Page 1 of 1

SVN 5776

Posted: Thu Aug 18, 2011 11:47 pm
by Azhrei
I've committed some more code updates. You can right-click on the maptool project and choose Team > Show History to see what the changes have been. In the mean time I'll distill the commit history into an entry on the wiki's ChangeLog page as time permits.

I'm particularly interested in testing of the saving-at-exit-while-auto-save-is-running-results-in-no-save-file bug. The autosave routine runs in a separate thread so it's difficult to reproduce every possible interaction. Please set your autosave duration to 1 minute and then load a large campaign and when you exit, try saving the campaign. The goal is to get the autosave to kick in while MapTool is saving and exiting, or to have MT try to save while the autosave is already running. Please do this on a backup copy of your campaign!! I'd hate to have you lose something important with this one!

I still have a few other bugs I want to fix, but I'd like to get b87 out because it fixes the Java 7 problem (or "might", at least). And we're getting a lot of reports of that particular error from users. The other bugs are things that can be continued to work on (meaning there *might* be a b88) but these are pretty important.

Thanks!

Re: SVN 5777

Posted: Fri Aug 19, 2011 12:40 am
by Azhrei
Update: Added another bug fix. Up to SVN 5777.

This is for the bug in which one or more layers of the map weren't displayed right after loading a campaign; instead, the user had to force the map to be regenerated, such as by zooming in or out.

Re: SVN 5776

Posted: Fri Aug 19, 2011 9:14 am
by neofax
If anyone wants me to make a build of this so they don't have to install Eclipse and such, let me know. I plan to make one anyhow so I can test this tomorrow.

Re: SVN 5776

Posted: Fri Aug 19, 2011 9:19 pm
by patoace
Any action about this bug?

viewtopic.php?f=3&t=18735

Re: SVN 5776

Posted: Fri Aug 19, 2011 10:19 pm
by Azhrei
Sorry, I posted this to my Google+ stream but not here. :(

This is not going to be fixed. The entire undo/redo function is pretty messed up. While it works in the simple case, there is no synchronization of what's in the queue between machines (which I think is where your symptoms are coming from). Unfortunately the whole implementation is wrong -- it's not synchronized across machines, it's not per-layer, per-map, or per-client, and the structures used to maintain the list are not easily adapted to any of the above.

So the only change I've made is to ensure that "Clear All Drawings" is synchronized across all clients (it wasn't before) but that's all that will happen for 1.3.

Folks in this forum are expected to have access to the SVN repository and the list of changes is in the history there (and I'll get them into the change log this weekend) so that's where to find the discussion of what has been fixed. I likely won't be responding to individual questions about this-or-that bug in this thread since that's not what it's for. :|

Re: SVN 5776

Posted: Sat Aug 20, 2011 10:42 am
by Azhrei
Btw, there's nothing in b87 that breaks compatibility with b86, even when connected to it during a game. I plan to use the development version (from within Eclipse) for my game tonight since DEVELOPMENT versions will connect to any other version...

Re: SVN 5776

Posted: Sat Aug 20, 2011 3:00 pm
by patoace
Azhrei wrote:Folks in this forum are expected to have access to the SVN repository and the list of changes is in the history there (and I'll get them into the change log this weekend) so that's where to find the discussion of what has been fixed. I likely won't be responding to individual questions about this-or-that bug in this thread since that's not what it's for. :|

Ok, I'm sorry :oops:

Re: SVN 5776

Posted: Sun Aug 21, 2011 12:16 am
by Azhrei
Another patch committed. It appears that I didn't include the xmlpull library dependency in the build path that was committed. So MapTool 5778 is the latest, along with rplib 5779.

(Both needed the new XML library since they both use XML load/save files and both were updated to the latest XStream library and that's where the dependency comes from.)