Hi,

> Hi there,
> 
> So... I finally got something working that resembles some degree of 
> debugable environment. I got Visual Studio Code working with the
> LLDB 
> debugger and a Vala Code plugin for Syntax Highlighting. This allows
> me 
> to step through the Vala code and inspect some of the variables.
> 
> If I also mangle the c-code a bit, I can also get insights to the 
> produced c-code - but usually that is not so interesting. See this 
> thread for details on that: 
> https://bugzilla.gnome.org/show_bug.cgi?id=734953

That's actually quite nice :)

> 
> Anyway... What I wanted to say is that, right now, I have one image 
> (more, actually, but I am concentrating on one for debug sake) for
> which 
> I cannot set the comment field. I can set the comment, and Shotwell
> will 
> show it, but upon next restart, Shotwell will remove that comment
> again.
> 
> So I thougt I would take this opportunity to see what is going on.
> So 
> here is where I'm at:
> 
> I have one photo which I added a comment to, restarted Shotwell with 
> debugger attached, set a breakpoint where the comments were altered
> on 
> the photo, and got a hit when Shotwell tries to remove the comment 
> again. Please see screenshot.
> 
> In the screenshot you can see that the current photo has the comment 
> "Test comment" and that the new comment being set is NULL. You can
> also 
> see the stacktrace coming from a photo monitor performing a re-
> import 
> (finish_reimport_editable). The job originates as a background job, 
> though, and I have not yet found where it originates.
> 
> My questions would be
> 
>   * What does the terminology "reimport" mean?

No idea, sorry

> 
>   * Where do these background jobs originate from?

They are created in src/PhotoMonitor.vala:1042, based on the
information created in check_for_editable_changes() in
src/PhotoMonitor.vala:419. Looking at the code this is because some of
the file info does not match (file size and mtime)
(src/db/PhotoTable.vala:1095)
> 
>   * If it means that data is re-imported from images on each start-
> up, then
>     a) Is this correct behaviour? If write-back (to image files) is
> not 
> enabled (which I have not (*) ) would the database not be the
> authority 
> on this info?

I would assume that as well.

>     b) Why does this (seemingly) only happen to some of my photos.
> I.e. 
> I could successfully alter the comment on another photo, and it was
> not 
> reverted.

No idea. You might be able to check that with
meson configure -Dtrace=monitoring,import

> (*) But I do have "Watch library" enabled.
> 
> And yes... I have found and read (some of) the documentation for 
> Shotwell on gnome.org. If the information is correct / up-to-date, I
> tip 
> my hat for good project documentation :-)
> 

Probably it isn't :-/

_______________________________________________
shotwell-list mailing list
shotwell-list@gnome.org
https://mail.gnome.org/mailman/listinfo/shotwell-list

Reply via email to