Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

To do May 2024 #206

Open
9 of 15 tasks
mtoeltsch opened this issue May 18, 2024 · 12 comments
Open
9 of 15 tasks

To do May 2024 #206

mtoeltsch opened this issue May 18, 2024 · 12 comments

Comments

@mtoeltsch
Copy link
Contributor

mtoeltsch commented May 18, 2024 •
edited

Martin Toeltsch
This list is just to manage the open issues (18-May-2024).

  • Append a number to output file not working.
    Link
  • If a colour TIFF file contains CFA exif data, it's interpreted as Bayer CFA. This causes a crash.
    Link
  • Adding a new file-group sometimes results in wrong group name.
  • The saved file list sometimes contains too few groups or wrong group names.
  • The entire EditStars functionality (including edit comets) is buggy.
    The green rectangles jump around in a weird manner, after some time DSS crashes.
    Link to mailing list
  • The magenta circle around the comet stays even after removing the image file.
    This seems to be a severe problem, as Bruce Sobut reports here.
    DSS is caught in comet-mode after loading a file-list with comet, calculating offsets fails (according to Bruce). The only way out is to load a file-list without comet.
    However, Bruce used a file-list with "no alignment" set, which was the reason for setting all offsets to "0.0". Also all other problems reported by him were user errors (he mistakenly defined a comet in one of his light-frames, etc.).
  • Occasional crashes in Align RGB Channels.
    Link to mailing list
  • Bruce Sobut reported still an inconsistent result file name (Autosave instead of <file list name>.
    Link to mailing list
    This is low priority, as Bruce was not able to reproduce it, nor was I.
  • 2 instances of DSS open, save a file-list into a directory, the other instance also defaults to this directory, even with unrelated projects.
    Link to mailing list
  • Comet+stars stacking: Remove remaining faint star trails in/behind the comet tail.
  • Bruce Sobut gets a reproducible crash (STACK_OVERFLOW) when registering a JPG file.
    This is his log file: DSSTrace
    Unfortunately, he cannot find the JPG file anymore.
    Fix: It was caused by a call to Progress2() from other than the main thread. Fixed in V 5.1.6.
  • Request for lower registration threshold than the 2%
    Link to request on mailing list by André Holzmeister
    User Jesse Goldba provided an example On Dropbox
  • Request by Bruce Sobut to insert a '-' between fileListFile and the increasing number.
  • Request by Bruce Sobut: When loading a file list when files already loaded, a reminder that files are loaded or a "Do you wish to clear existing files?" Link to request
  • Multiple reference frames can be selected in the file listing.
@perdrix52
Copy link
Member

perdrix52 commented May 18, 2024 via email

@perdrix52
Copy link
Member

perdrix52 commented May 18, 2024 via email

@mtoeltsch
Copy link
Contributor Author

mtoeltsch commented May 18, 2024 •
edited

In Comet editing mode the Green rectangle snaps to existing stars unless you use the Shift key which allows you to click anywhere. David

If you test it with that file:

https://groups.io/g/DeepSkyStacker/attachment/31369/1/C-2023%20A3%20_Tsuchinshan-ATLAS__2024-05-16_22-23-47_60.00s_ISO200_0.00_0030.nef

you'll notice that the green rectangle snaps e.g. to a star in the upper left corner while the mouse is in the bottom right part of the image. If you then mark it as a comet, it won't let you un-mark it anymore. With other stars, it works better, so it's a mixture of "normal" and "weird" behaviour.
After some marking/un-marking actions, DSS crashes with an EXC_ACCESS_VIOLATION, because from the vector EditStars::stars the item with index 2 shall be erased while the vector has a size of only 2.

This is an example: Instead of offering me the option to add the star near the mouse, it would let me remove a star at a completely different position.

grafik

@perdrix52
Copy link
Member

perdrix52 commented May 18, 2024 via email

@mtoeltsch
Copy link
Contributor Author

BTW, it's exactly the same in DSS 4.2.6, so that bug is not new.

@perdrix52
Copy link
Member

perdrix52 commented May 19, 2024 via email

@mtoeltsch
Copy link
Contributor Author

I'm in the middle of the work to debug the comet editing mode.
Question: There's a global bool g_bShowRefStars = settings.value("ShowRefStars", false).toBool();
It's always false, and in DSS there's no setting whatsoever to turn on a "Show Reference Stars" option.
There's a lot of code depending on this bool g_bShowRefStars, which is actually dead code. Can I remove it?

@perdrix52
Copy link
Member

perdrix52 commented May 22, 2024 via email

@mtoeltsch
Copy link
Contributor Author

I was able to fix the "rectangle jumping around" issue, it was a serious bug in function FindNearestStarWithinDistance().
But still the logic is incorrect, e.g. when removing one star, the star count falls from 12 to 9. After some time, there's still an access violation.
I'll continue to search for the problem, but the function EditStars::draw() is rather complex and obscure.

@mtoeltsch
Copy link
Contributor Author

I fixed the entire EditStars and CometMode issues. I streamlined the whole EditStars class to a large extent, but left all the old code in - just commenting it out what's not needed anymore.

@mtoeltsch
Copy link
Contributor Author

But now I notice that there is an almost 100% code duplication in file ImageSinks.cpp, the class is CEditStarsSink.
I did not change anything in ImageSinks.h/.cpp or so.
Question: What's the purpose of these ImageSink stuff? Do you think there's a possibility to remove the whole code duplication by e.g. just re-using the EditStars class for this?

@perdrix52
Copy link
Member

The ImageSinks code is almost ready for deletion. It is the old code, and I think there's one small piece of code that still uses it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@perdrix52 @mtoeltsch