| Summary: | Simulate filters out too many points with stop condition | ||
|---|---|---|---|
| Product: | UPPAAL | Reporter: | Erik Ramsgaard Wognsen <erw> |
| Component: | Engine | Assignee: | Alexandre David <adavid> |
| Status: | NEW --- | ||
| Severity: | minor | ||
| Priority: | P5 | ||
| Version: | unspecified | ||
| Hardware: | PC | ||
| OS: | Linux | ||
| Architecture: | |||
| Attachments: | A simple system to illustrate the bug | ||
Created attachment 257 [details]
A simple system to illustrate the bug
It didn't seem to keep my attachment the first time.
|
Tested in UPPAAL 4.1.14 (rev. 5212) With the attached simple system, the queries simulate 50 [<=1] {P.Done} simulate 50 [<=1] {P.Done} : P.Done simulate 50 [<=1] {P.Done} : wallclock >= 1 draw the expected result with only horizontal and vertical line segments in the plot. However, the queries simulate 50 [<=500] {P.Done} : P.Done simulate 50 [<=500] {P.Done} : wallclock >= 1 draw undesirables results with lots of sloped lines (some of them are vertical as expected, though). Presumably addPoint in server/modules/statmc/ProbaChecker.cpp considers too many points to be in the same "column" and skips them because the stop condition changes the effective overall plot width from 500 to 1 time unit. Changing the trace resolution in the Statistical parameters menu affects which lines in the plot become sloped and which ones that stay vertical as they were expected to.