1 |
On 11/08/10 00:08, Krzysztof Pawlik wrote: |
2 |
> On 11/07/10 13:07, Mike Auty wrote: |
3 |
>> On 07/11/10 02:40, Donnie Berkholz wrote: |
4 |
>>> I read it more closely and realized I was a little confused by the way |
5 |
>>> you listed all the bullet points mixing together benefits and problems. |
6 |
>> |
7 |
>>> So I'll try again: if you really want to do this change, you might want |
8 |
>>> to consider adding a mercurial-2.eclass instead. Eclasses of this nature |
9 |
>>> (svn, git, hg, etc) tend to be broadly used outside the tree as well as |
10 |
>>> within, so breaking backwards compatibility can be a real problem. A new |
11 |
>>> versioned eclass allows for a much more gradual transition. |
12 |
>> |
13 |
>> |
14 |
>> I've only just jumped into the conversation, but the obvious question |
15 |
>> is, why not just use ${S} as the location of the working directory |
16 |
>> (rather than "${WORKDIR}/${workdir}")? That way, if it's set old |
17 |
>> ebuilds still work, and if it's not set, new ebuilds get the benefit of |
18 |
>> using ${S}? I can only see a problem with this if there's somewhere |
19 |
>> that the value of the working directory needs to be known before any of |
20 |
>> the phases... |
21 |
> |
22 |
> Hm.. good idea :) I'm attaching modified patch that uses ${S} by default, so it |
23 |
> will improve the situation and at the same time it won't break existing ebuilds. |
24 |
> Thanks Mike for this suggestion. |
25 |
|
26 |
I've just committed this version. |
27 |
|
28 |
-- |
29 |
Krzysztof Pawlik <nelchael at gentoo.org> key id: 0xF6A80E46 |
30 |
desktop-misc, java, vim, kernel, python, apache... |