Closed jodavies closed 2 weeks ago
Should this check (*e < 0
) be in poly::argument_to_poly()
or poly::get_variables()
?
You are right, this check could equally go in get_variables
and makes more sense there. I will move it.
While here it would probably be fine, other areas of the code have been cumbersome to debug because you have to determine which of many goto
have actually caused a termination. In my opinion it could be beneficial to actually remove this pattern from the code entirely, but that requires some effort also...
How about giving Terminate a second parameter that indicates which call it was? Then one can change things easily, using grep.
On 8 Nov 2024, at 10:07, jodavies @.***> wrote:
While here it would probably be fine, other areas of the code have been cumbersome to debug because you have to determine which of many goto have actually caused a termination. In my opinion it could be beneficial to actually remove this pattern from the code entirely, but that requires some effort also...
— Reply to this email directly, view it on GitHub https://github.com/vermaseren/form/pull/570#issuecomment-2464185242, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABJPCEU344ZMQ7VLOMIODO3Z7R5O7AVCNFSM6AAAAABRIMFKASVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINRUGE4DKMRUGI. You are receiving this because you are subscribed to this thread.
When this PR is done https://github.com/vermaseren/form/pull/526 , Terminate
is replaced with a macro which prints the file and line information of every terminate call.
But still if you have several goto
which land at a Terminate
, you can't know which goto
caused the crash without stepping through in a debugger (hard if the crash is in a long running script with no useful way to set a breakpoint) or putting print statements at each of the goto
.
These created numerators of 0. Fixes #567