With limiting the rate of local preprocessors in #486 the change from #407
should be much less needed and IMO it's now generally a needless slowdown
that causes more system calls, possibly more context switches, etc.
So change back to using large chunks by default and use the smaller ones
only if ICECC_SLOW_NETWORK=1 is set.
With limiting the rate of local preprocessors in #486 the change from #407 should be much less needed and IMO it's now generally a needless slowdown that causes more system calls, possibly more context switches, etc. So change back to using large chunks by default and use the smaller ones only if ICECC_SLOW_NETWORK=1 is set.