Angband Forums

Angband Forums (http://angband.oook.cz/forum/index.php)
-   Vanilla (http://angband.oook.cz/forum/forumdisplay.php?f=3)
-   -   Current master post 4.2.0 (http://angband.oook.cz/forum/showthread.php?t=9659)

DavidMedley April 16, 2020 18:07

What if the throwing menus were separated like inv/eq/quiv? Like, you hit v to throw and the first menu it brings up has only things that are made for throwing. If there's nothing on that screen, or if you press left or right, there would be different lists, probably the same inv/eq/quiv lists we already have.

Nick April 16, 2020 23:56

Quote:

Originally Posted by DavidMedley (Post 144612)
What if the throwing menus were separated like inv/eq/quiv? Like, you hit v to throw and the first menu it brings up has only things that are made for throwing. If there's nothing on that screen, or if you press left or right, there would be different lists, probably the same inv/eq/quiv lists we already have.

That's an interesting idea.

Diego Gonzalez April 17, 2020 00:23

Then objects (flask of oil, throwing hammer, etc) should have an specific flag in the txt files for this.

Nick April 17, 2020 00:43

Quote:

Originally Posted by Diego Gonzalez (Post 144617)
Then objects (flask of oil, throwing hammer, etc) should have an specific flag in the txt files for this.

There is already a flag for throwing weapons that should work for this.

DavidMedley April 17, 2020 16:45

Playing the nightly (4.2.0 may be the same) and there was a spear under webs. I couldn't find any indication it was there except for the object window and moving over it. Seems like if it's hidden from sight it should be hidden from the object window, or if it's visible in the object window it should be visible on screen or at least to the look command. Or possibly I remembered it was there from before the webs covered it, but I still feel the same principle should apply.

DavidMedley April 17, 2020 21:48

What are the current plans for the next release? ETA, will it be 4.2.1 or 4.3, anything that needs to be in it that's not in yet, etc.

Nick April 17, 2020 23:50

Quote:

Originally Posted by DavidMedley (Post 144634)
Playing the nightly (4.2.0 may be the same) and there was a spear under webs. I couldn't find any indication it was there except for the object window and moving over it. Seems like if it's hidden from sight it should be hidden from the object window, or if it's visible in the object window it should be visible on screen or at least to the look command. Or possibly I remembered it was there from before the webs covered it, but I still feel the same principle should apply.

Yes, this is actually kind of tricky, I'm not sure of the best solution.

As for next release - once the nightlies are bug free is the general idea. With the new blackguards I wanted a little time for bugs to emerge and minor changes to be made too. Bugs are down under 20 now, so I think around a month is possible - but it could be longer, or theoretically shorter. And it will be 4.2.1 - the general idea being that 4.2 was a major change, and the 4.2.x series is essentially improvements on that, and maintains savefile compatibility.

DavidMedley April 18, 2020 00:35

Sounds good. Thanks for the update.

Diego Gonzalez April 18, 2020 02:10

Perhaps this is a problem of the android port but I'll report a possible bug here, just in case.

When inspecting Google's Play Console I see that there are several crashes in findpath, from anonymous users.

After trying to dissect the ndk coredump, the ndk-stack tool points me to this line:

Code:

pf_result[pf_result_index++] = '0' + (char)(10 - dir);
The code doesn't check for array overflow. This is a bug in Vanilla ?

I'll try to reproduce the bug.

Nick April 18, 2020 02:31

Quote:

Originally Posted by Diego Gonzalez (Post 144655)
Perhaps this is a problem of the android port but I'll report a possible bug here, just in case.

When inspecting Google's Play Console I see that there are several crashes in findpath, from anonymous users.

After trying to dissect the ndk coredump, the ndk-stack tool points me to this line:

Code:

pf_result[pf_result_index++] = '0' + (char)(10 - dir);
The code doesn't check for array overflow. This is a bug in Vanilla ?

I'll try to reproduce the bug.

A bug in this code would not surprise me at all. I'll do a check too.

EDIT: OK, one massive bug found here already. I'm rewriting the whole thing.


All times are GMT +1. The time now is 21:23.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, vBulletin Solutions Inc.