1 |
On 01/11/2014 02:11 AM, Ciaran McCreesh wrote: |
2 |
> On Fri, 10 Jan 2014 08:31:21 +0800 |
3 |
> Patrick Lauer <patrick@g.o> wrote: |
4 |
>> On 01/10/2014 08:16 AM, heroxbd@g.o wrote: |
5 |
>>> Igor <lanthruster@×××××.com> writes: |
6 |
>>> |
7 |
>>>> The ebuilds have approximately the same time to install, the |
8 |
>>>> failure rate is about the same, emerge is getting slower. |
9 |
>>> |
10 |
>>> I am curious about the slowness of emerge. |
11 |
>>> |
12 |
>>> How about profile the portage and rewrite the time-crucial part in |
13 |
>>> C/C++, or ideally, borrowing the counterpart from paludis? How |
14 |
>>> feasible is that? |
15 |
>> |
16 |
>> Last I checked paludis wasn't faster - on average portage was a few |
17 |
>> percents faster. |
18 |
> |
19 |
> Your benchmark was comparing uncached behaviour, where bash is the slow |
20 |
> part and which users don't see. |
21 |
Wrong - even the cached cases was showing the same timing proportions. |
22 |
|
23 |
And users see the uncached case whenever they use an overlay. |
24 |
|
25 |
> You were also not comparing like with |
26 |
> like -- any benchmarks of this nature should be taken with a heavy |
27 |
> pinch of salt, since Portage with everything turned on does less |
28 |
> validation that Paludis does with everything turned off... |
29 |
> |
30 |
Not my problem, bad code is bad. |