must be rebuilt in order |
20 |
> to run - right? |
21 |
> |
22 |
> The only reason that A wouldn't run anymore is that B's ABI changed? |
23 |
|
24 |
ABI has a fairly specific, technical meaning, which can be misleading |
25 |
in the general case. Is Python bytecode an ABI? |
26 |
|
27 |
-- |
28 |
Ciaran McCreesh |