Client rewrite β-testing

[loftar]

Found a very easy way to reproduce this after looking at thread dump: aggro an ant, start walking away and furiously spam all your combat buttons and mouse clicks. Client hangs without even actual combat.

This should be fixed now. Thanks for the useful report!

[ubersheva]

I haven't seen instantion-ids error since the last update, so I'd assume that it's fixed.

[ubersheva]

I haven't seen instantion-ids error since the last update, so I'd assume that it's fixed.

crash on stone milestone right-click

Code: Select all

java.lang.NullPointerException
	at haven.Widget.setfocus(Widget.java:549)
	at haven.Widget.uimsg(Widget.java:633)
	at haven.Window.uimsg(Window.java:237)
	at haven.UI.uimsg(UI.java:309)
	at haven.RemoteUI.run(RemoteUI.java:70)
	at haven.MainFrame.run(MainFrame.java:257)
	at java.base/java.lang.Thread.run(Thread.java:830)

[/spoiler]

[loftar]

Code: Select all

java.lang.NullPointerException
	at haven.Widget.setfocus(Widget.java:549)

Right, thanks! Should be fixed now.

On that note, as far as I'm aware all the critical bugs that I know of should be fixed now, so unless anyone can remind me of something to that effect, I'm very strongly considering switching to the new client very soon. I'm still having some minor trouble adapting the avatar rendering on the website to use the new client, but as soon as that is done, I'd like to try switching. Yes, there are still some minor graphical bugs left, but I'm not aware of anything so grave that I can't fix it after switching.

As for custom clients, the current resource URLs and similar things shouldn't stop working as part of the switch, so they should keep working transparently. However, I'd really, really like to stop maintaining dual resource trees as soon as possible.

[ubersheva]

Another one: the red circular band that indicate mining support radius does not allow right-clicking through it.

As for resources url, please make sure that all the mapping data, along with notes, is not lost when you make the switch. Ideally maps and notes saved under both clients should be merged.

[Oddity]

I'm still having some minor trouble adapting the avatar rendering on the website to use the new client

The website uses the game client?

[boshaw]

However, I'd really, really like to stop maintaining dual resource trees as soon as possible.

Can there be a few week notice as to when this is happening and what the future urls will be if they are changed?

[loftar]

The website uses the game client?

Well, yeah, it seemed far simpler to just reuse what I had instead of writing a completely separate rendering system for the website. In any case, I did eventually manage to solve that problem, but it seems that I, somewhat comically, fell prey to my own question of upgrading the client to OpenGL 3, because the software renderer I'm using on the server (since it lacks a GPU) only supports OpenGL 2. :/

I do have a few options for solving that problem, though, so I expect to switch in the coming days unless none of those turn out to work for some reason.

please make sure that all the mapping data, along with notes, is not lost when you make the switch. Ideally maps and notes saved under both clients should be merged.

Indeed that is a bit of a conundrum. My plan is to ensure that the new client, as permanently installed, will be using the mapping data that the current (ordinary) client uses, but that will indeed leave the mapping data accrued by those who've used the test client inaccessible (which includes me). The real solution would probably be to implement map import/export, but I'm not sure I want to wait for that to switch the client over. There are ways around it for those who can compile themselves, though.

Can there be a few week notice as to when this is happening and what the future urls will be if they are changed?

I'm not sure how to handle that, honestly. I mean, when I've switched the official client over, I'll pretty much just be waiting for custom clients to adapt before I stop maintaining the previous resource tree. I could set a deadline, of course, but from my perspective it would be completely arbitrary, and I'm not sure if it helps or hurts.

[boshaw]

The website uses the game client?

Well, yeah, it seemed far simpler to just reuse what I had instead of writing a completely separate rendering system for the website. In any case, I did eventually manage to solve that problem, but it seems that I, somewhat comically, fell prey to my own question of upgrading the client to OpenGL 3, because the software renderer I'm using on the server (since it lacks a GPU) only supports OpenGL 2. :/

I do have a few options for solving that problem, though, so I expect to switch in the coming days unless none of those turn out to work for some reason.

please make sure that all the mapping data, along with notes, is not lost when you make the switch. Ideally maps and notes saved under both clients should be merged.

Indeed that is a bit of a conundrum. My plan is to ensure that the new client, as permanently installed, will be using the mapping data that the current (ordinary) client uses, but that will indeed leave the mapping data accrued by those who've used the test client inaccessible (which includes me). The real solution would probably be to implement map import/export, but I'm not sure I want to wait for that to switch the client over. There are ways around it for those who can compile themselves, though.

Can there be a few week notice as to when this is happening and what the future urls will be if they are changed?

I'm not sure how to handle that, honestly. I mean, when I've switched the official client over, I'll pretty much just be waiting for custom clients to adapt before I stop maintaining the previous resource tree. I could set a deadline, of course, but from my perspective it would be completely arbitrary, and I'm not sure if it helps or hurts.

What if to avoid breaking maps and having to come up with an import/export solution or waiting for custom clients to adapt we just start W13 off with the new render client and removal of the old resource tree? Altho that may be a long time away I guess.

[loftar]

What if to avoid breaking maps and having to come up with an import/export solution or waiting for custom clients to adapt we just start W13 off with the new render client and removal of the old resource tree? Altho that may be a long time away I guess.

That would indeed be much too far off. I'm also not sure why that would even help. Why would it be easier for custom clients to adapt at the start of a new world than in the middle of one?

Also as far as breaking maps, I'm not really too worried. The number of people who've been using the test client and don't compile themselves seems to be pretty limited.