#13 Error message "invalid" etc.

[Steve Keen]

Here's a weird one! I was stepping through countries on this Ravel and the error message about an invalid access turned up. I pressed on the OK and then the Ravel looked like this: two axes imposed upon each other, along with a "no data" message. I'll try to reproduce it and record the sequence.

[Steve Keen]

Bingo! Caught it.

[High Performance Coder]

Invalid data was a message from EcoLab, pretty much meaning the same thing as "No data", but written sometimes when the xrange or yrange is not automatically set. I removed that message, as "no data" is more meaningful. And there is no data sometimes, because in 1951, only a small number of nations have GDP values in this database, and if you have a 2D ravel, only the first 10 countries are plotted, or alternatively if you have a subselection of countries, maybe none have data for that year.

The other situation of two handles landing on top of each other occurs, I think, when one handle of a locked ravel is not present in another, although that doesn't seem to apply in this specific case. Anyway, I've now do an implicit redistributeHandles when applying state from one ravel to another, and filtered out the case of a ravel applying its state to itself to avoid redistributing handles on the master Ravel.

We'll see if this solves the problem...

[Steve Keen]

Unfortunately not! I've done a 40 minute movie of manipulating the Data Dot up and down an Arm of Coupled Ravels, and it turns up effectively at random, and not related to the data itself (ie, it might happen at 1973-Q1, then you backtrack over 1973-Q1 again, and it's fine). When it happens, ALMOST all of the time the Slave Ravel resets to the beginning of the data set (1952-Q1) while the Master is still on 1973-Q1. Then the next movement of the Dot resets the Slave and everything is fine again until the next time.

ONCE, the Slave Data Dot didn't reset to 1952-Q1, meaning that this bug is even more elaborate than it seemed.

You will now have to listen to me ramble through a 40 minute video, but I think it's worth it! The video is far too large to post here, so I'll send it by email as a Google Doc link

[High Performance Coder]

Ticket moved from /p/minsky/tickets/927/

Can't be converted:

• _priority: 4normal

[High Performance Coder]

There doesn't seem to be any data attached to these mky files, so I can't investigate this further. I guess this ticket has been around for a while, from before we included the data in the model files. For reference, the datasets are:

• D:/Documents/Ravel/Data/PrivDebtNationalCurrency.csv
• D:/Documents/Ravel/Data/GDP_NationalCurrency.csv

[High Performance Coder]

Yeah - these files date from the old days when Ravels used to import the data directly, not parameters. If you could find those two datasets, I should be able to reproduce this problem (or not if it has gone away).

[Steve Keen]

This looks very dated, so let's hope it's been fixed by other changes. Might as well change the statud of this one.