#417 closed defect (fixed)
Whizard job restart doesn't work
Reported by: | fbach | Owned by: | kilian |
---|---|---|---|
Priority: | P2 | Milestone: | v2.0.6 |
Component: | core | Version: | 2.0.5 |
Severity: | major | Keywords: | |
Cc: |
Description
Using the latest trunk, when I start a Whizard job (e.g. foo.sin attached), then interrupt during event generation (foo1.log) and restart Whizard with the identical sindarin file, Whizard claims that the cut configuration has changed and discards the grid (foo2.log).
Attachments (4)
Change History (11)
Changed 14 years ago by
Changed 14 years ago by
Changed 14 years ago by
comment:1 Changed 14 years ago by
comment:2 Changed 14 years ago by
Resolution: | → fixed |
---|---|
Status: | new → closed |
Done in r3266. The cause was the missing initialization of the 'known' flag in a value in the cut expression. It went through unnoticed because the initial value was undefined but mostly happened to be correct. However, here it changed between first and second run.
comment:3 Changed 14 years ago by
Resolution: | fixed |
---|---|
Status: | closed → reopened |
Admittedly, now Whizard recognizes the old grid after restart, but rather than carrying on with event generation, it just stops after reading the evx file (cf. foo3.log). This is independent of the requested output format (at least I checked with evx and lhef).
Changed 14 years ago by
comment:4 follow-up: 5 Changed 14 years ago by
Resolution: | → fixed |
---|---|
Status: | reopened → closed |
This time you are wrong. I killed the file after having generated 48 events. Started again, it read in everything and generated the missing 952 events. All as expected for.... Closing. (Seems your old event file was already complete)
comment:5 Changed 14 years ago by
Replying to jr_reuter:
This time you are wrong. I killed the file after having generated 48 events. Started again, it read in everything and generated the missing 952 events. All as expected for.... Closing. (Seems your old event file was already complete)
Sorry, you are right with evx format (seems like I was too slow with manual interrupt while requiring just 1000 events). However, I can still confirm the behaviour I described above when producing lhef format: Whizard didn't restart event generation although there were only 131 out of 10000 events in the lhef file. Did you check that too? In any case, at least for me there's a simple and feasible workaround for that, i.e. generate events in evx format first, then rerun the sin file with lhef flag turned on for conversion, so I don't know if this is worth reopening the ticket...
comment:7 Changed 14 years ago by
Note that Whizard can't read from LHEF, just write. For recovering, only .evx is useful (and that's the purpose of that format).
JRR confirmed this behavior.