Repository navigation
Bug in macro replacement in display #3950
Description
I think I have found a bug in macro replacement in display files. In my case these are files that are converted from medm .adl files, but I think the problem is more general.
I originally saw the problem using the adl converter from Phoebus 4.6.10 and Phoebus display 4.7.3. But I have reproduced it using both the converter and display in Phoebus 6.0.0.
This is an example of the problem.
quadEM.bob loads the related display file NDStats, and passes the macro parameters P=QE1:FX4:, R=Current1:
NDStats.bob loads the related display file NDStatsTimeSeriesPlot, and passes the parameters P=$(P), R=$(R)TS:, RN=$(R)TSTotal
The title of this plot is $(P)$(RN)
After substitution the plot title should thus be QE1:FX4:Current1:TSTotal.
However, it is not. As seen below it is QE1:FX4:Current1:TS:TSTotal.
The reason is clear. This is the screen for the Actions of the menu that calls NDStatsTimeSeriesPlot:
What is happening is that it is not using the original value of R when it builds RN, it is using the modified version from the first line.
I can work around the problem by swapping these 2 lines in the .adl file, so that RN is built before R. But it seems that this should be fixed in the translation or runtime engine.
Activity
shroffk commentedon Oct 14, 2026
Thanks for the example and the diagnosis.
I believe that the ordered and immediate resolution of the macros was a deliberate choice and this "feature" is being used at some facilities. ( The same is true when it comes of rules, where the conditions are all checked sequentially and the order matters )
We should discussion if this is still the best approach, and clearly document the decision.
@MarkRivers I will assume that your suggestion is that all the macros should be resolved "simultaneously" when the action is executed and so that their value is not dependent on the order of their declaration.
MarkRivers commentedon Oct 14, 2026
@shroffk , yes i am suggesting that it should not matter what order macro arguments are listed. That is how medm and perhaps other OPIs work. Lots of areaDetector screens don’t work correctly in Phoebus now.