--- Comment #8 from
daedalus...@gmail.com 2012-10-06 07:21:52 UTC ---
(In reply to comment #7)
> Whenever you open a document, evince launches a new window.
>
> So, you would need to consider that situation.(In reply to comment #4)
> > Personally, I prefer using CTRL+O to open up a file dialog (and I would imagine
> > most advanced users would also). In fact, I hardly ever use the menu bar.
>
> Still, the evince behaviour is to open any new file in a new window, not matter
> if the current one does not a document.
What do you mean by a "new" window? If I start a new Evince session (with a
blank window), and then I open a file, the file appears in the 'same' window.
Furthermore, if the original (empty) window was fullscreen, then the window
with the new file is also fullscreen, so even if the underlying mechanism
actually does 'launch a new window', it still transfers a nontrivial amount of
the state from the 'old' window into the 'new' one.
Lastly, I don't see how this pertains to my point.
> > Additionally, I think it would be good to change it because:
> > 1. Consistency: having --fullscreen deviate from the usual behaviour in this
> > one case increases complexity.
>
> A fix for that is to require an argument (filename) when using this option.
This makes sense, assuming you also disable the option to fullscreen an empty
window once Evince is loaded. In essence, my point is:
1. Allowing presentation/fullscreen display if and only if a file is loaded is
consistent (provided you explicitly state this fact in the man page).
2. Alternatively, allowing presentation/fullscreen display at all times is
consistent.
3. Allowing fullscreen not from the command line without loading a file, but
allowing it from within the GUI without a file, and allowing it at all times
with a file loaded is unnecessarily complicated.
Alternatively, I feel like the two statements are equivalent:
1. There exists a use case for allowing fullscreening an empty window while the
application is loaded.
2. There exists a use case for allowing fullscreening from the command line
without a file.
> > 2. User intention: If the user explicitly specified --fullscreen, then I think
> > we shouldn't make the decision for them to un-fullscreen it. (If they really
> > made a mistake, then they can disable fullscreen anyways.)
>
> If the user can explicitly launch in full screen, why the user does not provide
> the file name at the same time?
I'm not interested in what the user 'can' do, but what the user did explicitly,
verbally, specify. (However, this issue can be resolved if we define launching
fullscreen without a file as invalid, and inform the user about this fact.)