1 |
On Wed, Aug 03, 2011 at 12:34:21PM +0100, Ciaran McCreesh wrote: |
2 |
> On Tue, 2 Aug 2011 17:29:29 -0700 |
3 |
> Brian Harring <ferringb@×××××.com> wrote: |
4 |
> > That's not a "massive change" to vdb behaviour either; file |
5 |
> > collisions aren't supposed to occur, as such ownership of the file is |
6 |
> > basically guranteed back to a single package. Throw in |
7 |
> > CONFIG_PROTECT for adjusting the behaviour, and you have a far more |
8 |
> > preferable norm than "lets just leave a shit ton of .pyc/.pyo on the |
9 |
> > fs". |
10 |
> |
11 |
> It is a massive change, since if the feature is there then people don't |
12 |
> feel bad about writing lousy pkg_ functions that leave a load |
13 |
> of .pyc / .pyo files all over the place. |
14 |
|
15 |
Quoting the good spec: |
16 |
|
17 |
"The unmerge process removes an installed package's files. It is not |
18 |
covered in detail in this specification." |
19 |
|
20 |
Aka, ebuild's should be written to assume the files they install get |
21 |
wiped; there is *zero* mention of mtime, nor could any ebuild rely on |
22 |
it and be compliant. |
23 |
|
24 |
Background as to why we ever relied on mtime- it was a hack to work |
25 |
around a bad implementation in portage (treewalk function); it didn't |
26 |
actually know if it was replacing or what not, so mtime was what was |
27 |
relied on- afaik, that being the sole reason we shoved mtime into |
28 |
the vdb also. |
29 |
|
30 |
At least from the portage standpoint, shifting away from mtime |
31 |
reliance was on the radar since '05 and implemented at least |
32 |
initially by '06... exact date it was released from a stable branch I |
33 |
couldn't tell you, but it's been there a long while. |
34 |
|
35 |
~brian |