Repeatable freeze in PI 01.08.03.1123 under OSX Mavericks

[rgbtxus]

Hi, I just started a trial of PI a few days ago and seem to have run straight into a bug.  The following information should allow you to reproduce this bug.
Test system: MacBookProRetina 2.6GHz 4-core i7, 16GB ram, I doubt this matters - the system was not under memory or cpu pressure
To reproduce the bug:  (I have the steps close, but not exact, you may need to try these steps in different orders, but I have always been able to freeze PI within a minute or two using these steps)
1) load 2 master dark images (I can send you mine if it should matter, but I doubt it and I doubt the image content matters, I just happened to be using master darks from a QHY23)
2) STF them
3)use PixMath to create a new image that is one minus the other
4) STF new image
5) use image statistics with unclipped option (this seems to be key) and click around on the images
6) rerun the PixMath
7) freeze  -- oops

[Juan Conejero]

Hi,

I cannot reproduce this problem with our data sets on any platform. I have repeated your steps multiple times with different images on Mac OS X 10.9.5 (also on Windows 8.1 and Linux), and I've got no freeze or problem at all.

Could you upload the two images that are giving you this problem?

[rgbtxus]

Sure, I'm new here and do not know the procedure for uploading images.  Can you point me in the right direction?

[rgbtxus]

Actually, you can just grab them from my google drive folder using the two links below.  I am in the midst of something that will take about another 30 minutes but then I will see if I can document an exact series of steps to cause the freeze.
FWIW: my java info is :
Richard-Bratts-MacBook-Pro-4:2process rgb$ java -version
java version "1.6.0_65"
Java(TM) SE Runtime Environment (build 1.6.0_65-b14-462-11M4609)
Java HotSpot(TM) 64-Bit Server VM (build 20.65-b04-462, mixed mode)

URLs to the two master dark files:
https://docs.google.com/file/d/0B0D0ZS2PaFEEY2YwRXZrTG9ZVnc/edit
https://docs.google.com/file/d/0B0D0ZS2PaFEEdl94Z2ZRNGZmVms/edit

Thanks for the extremely prompt response!

[rgbtxus]

Ok, these steps (which may not be minimal) cause PI to freeze on my machine using the 2 images I gave links for in the prior message
1) load both images by dragging and dropping them in PI
2) activate STF and click on both
3) Pixel Math creating new image: 300s image - 60s image
4) STF that
5) load Process/Image/Statistics, uncheck normalized, check unclipped
6) apply to the three images in the order new, 60s, 300s (I doubt the order matters, but that is what I did)
7) execute the PixelMath again creating a fourth image on the screen
8) STF the result --- bang I'm frozen

[Alejandro Tombolini]

Hi, I made a try on linux and all was fine. Maybe is Mac related. Better wait Juan to test it.
Saludos, Alejandro

[rgbtxus]

Hi Alejandro,
Wierd.  Reliably fails on my machine.  Tried this basic approach 4 times (slightly different order each time), 4 application freezes requiring a force quit.  We'll see what Juan finds.  Let me know if I can do anything to help.  It is not like this is holding me up -- just seemed like something that should be reported.  My system is used heavily and is quite reliable so I'm not inclined to think the problem lies there -- but who knows, perhaps it does.
Regards,
Richard

[Juan Conejero]

Hi Richard,

Bug confirmed. You have discovered a rare race condition which happens on all platforms, with any image. We have found a minimal reproducible sequence:

- Open any image.
- Open the Statistics tool.
- Check Unclipped on Statistics.
- Check the Track View icon on Statistics.
- Press Ctrl+A (Cmd+A on OS X) - > freezed!

I'll try to fix this problem as soon as possible with an update to the Statistics tool.

In the meanwhile, the workaround is: Make sure that the Unclipped checkbox is unchecked when you press Ctrl+A/Cmd+A to apply an SFT AutoStretch function. Once the AutoStretch has been applied, you can check Unclipped if you want to see unclipped statistics. Alternatively, you can also unckeck the Track View icon on Statistics to prevent the problem.

Sorry for the inconvenience.

By the way, why do you provide Java Runtime versions? PixInsight does not use Java at all (it has a JavaScript engine, but this has nothing to do with Java).

[rgbtxus]

Juan,
No problem, happy I could help.  I am totally new to AP (just bought a telescope and camera last week and started a PI trial a few days ago) so did not know anything about the internals of PI.  Someone on astronomynet.com suugested (perhaps I misunderstood) that PI used java and suggested I provide java version info with bug report.
Richard

BTW: I really like what I've seen so far.  PI looks like a fantastic tool!

[Juan Conejero]

Hi Richard,

This bug is now fixed. We have just released an update for all platforms. Thank you so much for your help and patience.

[rgbtxus]

Thank you.  Very impressive responsiveness!!

[rgbtxus]

Juan,
FYI I just logged into the SW distribution system and did not see a new version (could that be because I'm running a trial?).  I'm not in a rush for this fix, I just thought I'd let you know since the facts seem to contradict your statement that a new update was just released.
Richard

[NGC7789]

My guess is that this is the kind of fix distributed from inside PI and not through the SW distribution site.

[Juan Conejero]

Richard,

Welcome to our automatic updates system! We regularly release updates to modules and scripts to fix bugs and provide new versions and improvements. These updates don't require installing a new version of PixInsight.

If you have PixInsight running, select Resources > Updates > Check for Updates from the main menu. Otherwise when you run the PixInsight Core application the available updates will be announced automatically (unless you have disabled this feature through Preferences).

[rgbtxus]

Thanks my rookieness is showing (just started AP and using PI a few days ago).  I did not realize there were two SW distribution methods.  Worked like a champ and found the updates.  Thanks for setting me straight!