These are chat archives for astropy/astropy

18th
Feb 2016
Erik Tollerud
@eteq
Feb 18 2016 15:28
@keflavich: you around right now?
Adam Ginsburg
@keflavich
Feb 18 2016 16:31
aye
Erik Tollerud
@eteq
Feb 18 2016 16:31
@keflavich: are you planning to do the regression test for #4612 ?
Adam Ginsburg
@keflavich
Feb 18 2016 16:31
can't promise to; it is very important but I have some rush work to get done before heading home
Erik Tollerud
@eteq
Feb 18 2016 16:31
ah, ok - I’m hoping to double-check that this isn’t a problem in 1.0.x
although if you’ve checked that then my job there is done ;)
Adam Ginsburg
@keflavich
Feb 18 2016 16:32
hmm
well, a regression test will definitely answer that, but I have not checked directly. I think I would have caught it if such a problem existed, though
Erik Tollerud
@eteq
Feb 18 2016 16:33
ok, that’s what I was thinking. It’s also a bit hard to tell from the WCSLIB changelog but it looks to me like it was a 5.x change that caused it in the first place
Adam Ginsburg
@keflavich
Feb 18 2016 16:33
I believe that's right - probably 5.1x
Erik Tollerud
@eteq
Feb 18 2016 16:34
can you let me know (either here or gchat) if you have to leave before getting the test in? That way I can try to quick hack it if need be but can use your test if it’s ready by then
Adam Ginsburg
@keflavich
Feb 18 2016 16:34
yeah
Erik Tollerud
@eteq
Feb 18 2016 16:34
:thumbsup:
Adam Ginsburg
@keflavich
Feb 18 2016 16:45
@eteq where should I PR it? 4612 or a new PR, linked to 4613?
Erik Tollerud
@eteq
Feb 18 2016 16:47
@keflavich: new PR
Adam Ginsburg
@keflavich
Feb 18 2016 16:50
k, #4614 is in, but it is a WIP
I want to get travis running on it because I don't trust my build env
…something is very weirdly inconsistent with this test....
Erik Tollerud
@eteq
Feb 18 2016 17:02
OK, lets see what Travis says. This might also be related to #4555 ? There, @nden points out that CDELT is a bit different in that you need to set certain flags to get maximal precision
If you think you can work out the fix to that (i.e. close #4555), feel free to bundle that into #4614. @nden might also be able to do it (but we all seem a bit pressed for time today/tomorrow…)