Project

General

Profile

Bug #1527

Inadequate error message on incorrect list access

Added by Julian Danner over 3 years ago. Updated over 3 years ago.

Status:
Closed
Priority:
Low
Assignee:
John Abbott
Category:
enhancing/improving
Target version:
CoCoA-5.4.0
Start date:
30 Oct 2020
Due date:
% Done:

100%

Estimated time:
2.33 h
Spent time:
2.35 h

Description

When trying to access a list using another list, the error message seems to be incorrect.

Here is an example:


L:=1..10;
L[[1]];

which gives the error
ERROR: Expecting type INT or STRING, but found type LIST.

AFAIK a list cannot be accessed by a string, at least I could find nothing in the manual about it.

History

#1 Updated by John Abbott over 3 years ago

  • Category set to enhancing/improving
  • Status changed from New to In Progress
  • Target version set to CoCoA-5.4.0
  • % Done changed from 0 to 10

The relevant function seems to be IndexedAccessExpression::implEval in Interpreter.C around lines 3677--3695.

Now sure how to fix it. Need to find out whether the indexed object is a RECORD; and if not, restrict indexes to being just integers.

#2 Updated by John Abbott over 3 years ago

  • Assignee set to John Abbott

I have a partial fix: the error mesg now says "Expecting type STRING" if the value being indexed is a RECORD otherwise it says "Expecting type INT".

This is not entirely satisfactory, because there are other "weird" cases where the error mesg is still unhelpful:

I := ideal(x);
I[I]; --> ERROR: expecting type INT

Really it should check if the value being indexed is indexable, and if not then say "ERROR: type IDEAL cannot be indexed"

#3 Updated by John Abbott over 3 years ago

There is a fn canBeIndexedByBigInt, but there does not seem to be a canBeIndexedByString*.

Maybe I should make the string version (it is a mem fn), and then test if both give false then complain that the object is not indexable.

That is probably more helpful to the user... do I want to spend more time deciphering the "self-documenting" interpreter? :-/

#4 Updated by John Abbott over 3 years ago

Na toll:

--> ERROR: The type LIST cannot be indexed, but there is an index

Time to go to sleep...

#5 Updated by John Abbott over 3 years ago

I do not understand lines 146 and 147 in AST.H.
They seem to say that everything can be indexed by both INT and STRING... seems like dodgy code to me!

#6 Updated by John Abbott over 3 years ago

  • Status changed from In Progress to Resolved
  • % Done changed from 10 to 70

I have made a change (by adding a call to mem fn asRightValue).
Now the code seems to be doing more or less what I had hoped/expected.
Perhaps I'll wait until tomorrow before checking in.
I make no claims about correctness nor efficiency... it compiles and the CoCoA-5 tests passed!

#7 Updated by John Abbott over 3 years ago

This is what happens now:

--> ERROR: Expecting type INT, but found type LIST
--> L[[1]];
-->   ^^^

Also we get with an ideal:

/**/ I := ideal(x);
/**/ I[1];
--> ERROR: The type IDEAL cannot be indexed, but there is an index
--> I[1];
-->   ^

#8 Updated by John Abbott over 3 years ago

  • Status changed from Resolved to Feedback
  • % Done changed from 70 to 90

I have checked in the code.

#9 Updated by John Abbott over 3 years ago

  • Status changed from Feedback to Closed
  • % Done changed from 90 to 100
  • Estimated time set to 2.33 h

Also available in: Atom PDF