X11 forwarding

[Gwarl]

Originally Posted by t4nk

You mean like Dungeon Crawl's webtiles?

Ah, I already set up a service to play angband in a browser a while back, but I ran into difficulties getting all of the features from dgamelaunch working and inexplicable segfaults when trying to use anything but vanilla. I wiped it recently to start again, I'm asking about the x11 stuff because if it had worked and included graphics it would be a cheap way to add support for tiles.

[gameplay appreciator]

It's not a good idea to allow public access to a X windows server.

[Pete Mack]

Originally Posted by gameplay appreciator

It's not a good idea to allow public access to a X windows server.

No lie. They shouldn't even be visible outside a firewall, let alone accessible.

[Gwarl]

Looks like the x11 thing is a non-starter, but does anyone know what the problem might be with spawning extra terminals? All of the subopts for -mgcu (i.e. -B) give me the same error

unable to prepare any 'display module'

I have the gcu ports for angband and poschengband running nicely in the browser (But I'll have to implement a login system before I make it public) but the extra terminals are a massive QoL improvement to playing angband - it should be possible so I want to do it.

You can get 'virtual terminals' in V if the main terminal is large enough, but I'd rather have the seperate terminals individually if at all possible.

[takkaria]

Originally Posted by Gwarl

You can get 'virtual terminals' in V if the main terminal is large enough, but I'd rather have the seperate terminals individually if at all possible.

If I understand you right, the virtual terminals are all the curses supports. There's no way from inside a piece of software using curses to open new windows - you're just given a single grid of cells to update.

[Gwarl]

so this:
http://pubs.opengroup.org/onlinepubs...es/newwin.html
refers to what I called 'virtual windows' in my last post?

Odd that it refuses to run with any of the -mgcu subopts if so.

Usage info for -mgcu is as follows:

* angband -mgcu -- [-a] [-b] [-B] [-nN]
*
* -a Use ASCII walls
* -b Big screen (equivalent to -n1)
* -B Use brighter bold characters
* -nN Use N terminals (up to 6)

But I know from experience and reading through old threads here that there's only support for 4 terminals using a single curses window. So where is the 6 coming from there?

I could probably eventually hack what looked like real subwindows out of it, I can at least try and make sure the virtual subwindows are available though.

[Pete Mack]

For this T4NK's new terminal would be best. It uses panels rather than screens, which is easier for a web interface. In any case, i'd recommend sticking to a single screen.

[takkaria]

Originally Posted by Gwarl

so this:
http://pubs.opengroup.org/onlinepubs...es/newwin.html
refers to what I called 'virtual windows' in my last post?

Yes, curses windows are 'virtual' windows - curses has to operate on any text output, including ones where there are no graphical windows.

[t4nk]

Originally Posted by Gwarl

Looks like the x11 thing is a non-starter, but does anyone know what the problem might be with spawning extra terminals? All of the subopts for -mgcu (i.e. -B) give me the same error

unable to prepare any 'display module'

You're getting strange errors... The only thing I can think of is you're somehow entering "-mgcu -- -n5" as a single string, so Angband tries to find a module called literally "gcu -- -n5" (instead of "gcu")?

[Gwarl]

Yup, you were right it was just my syntax. gcu subopts no longer cause a crash.

To get this clear though, to have angband run in multiple windows requires either the x11 port (which seems defunct?) or the windows port? I know SDL happens on a single screen and the subwindows there are 'virtual'.

How do people do it on linux? Does x11 work if you're running it locally?

I'll make a note to look at textui2 at some point, skimming the thread it seems like a lot of the problems it solves would be exacerbated over a network.

Could it be compiled into other variants? I think the ui code it's replacing is quite old, but I don't know if it's *that* old... Both the O and Z lines forked from 2.9.3 IIRC.