Welcome to Cadabra Q&A, where you can ask questions and receive answers from other members of the community.
+1 vote

While using split index on an expression (named as 'ex' in the current case), the expression name C_aA repeats itself in the RHS with an equals sign.


Is this a bug or is there information I'm missing?

asked in Bug reports by (150 points)

2 Answers

+2 votes

There are some known issues with split_index on expressions which are not fully distributed; can you try distribute on your expression first?

answered by (54.1k points)

Dear Kasper,

The error remains the same even after I distribute.


Oddly, this bug does not occur for some expression like the Kretschman scalar (self-contraction of the Riemann tensor).

Can you send a (possibly minimal) notebook which reproduces this problem? It's difficult to guess what is going on based on that screenshot.

Dear Kasper,

This bug seems to be occurring all the times I have something like

expr:= K_{a b ...} = "some expression"

If I simply put

expr:= "some expression",

the split index bug does not occur. However, if I need to use K_{a b ...} later in some other calculation, the latter method doesn't help.

Let me have a look.

+2 votes

I have pushed a fix for this to github master. Are you able to build from source?

answered by (54.1k points)

I've built my current version from source. Do I update or rebuild it?

git pull, and then run make and sudo make install again.

Dear Kasper,

It works perfectly. Thanks a lot.

Best regards.

Very good. Just for completeness: your notebook has a few weird things in it, e.g. Coordinate with a name as argument (it doesn't have one).

So the subspace specification is redundant, I see. I was trying to break a spherically symmetric space into (t,r) and a 2-sphere. Does the Metric, InverseMetric, and KroneckerDelta take names as arguments or is it the same as Coordinate?

They don't either. The manual pages are your friends (press F1 when your cursor is on a property name or algorithm name).