Changes between Version 57 and Version 58 of ticket/1851/General
- Timestamp:
- 2017-03-09T09:50:11+01:00 (7 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
ticket/1851/General
v57 v58 445 445 446 446 Add more writes to make certain. Also write out all other pertinent scalars. 447 448 OK well all inputs to this routine are identical (or should be overwritten) and the scalars are indeed OK and teh neuler condition is false as expected. But what we've missed is a call 449 to "div_cur" lurking in amongst the apparently benign code. This hides a multitude of horrific possibilities since it does calculations of its own and potentially calls at least 4 other 450 subroutines which can modify hdivn with a myriad of other arrays and scalars which are completely invisible. So as usual, having spent days narrowing things down to one or two particular arrays 451 the whole thing just explodes in our face again with dozens of other potential sources of trouble.