[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: new version buggy

> Julian Taylor <jtaylor.debian@googlemail.com> writes:
>> the latest version of swarp has introduced regressions related to file
>> descriptors and weight maps. I would advise against updating before
>> having them fixed.
> I uploaded the new version to experimental, and fixed (as far as I
> understand it) the bug #748970. Since this version is past the ChangeLog
> entry that mentiones you (see my other mail), I could imagine that some
> or all of your regressions are gone.
> Can you test this please and open a bug report if this is not the case?
> If not, I would like to pass this package into Ubuntu as well to appear
> in 15.04.

assuming you have taken the newest version (r145) the issues are not fixed.
The issue mentioned in the changelog was a fix I submitted to solve the
file descriptor issue but it was promptly broken again in a worse way in
a latter commit (I think r144).

the weight issue only appears if you use WEIGHT_TYPE MAP_RMS and is a
regression introduced in r125, so not related to the forum post.

> When WEIGHT_TYPE is MAP_RMS the threshold is treated differently in coadd.c and weights.c
> In weights.c:weight_to_var the rms is loaded and compared against a weight, while in coadd.c:1984 the data is loaded as a weight and compared against the weight.
> This always causes one of the two paths to always reject all pixels. 

I'll try to find my testcase later, it seemed to have disappeared in the
depths of my harddisk.

Also please keep me in CC. I'm not subscribed to list or bug.


Reply to: