GLS ramp fitting bad input data should result in zero output, but results in NaNs.

[stscijgbot-jp]

Issue JP-2377 was created on JIRA by Kenneth MacDonald:

A number of tests are failing to return expected zeros when bad data is inputed.  But test_nocrs_noflux_firstrows_are_nan and test_bad_gain_value should return zeros, but return NaNs.

[stscijgbot-jp]

Comment by David Law on JIRA:

Kenneth MacDonald Is this still relevant?  Generally I'd expect NaN slopes when there is no valid data to fit?

[stscijgbot-jp]

Comment by Kenneth MacDonald on JIRA:

David Law the GLS algorithm is still under development.  I have set of tickets for known problems with the existing code.  The GLS algorithm is low priority, so has not been worked on in a long time as other things have taken priority.

[stscijgbot-jp]

Comment by David Law on JIRA:

Kenneth MacDonald In that case I'll set this ticket to 'On Hold', agreed that there are higher priorities.

[stscijgbot-jp]

Comment by Melanie Clarke on JIRA:

Closing, since we will remove GLS code in JP-3871