| 1 |
On Sunday 17 May 2009 09:40:14 Tiziano Müller wrote: |
| 2 |
> Am Sonntag, den 17.05.2009, 01:50 +0100 schrieb Ciaran McCreesh: |
| 3 |
> > On Sun, 17 May 2009 00:35:45 +0000 (UTC) |
| 4 |
> > |
| 5 |
> > Duncan <1i5t5.duncan@×××.net> wrote: |
| 6 |
> > > As for ciaranm's argument that you're restricting changes to the |
| 7 |
> > > version string, say allowing -rc where _rc is now required, one-time |
| 8 |
> > > restriction of a year or two, yes. However, if the spec is crafted |
| 9 |
> > > such that the EAPI must be checked FIRST |
| 10 |
> > |
| 11 |
> > ...then the package manager has to inspect the metadata for every |
| 12 |
> > version of a package before it can do anything, rather than just |
| 13 |
> > starting at the best version and working downwards until it finds |
| 14 |
> > something usable, which is a pretty hefty price to pay. |
| 15 |
> |
| 16 |
> ... if the cache can be parsed at all. With GLEP-55 we might even choose |
| 17 |
> to change the cache format. |
| 18 |
|
| 19 |
We might actually decide to change the cache format with or without GLEP55. |
| 20 |
It's trivial to think of a situation where the cache location is changed so |
| 21 |
that older package managers don't see it at all, thus falling back to the slow |
| 22 |
but reliable sourcing strategy, whereas newer versions are aware of that |
| 23 |
format and will be able to use it. |
| 24 |
|
| 25 |
Please stop mixing issues, it makes it hard to discuss things. |
Archives