Bug: nnz_niters counter doubles instead of incrementing

[Neplyakh]

Hi drelyea,

Thanks for maintaining this library. I’m using it for compressed sensing reconstruction in Scanning Tunneling Microscopy and it’s been very useful.

While skimming the source, I noticed what looks like a bug in the function spgl1 in the active-set stagnation counter. The code currently doubles nnz_niters instead of incrementing it:

if nnz_diff:
nnz_niters = 0
else:
nnz_niters += nnz_niters
if nnz_niters + 1 >= active_set_niters:
stat = EXIT_ACTIVE_SET

Since nnz_niters is initialized to 0, this update keeps it at 0 indefinitely (so EXIT_ACTIVE_SET would never trigger). I think the intended behavior is likely:

nnz_niters += 1

Is the doubling intentional or would you accept a fix/PR?

With kind regards,
G.N

[drrelyea]

This is definitely a bug. This code is old as dirt and I'd want to look it over to make sure you're right, because that behavior is also present in the original Matlab code, I think?

I should really look through this code and understand it to make sure. If you want to open a PR, be my guest.