sim: fix emulator after introduction of new submit RPC

[SteVwonder]

The main problem was a pre-mature destruction of the flux_future. The unpacked values were later accessed in the packing of the wreck event. While making changes in the simulator, went ahead and added some additional error checking/messaging and refactored to take advantage of the new job.submit-nocreate RPC.

Fixes: #303

[SteVwonder]

@grondo: just want to confirm that I interpreted the job.submit-nocreate control flow correctly. It does not change the state of the job in the KVS to "submitted", correct?

[garlick]

(Sorry if I introduced the premature future destruction bug when updating the KVS API!)

[codecov-io]

Codecov Report

Merging #304 into master will increase coverage by 6.37%.
The diff coverage is 70%.

@@            Coverage Diff             @@
##           master     #304      +/-   ##
==========================================
+ Coverage   67.74%   74.11%   +6.37%     
==========================================
  Files          46       49       +3     
  Lines        8692     9287     +595     
==========================================
+ Hits         5888     6883     +995     
+ Misses       2804     2404     -400

Impacted Files	Coverage Δ
simulator/submitsrv.c	77.5% <70%> (ø)
simulator/simsrv.c	76.64% <0%> (ø)
simulator/sim_execsrv.c	84.77% <0%> (ø)
resrc/resrc.c	83.05% <0%> (+1.27%)	⬆️
sched/sched_fcfs.c	94.31% <0%> (+5.68%)	⬆️
src/common/libutil/shortjansson.h	88.33% <0%> (+10.36%)	⬆️
sched/sched.c	73.95% <0%> (+12.25%)	⬆️
sched/rs2rank.c	94.82% <0%> (+14.65%)	⬆️
sched/rsreader.c	96.52% <0%> (+20.13%)	⬆️
... and 2 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 02b17e9...fa44e05. Read the comment docs.

[coveralls]

Coverage increased (+6.9%) to 75.851% when pulling fa44e05 on SteVwonder:fix-emulator-new-submit into 02b17e9 on flux-framework:master.

[dongahn]

Thanks @SteVwonder. LGTM!

I will merge this after flux-core 1409 gets in.

[garlick]

flux-framework/flux-core#1409 went in just now, so kicking off tests again to run against the new master.

[grondo]

@grondo: just want to confirm that I interpreted the job.submit-nocreate control flow correctly. It does not change the state of the job in the KVS to "submitted", correct?

@SteVwonder, oops I missed your question yesterday, sorry. The job.submit-nocreate probably should have updated the job state to submitted for symmetry with job.submit. However, now that you point it out, I left that out which was probably an oversight. Sorry!

We can fix that up later if it makes more sense. (Current behavior might be confusing because it emits the submitted event, but doesn't change the state)

[grondo]

Opened flux-framework/flux-core#1410 on the question of submitted state for job.submit-nocreate RPC.

[garlick]

Thanks! Merging.