Manager cannot handle corrupted scenario files. about rfswarm

[damies13]

No it's not zero, because if the first row in the GUI had an index of 0 it would look weird, and I wanted the index in the file to match what it shows in the GUI.
I actually thought long and hard about that back in the early versions, probably spent most of a day if not more thinking about what it should be.

[damies13]

I slightly disagree with the expected behavior, what you've shown here is what I would expect as the behavior for this scenario file, as long as the GUI doesn't crash and remains usable this working as it should.
By remains usable, I mean (in either first load or second load examples):

• user is able to make changes
• user is able to add and remove rows
• user is able to operate all buttons, fields and menus
• user is able to save and close
• User can close without saving and scenario is unchanged

Because the the data for "Example Test Case" is row 7 but the scenario says only 3 rows should be there, row 7 should not be loaded, If a user manually modifies a scenario to remove rows, they also should update the row number of the row data to reflect the row position they want it to have
If scenario says only 7 or more rows, and the first 6 are missing, I'd expect the 7th one to be loaded first, 8th one to be loaded second, etc, and they should only get renumbers if the user saves the scenario.

• Did you actually experience a crash? by crash I mean the UI was locked up.
• Did you get stack trace errors on the console?
• Perhaps this could be documented better? but this is not a normal usage scenario, the GUI should never produce a scenario file like this.

[ArekKuczynski]

Hmm i think you're right, the Manager should work like this and if the user decides to modify the scenario file, he must be careful with the structure. I also found one more thing: If i load this scenario file, the Manager will add extra row.

[Scenario]
uploadmode = err
scriptcount = 8
graphlist = 

[4]
robots = 22
delay = 23
rampup = 24
run = 25
test = Example Test Case
script = example.robot

[8]
robots = 71
delay = 72
rampup = 73
run = 74
test = Example Test Case
script = example.robot
excludelibraries = row7el
robotoptions = row7ro
testrepeater = True
injectsleepenabled = True
injectsleepminimum = 70
injectsleepmaximum = 75
disableloglog = True
disablelogreport = True
disablelogoutput = True

[ArekKuczynski]

Ah it should also work like this the first index is not zero :)

[ArekKuczynski]

• Okay, I was testing corrupted scenarios again but this time with correct scriptcount, and when I created a file like this:

[Scenario]
uploadmode = err
scriptcount = 7
graphlist = 

[4]
robots = 22
delay = 23
rampup = 24
run = 25
test = Example Test Case4
script = example4.robot
excludelibraries = row4el
robotoptions = row4ro
testrepeater = True
injectsleepenabled = True
injectsleepminimum = 70
injectsleepmaximum = 75
disableloglog = True
disablelogreport = True
disablelogoutput = True

[7]
robots = 71
delay = 72
rampup = 73
run = 74
test = Example Test Case7
script = example7.robot
excludelibraries = row7el
robotoptions = row7ro
testrepeater = True
injectsleepenabled = True
injectsleepminimum = 30
injectsleepmaximum = 60
disableloglog = True
disablelogreport = True
disablelogoutput = True

This is a result, also the settings form first row disappeared and Manager did not crash:

Looks like the core.OpenFile() function is generating an error message when trying to get settings data from row.
And i think the Manager should work well with transferring indexes to right order, but if you add settings that error occurs.

• When you create scenario file without settings like this:

[Scenario]
uploadmode = err
scriptcount = 7
graphlist = 

[4]
robots = 22
delay = 23
rampup = 24
run = 25
test = Example Test Case4
script = example4.robot

[7]
robots = 71
delay = 72
rampup = 73
run = 74
test = Example Test Case7
script = example7.robot

There will be not any errors in terminal but Manager will add extra row.

• One thing i wonder about is if anyone will ever create this type of scenario. When someone modifies it, they should be aware of any small mistakes in scenario structure.

[damies13]

This is something the manager should handle more gracefully, it should check first if the row index data exists, it looks like it is for the options in the run screen but not for the options in the settings screen.

Good find 👍

[damies13]

Merged into release branch v1.3.1 and all test passed