"no method found" points to the wrong line of code

[ssiccha]

When I run

f := function() 
  local l;
  l := 0 * [1..6];
  l[[1..3]] := 1;
end;
f();

I get the error msg

Error, no method found! For debugging hints type ?Recovery from NoMethodFound
Error, no 1st choice method found for `[]:=' on 3 arguments called from
0 * [ 1 .. 6 ] at stream:3 called from
<function "f">( <arguments> )
 called from read-eval loop at line 5 of stream
...

although l[[1..3]] := 1; is the offending part of code.
This does not happen with my version of GAP 4.7.9.

[ChrisJefferson]

This has actually also got worse in the GAP 4.8 betas. I will investigate.

[ChrisJefferson]

Further investigation (for my own records mainly). The problem is that this used to be invalid because you can only index lists with integers. Now it's invalid because we can index lists with anything, but l in particular doesn't do indexing by a range, so the underlying error was always there (wrong marking of place where error occurs), but it didn't show up before.

[fingolfin]

This problem also occurs in plain GAP.

[fingolfin]

For some background: this is because we don't call SET_BRK_CALL_TO enough. I do have a fix for that ready; the main thing missing is a collection of tests which verify this doesn't break other things, and also ensuring this doesn't regress.
I.e. I want a bunch of tests which record the reported error, in different scenarios.