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

If the out expression of a computation is "too long", then Cadabra 1 deals with it in a particular manner. It would say that the output is suppressed within the notebook without giving an error. Presumably, in this case one can safely proceed to the next level of the computation. That's how I've been dealing with it so far.

However, Cadabra 2 gives out an error:

"Output cell too large (breqn dimension too large), output suppressed. See /tmp/cdbz76OR0.tex to debug this."

in a separate window with the title: "Generic TeX error". Moreover, when you go back to the notebook (upon clicking OK on the error-window), the progress wheel at the bottom continues to revolve.

It's not clear to me at this stage how to proceed further. How can I reliably go ahead and proceed to the next level of computation?

It seems to me that in some computations I can avoid such a long output only if I'm able to use "unwrap" before "distribute". However, this approach has a problem as I described in a separate question.

Therefore, at a case-specific level, my problem is the following: Being able to use "unwrap" is vital for my computation. This works only in V.2 which doesn't seem to be efficient in handling long output expressions.

in Bug reports by (300 points)

I always encounter such a problem too. I usually have to use weight or substitute to handle the long output expressions. ( FYI: drop_weight or keep_weight to simplify the expression, also substitute(_, $ [one common factor of some terms ] -> 0$) )

Yeah, sure. One should try to avoid it, but it may not always be avoidable. The problem I'm pointing to is that suppressing long output is OK, but Cadabra 2 should not give an error. If it did not give any error, as the case with Cadabra 1, then presumably you could move to the next step with reliability.

Thanks for your comments.

The current version on github has undo facilities. They may still need a bit of tweaking but at least they should prevent you from getting work lost because of a wrong key press.

2 Answers

+1 vote

Can you send me (privately, by email to info@cadabra.science) a notebook which triggers this problem? Thanks.

by (82.0k points)

I sent the details as you asked for.

Another small thing I noticed: Undo in Cadabra 2 notebook does not work.

An update: If you do not ask Cadabra 2 to display the output by ending the command with ; and follow it up with other commands which end with a shorter output, then the program displays the end result without complaining. So for the time being I'm hoping that this is a reliable procedure to follow (?).

That is always fine.

Yes, undo is still on the todo list; most of the logic is there but it still needs a bit of plumbing to get it all to work together.

+1 vote

For people searching this forum: the current version of Cadabra should no longer have any issues with very long output (it no longer relies on breqn.sty). None of the workarounds discussed in this thread should be necessary any more.

by (82.0k points)