| 1 |
Amadeusz Żołnowski wrote: |
| 2 |
> Excerpts from Dale's message of 2011-11-14 13:17:28 +0100: |
| 3 |
>> Here is some feedback then. I liked it the way it was. When a build |
| 4 |
>> fails, I do a one of install of that package and I like to see the |
| 5 |
>> output. Why, because sometimes it gives me a hint as to why it failed |
| 6 |
>> or something I can google for. |
| 7 |
> If it fails you get tail of build.log, so you see it anyway. |
| 8 |
> |
| 9 |
|
| 10 |
That doesn't always go back far enough tho. I have on my new rig seen |
| 11 |
the failure be as far back as a couple hundred lines. There are a |
| 12 |
number of times that I have had to use the Find function to even find |
| 13 |
the original failure because it is so far back. I have even set Konsole |
| 14 |
to have unlimited history. Unless it is going to tail -n 500 or more, |
| 15 |
that may not go back far enough. This is the age of multiple cores and |
| 16 |
enough ram to use tmpfs. I have both. 4 cores and 16Gbs of ram. |
| 17 |
|
| 18 |
MAKEOPTS="-j10" |
| 19 |
tmpfs on /var/tmp/portage type tmpfs (rw,noatime) |
| 20 |
|
| 21 |
The MAKEOPTS is increasing. I'm looking for that sweet spot. ;-) |
| 22 |
|
| 23 |
Dale |
| 24 |
|
| 25 |
:-) :-) |
Archives