Trunk discussion / complains


#301

Okay, this probably means we have to download and place correctly some files, as atm map crashes at launch.


#302

That’s strange. Which file(s) are missing?

What we had in core.zip before:

  • new creature window
  • new quest window
  • stack queue for battles
  • some bonus icons
  • extra resolutions, including extra buttons for adv. map
  • test map

First three are already in git - so vcmi shoud work without extra graphical package.

However non-standard resolutions would need that package and enabling its mods. (they should appear in Launcher as submods of “vcmi” mod)

Oh - and I guess we should add fallback “if selected resolution was not found - fall back to 800x600”


#303

Thanks for clarification.

Forget about previous post. After merge I’m getting crash when launching Arrogance with default settings. Logs show something about missing player info.
This is so weird I don’t know where to look, it doesn’t seem any merged code could cause such behavior :?

I’d also like to split “develop” branch into two: one where other branches are merged, tested and fixed and the other which actually compiles and works on every platform. This way the issues I have now wouldn’t stop main branch (and possibly everybody who uses it for update) from working.


#304

could you add an graphic scaler module?
for example it could run internally 800x600 graphics on 400x300 rectangle of screen (integer proportion scaling should go fast).

I’m sad because I have a linux (not Android, but true linux) phone, but it have screen of 800x480, which makes it can’t run VCMI

PS: of course if we have an 1600x960 resolution supported in VCMI half scaling would work on whole screen (1600/2x960/2=800x480)

PPS: how hard would be to implement this? (and i hope i used right topic on the forum)


#305

We had two major changes:

  • VCMI Dirs change. Which can affect server but not like that
  • My GUI branch. Which contains no/almost no change to lib/server. So it shouldn’t affect it either. Weird.

I’d wish we could set up automated builds for MSVC as well but apparently this is not an option :frowning:

We can try to create separate “stable” branch that will be used for releases & daily builds. But this means that we must constantly update it so I’m not sure if this is a good idea.

majaczek, I think SDL 2 has this feature already. And I think that AVS also enabled it but you’d better ask him about this to be sure. IIRC there were some problems with this.


#306

Apparently player issues were just incorrect logs (though they were not present before).

Actual crash occurs at this line:

	// Create enough room for the whole map and its frame
	ttiles.resize(sizes.x, frameW, frameW);

Will investigate it now. Debugger gets confused with this code and doesn’t show correct info.


#307

With SDL2 vcmi scaled in fullscreen mode to desktop resolution (but with Opengl scaling in not proportional - cant fix this, it comes from SDL internals.).


#308

@up yay! thanks!

could you add a deb with trunk vcmi to maemo distribution repository (there is one but it’s old 0.86 with custom changes for the resolution or something) ?


#309

Yes, now I’m certain the crash occurs at the line mentioned above. Looks like uninitialized values:

11:57:19 INFO animation [1b94] - CMapHandler::roadsRiverTerrainInit(): frameW 27378054, frameH 15239566, sizes(36,36,2)

#310

Wait a minute - are you using non-800x600 resolution? In this case - do you have that package I posted above extracted and enabled? Because othervice you will have a crash due to no longer present resolution - in trunk we now have only 800x600 resolution, extra resolutions are in separate package since I’d rather avoid derivative graphics in repository.

I guess now I need to implement fallback to 800x600 resolution if selected resolution is missing. Will do it today.

Moved some posts here. Ivan forum.vcmi.eu/viewtopic.php?p=13147#13147


#311

Cann’t compile developer branch rev f78821. Is ERM still supported?


#312

Kind of. ERM module was never actually used by engine so by default ERM is disabled. As result - from time to time things like this may happen. Create pull request if you wish to fix this but we can’t use ERM till proper scripting support.


#313

Hi,

I’m a new “user” of vcmi and started with latest release on ubuntu. Due to segfaults and very long time at each “end of turn” did I switch to the daily build (pre built package).

After playing for a while did I get yet another segfault and it repeats as soon as I load my saved game. Is there any instructions on how to provide some good feedback for the developers on this issue?

Below is the log just before crash:

Warning: an orphaned child!
Warning: an orphaned child!
const PlayerState* CGameInfoCallback::getPlayer(PlayerColor, bool) const: Cannot find player 254info!
Dwarf doesn’t have a base stack!
Opening StupidAI
Loaded Stupid AI 0.1
Unrecognized trigger effect type 3019
Unrecognized trigger effect type 3019
Unrecognized trigger effect type 3019
Unrecognized trigger effect type 3019
Warning: an orphaned child!
Warning: an orphaned child!
Segmentation fault
read: Connection reset by peer
Ended handling connection


#314

Anyone that have some input on above?

I have tried several campaigns and I always hit a segfault sooner or later…


#315

Most certainly. bugs.vcmi.eu/my_view_page.php


#316

hi guys
the trunk doesn’t compile on VS2013 :

  • ERM project doesn’t compile (but apparently it’s normal!)
  • Launcher proj doesn’t compile :

I guess some of the config files are badly configured because it’s using someone else’s absolute filepaths


#317

To compile Lancuher under MSVS, you need to run the compilation twice for the first time.


#318

Can somebody accept those two 9-day-old pulls on github? Both look OK for me.
github.com/vcmi/vcmi/pull/42
github.com/vcmi/vcmi/pull/43


#319

Done.


#320

Hi guys
is there a place I’m missing where we can ask how / what specific part of the code does?
THere are some stuff I’d like to refactor but I can’t understand what some stuff is supposed to accomplish.
For instance ObjectTemplate.setSize() is called twice during the loading of the objects and I cant tell if that’s intended or not, because setSize has a different effect depending whether the size is moved up or down