recent regression in fitsdata

[Jasem Mutlaq]

Hello Hy,

Thanks for the report. The last commit tries to improve updateFrame as I
found it was getting multiple times (sometimes 3 times in quick succession)
even though this function is CPU intensive. A throttle was added that would
wait 250ms before it executes the actual call.

Testing yesterday with the units tests and just regular testing in all
modules didn't show any obvious bug, but it would be great to get more
feedback on this. Are there any circumstances where this doesn't work? I
suppose one condition is that if the updateFrame is called every 250ms or
less then it never fires, i.e. in case of a loop, but I haven't found a
case where loading FITS + processing takes less than 250ms. So there might
be room there for more improvement.

--
Best Regards,
Jasem Mutlaq



On Sat, Mar 13, 2021 at 9:01 AM Hy Murveit <murveit at gmail.com> wrote:

> Looks like I was missing the last two commits from a couple hours ago, and
> Jasem has already
> fixed the issue--somehow processData() got commented out.
>
> Thanks Jasem!
>
> On Fri, Mar 12, 2021 at 9:03 PM Hy Murveit <murveit at gmail.com> wrote:
>
>> I noticed imaging tonight, something I haven't done for a bit, that the
>> fits statistics
>> aren't being computed for my captures. No HFR, no median, no
>> num-stars-in-image, ...
>>
>> Without changing any parameters, I rad
>>   git checkout 66d10a8241b21f35ff5e95b999207835d7dce5a3
>> recompiled and ran, and all was well.
>>
>> This seems to be the culprit
>>
>> Thanks,
>> Hy
>>
>> commit 66d10a8241b21f35ff5e95b999207835d7dce5a3
>> Author: Jasem Mutlaq <mutlaqja at ikarustech.com>
>> Date:   Thu Mar 11 02:41:06 2021 +0300
>>
>>     When Adaptive Sampling is on, m_PreviewSampling is changed
>> dynamically depending on available memory. This MR fixes an issue whereas
>> m_PreviewSampling keeps increasing exponentially until it resets to zero
>> which then causes KStars to crash. This also fixes scaling issue when
>> dealing with large images with overlay since m_PreviewSampling can be
>> larger than 1.
>>
>>
>>
>>
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kstars-devel/attachments/20210313/95086714/attachment.htm>