| 1 |
On Monday, 15 October 2018 19:49:59 BST Philip Webb wrote: |
| 2 |
> 181015 Dale wrote: |
| 3 |
> > Just curious, did you notice this little part? |
| 4 |
> > "LLVM ERROR: IO failure on output stream: No space left on device" |
| 5 |
> > You may want to make sure you are not out of disk space |
| 6 |
> > wherever your tmp directory is or out of ram if you use tmpfs. |
| 7 |
> |
| 8 |
> Yes, I did, as I said, & added 2 lines to 'package.env'. |
| 9 |
> That solved that problem, which was surprising : |
| 10 |
> my explanation is that FF itself is too big to use 'tmpfs' |
| 11 |
> & this then squeezes out any other pkgs to be compiled along with it, |
| 12 |
> even a tiny virtual. Otherwise, the 1st problem was USE flags. |
| 13 |
> |
| 14 |
> The new FF requires some very big items, which took a long time to emerge : |
| 15 |
> Rust (59), Clang (11), Llvm (15), FF (33) : total 118 min . |
| 16 |
> The total download was c 500 MB . LO is modest in comparison. |
| 17 |
> |
| 18 |
> Now to get some groceries, then I'll try it out. |
| 19 |
> The big question is whether I can still group tabs, |
| 20 |
> whether directly with FF or via some add-on (whatever they're now called). |
| 21 |
> |
| 22 |
> Thanks for offering a bit of help. |
| 23 |
|
| 24 |
I've noticed the same both in terms of the dependencies now being drawn in and |
| 25 |
in terms of how much RAM the compile consumes. On systems with low RAM I set |
| 26 |
lower MAKEOPTS jobs and average values and add plenty of swap. This keeps |
| 27 |
emerge in check and stops it from swapping in and out continuously thrashing |
| 28 |
the disk. |
| 29 |
|
| 30 |
More than a year ago I'd noticed similar uncontrolled consumption of resources |
| 31 |
by emerge on Chromium. Interestingly a few versions later something must have |
| 32 |
changed (some hardware limit checks added by devs?) and Chromium became much |
| 33 |
less hungry for resources. |
| 34 |
-- |
| 35 |
Regards, |
| 36 |
Mick |
Archives