Pale Moon forum

I know there's probably a genuinely valid reason for this, but the fact is, pale moon takes a while to free itself from memory after it's closed.
If I close then remember something and reopen the browser 5-10 seconds later, I don't want to have to click on the 'ok' for the "pale moon is already loaded but is not responding" dialog, then wait another 10 seconds.
Just let it load a second instance?
Thanks,
Matt

Hello Matt!

I think that this because Pale Moon just doesn't close instantly when you quit. But also it needs to clean the plugin container and RAM cache, so this time may vary depending of your session. That's it, if you see a palemoon.exe or plugin-container.exe for long time after you closed Pale Moon, then that is a hang and may need to be investigated.

Other browsers will do the same thing, a certain amount of housekeeping has to be done when the browser is closed. (I recall FF generally taking longer actually)

"Just letting it load a second instance" isn't possible. You'd have 2 programs trying to read and write the same files at the same time in your profile.
If you close the browser only to immediately reopen it again, why did you close it in the first place?

Of course it's possible - just have it poll until the second instance isn't detected. Simple as that.

metamorphosis wrote:Of course it's possible - just have it poll until the second instance isn't detected. Simple as that.

And risk profile corruption.. Nah.. RESOLVED INVALID

metamorphosis wrote:Of course it's possible - just have it poll until the second instance isn't detected. Simple as that.

If it's that simple according to you, feel free to experiment with your own builds -- and if you can find a robust way to:

1. Determine what to poll (people may be having multiple instances running, having it running under a different user, etc.)
2. Do this cross-platform
3. Determine how and how often to poll
4. Deal with disk cache latency after closing (locked files)
5. Not leave the user "hanging" without feedback when profile cleanup maintenance takes longer than expected

then I'm very interested in seeing your patches

Moonchild wrote:

metamorphosis wrote:Of course it's possible - just have it poll until the second instance isn't detected. Simple as that.

If it's that simple according to you, feel free to experiment with your own builds -- and if you can find a robust way to:

1. Determine what to poll (people may be having multiple instances running, having it running under a different user, etc.)
2. Do this cross-platform
3. Determine how and how often to poll
4. Deal with disk cache latency after closing (locked files)
5. Not leave the user "hanging" without feedback when profile cleanup maintenance takes longer than expected

then I'm very interested in seeing your patches

Oh come now. You're already detecting that it's running or not running. That's why you're displaying that message. It's simply a matter of doing that, not displaying a message, waiting a second, doing the same detection, rinse and repeat until undetected. But at least don't be sarcastic please, it's unbecoming.

So your solution would be to leave the user hanging. Nope, sorry, not gonna happen.

Moonchild wrote:So your solution would be to leave the user hanging. Nope, sorry, not gonna happen.

Yeah, that's what you're already doing, except the user has to repeatedly open and click your message and wait, whereas the program would do so automatically and waste less time. It's up to you, but don't be misleading about it.

metamorphosis wrote:It's up to you, but don't be misleading about it.

I'm not being misleading about anything, so please don't make it personal. There is some actual logic behind doing it this way.

There are a few other options in a slow shutdown and user restart situation, all of which will interrupt the user one way or another.
Not closing the main window until everything has been cleaned up upon shutdown is one - people will think the program is unresponsive/hangs/has crashed.
Letting the user hang on startup (your proposal) - people will think the program fails to start, may even try to start a second copy; that may end up waiting for each other -- perfect race condition.
Attaching to the process already found running -- not a viable option because it's in a shutting-down state that cannot be cancelled.

So, informing the user that the process has been found in a still-running state is the sane thing to do. If this takes enough time in your case that you can make several attempts at launching the browser before it's done, then you may need to look into cleaning your history, session, extensions, and other factors that may influence the updating/cleanup of work files in your profile folder.

If you so rapidly want to exit and restart the browser in your workflow, then you should think of a different solution that doesn't actually shut down the browser at all avoiding your "minor gripe".