Sarah Spall

I wanted to write stdout/stderr and the output of rattle to the same file; to emulate someone doing '&> file'. But I'm not sure how/if the output from the 'rattleUI'...

I agree that with Rattle's current scheduling policy there is no distinction between a "recover" and a "restart", but I believe that there is a distinction between the two situations...

I wrote a new scheduling policy for Rattle, which takes advantage of the concept of recoverable hazards to avoid halting all threads and restarting when a recoverable hazard has been...

** | r-cmd read | r-cmd write | s-cmd read | s-cmd write | 2nd ---- | ----- | -----| -- | -- | -- **r-cmd read** | OK |...

I think in the worst case a pathological command has a hazard with a new cmd each time the build is re-run. But, because speculation decisions are made using all-previous...

I agree option 1 seems the best. In the model I followed something more closely to option 2, but that is just because that fit the model better, not because...

> Should the scheduler have access to the parallel information? Or should these just look like commands and the scheduler is unaware of their topology of branching? @spall - do...

> The queue of what is to be run would have to understand what blocked on what, and thus would automatically take care of running things when their dependencies had...

Yes, I can try that and if I run into major difficulties I'll document them here and we can deal with them later.

write before read: This looks good to me, and i'm gonna type out my logic here in case I am wrong. There are 4 possiblities: 1. write speculated -> read...