×

INDI Library v2.0.7 is Released (01 Apr 2024)

Bi-monthly release with minor bug fixes and improvements

Produced .fits files by EKOS / Kstars do not follow the fits file specification

  • Posts: 29
  • Thank you received: 1
I only just noticed this because I am implementing a .fits file library in C# when I was using my own .fits produced through INDI /EKOS for testing.

The fits file standard specifies, that the file always contains blocks of 2880 in length, meaning the file size should be a multiple of 2880.
Now I noticed that something was off when testing my library using fits files produced by EKOS / Kstars.

The file I was using for testing, only had the Primary HDU, however, the file still had bytes after the last 2880bytes block from the images content.
So my library went ahead and thought "oh nice, here comes more data / a new HDU" but the remaining block was less than 2880bytes in length.

I then uploaded some fits files from EKOS / kstars to the online validator found here: fits.gsfc.nasa.gov/fits_verify.html
And sure enough it throws an error:
*** Error:   File has extra byte(s) after last HDU at byte 32803200.

Now I can simply do a check in my own fits reader to check if the remaining file bytes are not of at least the length of a new 2880 byte block but this seems kind of strange.
1 week 6 days ago #100377

Please Log in or Create an account to join the conversation.

  • Posts: 29
  • Thank you received: 1
Also for some reason I can not create an account at KDE to report it at their git repo directly. It keeps rejecting my client due to anti spam measures
1 week 6 days ago #100378

Please Log in or Create an account to join the conversation.

  • Posts: 362
  • Thank you received: 114
Then this may be problem with cfitsio library because that is used to write FITS files.
1 week 5 days ago #100396

Please Log in or Create an account to join the conversation.

  • Posts: 29
  • Thank you received: 1
1 week 5 days ago #100398

Please Log in or Create an account to join the conversation.

  • Posts: 362
  • Thank you received: 114
Tried this with simulator but I can't reproduce it. My FITS file is multiple of 2880 and I can see zero padding at end of the file.
1 week 5 days ago #100402

Please Log in or Create an account to join the conversation.

  • Posts: 29
  • Thank you received: 1
That's interesting, every single fits file I have from EKOS (and I have 1000's) that I have tested had this issue.

That would imply that something else is causing this, maybe the filesystem?
1 week 5 days ago #100404

Please Log in or Create an account to join the conversation.

  • Posts: 362
  • Thank you received: 114
Just checked my archive. I have around 9000 files that have correct 2880 size multiple and 987 that are not. All bad ones were produced around august - september 2022. No idea why it happened. Files after that are good.
Last edit: 1 week 5 days ago by nou.
1 week 5 days ago #100406

Please Log in or Create an account to join the conversation.

  • Posts: 29
  • Thank you received: 1
Just went through my fits files, it appears that the issue started somewhere around Mid 2022, fits before that are fine, but fits files that are younger all have this issue.
The "youngest" fits file I tested it with is one from sept. 2023 which also had this issue.
1 week 4 days ago #100417

Please Log in or Create an account to join the conversation.

  • Posts: 302
  • Thank you received: 46
The version I have installed is 4.0.0 and the latest version from github is 4.4.0 since about 2 months ago.

Br,
/Markku
1 week 4 days ago #100421

Please Log in or Create an account to join the conversation.

  • Posts: 212
  • Thank you received: 104
The date correspond to this old issue: github.com/indilib/indi/issues/1719
It was quickly fixed and is probably present only in INDI version 1.9.7
1 week 4 days ago #100425

Please Log in or Create an account to join the conversation.

Time to create page: 0.685 seconds