new history problems
Issue #197
resolved
I will use this bug report to track all the problems I observe with the new history display. K.O.
Comments (20)
-
reporter
-
reporter
missing horizontal rules. compare https://daq16.triumf.ca/?cmd=oldhistory&group=Trigger&panel=Clk625
and https://daq16.triumf.ca/?cmd=History&group=Trigger&panel=Clk625
K.O.
-
reporter
-
reporter
this gets stuck https://daq16.triumf.ca/?cmd=History&group=TMB&panel=test
-
-
assigned issue to
Stefan Ritt
-
assigned issue to
-
reporter
plot shows wrong colors: all lines are gray, after zooming in twice, they are very light gray and I can barely see the plots: https://daq16.triumf.ca/?cmd=History&group=PWBcol0&panel=v_sca12
K.O.
-
reporter
history plot odb settings are destroyed: compare with sca34 plot (or any other column plots, they are all populated the same way by a script). note how “colors” are missing and “factor” is set to zero. https://daq16.triumf.ca/?cmd=odb&odb_path=History%2FDisplay%2FPWBcol0%2Fv%5Fsca12
-
reporter
andrea reported that his plots do not work because “factor” is set to zero. indeed, the “factor” column was removed from the new history plot editor and one must use the odb editor to see why plots do not work and to fix the wrong “factor” (set it to 1). K.O.
-
reporter
the history plot editor column used for reordering and removing individual entries has been removed. K.O.
-
reporter
just looking at history plots using new viewer changes the plot definitions in odb. (I do not enter the history plot editor!). the changes in odb are not the same every time: “factor” can change from an array of 8 float to a single float (value is usually “1”) or stay an array of 8 float but value is “1,0,0,0,0,0,0,0” (factor values 0 is wrong, of course). “offset” seems to change the same way, but usual value is zero, so hard to tell. “colour” is sometimes erased, sometimes not. “formula” is sometimes added, sometimes not. K.O.
-
reporter
request: just looking at plots should not change anything in ODB! K.O.
-
reporter
request: new history plot viewer and editor should be compatible with the old one until all bugs in the new system are fixed and nobody needs to use the old viewer. (this is not the case yet).
specifically, new editor should not change plot definitions in odb in a way that will break the old history viewer. (and the new history viewer should not change anything in odb at all).
K.O.
-
reporter
switching from some other tab to a history viewer plot tab takes a very long time (5-10-15 seconds). K.O.
-
There is no “new” editor. The editor is still the old one written by KO. But he does not have time to look into that, so you have to be patient.
-
Let me reply on the other raised issues:
- If history lines are grey, one uses more than 16 variables which is discouraged.
- I could not reproduce the history display changing ODB entries. The only change I see is if one activated the history panel editor, the “Factor” is added. This is used in the new system instead “Offset/Factor” because it’s more flexible. If the problem persists, I need instruction how to reproduce it from a fresh midas installation (or empty /History/Display ODB tree)
- I kind of accidentally removed the re-ordering of history variables, because I didn’t realize this functionality. Why should one need it? All variables are plotted on top of each other anyhow. If one wants different colors, one can edit them directly.
- I think the panel editor should be rewritten some day as a floating dialog box like we do on many other pages. If we still need re-ordering, we can do that via drag-and-drop like everywhere else.
- I still cannot access daq16 so I cannot comment on those issues
Stefan
-
reporter
I confirm that I do not see any code that writes into odb to destroy the display settings. Very strange. Definitely happening in the agmini system. K.O.
-
reporter
I will keep this issue as a grab bag of problems as I see them. To avoid confusing many problems in one issue trackers, I will open specific bug reports for the problems I can reproduce that need fixing. K.O.
-
reporter
I see that the agdaq/agmini history plots for PWB board are wrong - offset and factor are not arrays of 8 elements. These plots were created by a script (I cannot find it!), so the mistake is probably there. After these wrong settings are loaded into the history plot editor and saved, they become even more broken. This is probably what I have seen back then. So false alarm. K.O.
-
reporter
All problems described in this bug are disposed of. Closing it. K.O.
-
reporter
- changed status to resolved
fixed midas-2019-09-e. K.O.
- Log in to comment
https://daq16.triumf.ca/?cmd=History&group=ADC&panel=fpga_temp
select “clock”, “7d”, page displays “updating data…” and is completely unresponsive for a long time. cannot click to “status” button to return to status page, cannot right-click to enter “inspect” and JS debugger, cannot reload page using browser “reload” button. Entering a new URL in the browser URL bar also does nothing. The only thing that works is the killing of the tab. K.O.