VCMI 0.68 - development version

I tested this version (vcmi_068.7z and vcmi_r679.7z ) - crash
but version vcmi_r631.7z work.

I have vcredist 2005 and 2008, I also deleted msvcp90d.dll and msvcr90d.dll from vcmi folder.

  • crash when i spell “Magic Arrow” (Magiczna Strzała) - crash {“Memorry can’t be read”}

  • i try edit resolution but alvays run in 800x600

  • fullscreen not work
    ^^ I edited settings.txt

What work in Genius AI ? Monster stil doing nothing.

I also try to compile code form trunk (CodeBlocks + mingw + boost + sdl + lua) and it’s seems not complete source.

What does it mean exactly? What happens during a battle? Is there any activity?

Unfortunately, there are some bugs but AI should work. I’ve only compiled it in VS 2008, I don’t know how it works on other compilers.

Are you running 0.68 or r631? If r631, then I believe it’s normal that AI doesn’t work in battles.

It was hard but I believe I’ve finally fixed that strange bug. It was problem with vcmi interprocess memory mapping. New build works for amvmichael and I hope it’ll also help everyone else who suffered from that bug.
I attach package with new binaries. Except of evil bug it fixes invisible hero bug reported by Steven and brings improvements for AI and battle console.

Revision 631 is basically VCMI 0.64, so all new features are not supported there.

Well, it only seems. They are complete and working, though I’ve not tested building with MinGW. I can provide help for building VCMI in MSVC.
vcmi_069.7z (1.53 MB)

Hehe… for a second I thought this one is crashing as well. Luckily it was only a map…

0.69 bugs:

69#1 A Terrible Rumor crashes at start. No relevant message in the console. Actually the last line says: “Server opened map properly”

Indeed ironic. :wink:
I’ve reproduced it, this is bug in reading info about spell scroll. Should be fixed within a few mins.

Fixed. Should I upload binaries with that fix?

Not sure what a “binaries” are exactly (the files / the package ?). You can, but maybe there are more bugs to be fixed before a 0.69a. As you feel…

Meanwhile here’s two others (well, more in the functionality area than bugs):

69#2 If hero is visiting a map object (Magic Well, Witch Hut etc), and you you give him a path going up afterwards, the first left/right arrows of that movement path are not yet implemented (but the hero takes the path correctly otherwise)

69#3 To attack monsters on Adv.Map, we really have to walk “over” them to trigger the battle. Normally any tile surrounding the monster should trigger the attack already. Though probably this implementation is delayed intentionally (to give us more freedom to move around monsters for testing purpose).

0.69 - Viking We Shall Go! Stronghold Tavern is crashing 90 - 95% of the time. Attaching the save game where I experienced this. It may be when an extension hero appears (one not normally enabled) that the game is likely to crash.

Let me know if you’d like saves uncompressed.

Best regards,
Steven.
0-69No5.7z (71.2 KB)
0-69No1.7z (87.3 KB)

Big 0.69 bug. Artifacts on map don’t disappear when you pick them up, so you can get as many copies of it as you want. It seems to happen in all maps.

Best regards,
Steven.
0-69No6.7z (126 KB)

For me it seems to happen every time I try to save the game AFTER the hero has picked at least 1 Treasure Chest (does not matter if I choose Gold or Experience). It doesn’t matter which map, I’ve had it in all 3 maps tested lately. But just in case it’s relevant, the last one was “Good to Go”.

In terms of behavior, I noticed that after “step 3” in my description above, some functionalities of the game still work (ie. I can open/close hero screen w/o crash). Other functionalities however trigger the crash at step 4 (ie. open town screen, or hero trying to move from his spot).

I checked in the “games” folder to see what happens with the files, and it seems after this crash, the .vsgm1 file is not created. More exactly:

  • at game start I saved successfully under the name “Tyraxor”
  • 2 files were created in the Heroes/games folder: Tyraxor.vlgm1 & Tyraxor.vsgm1
  • I went then to pick a treasure chest, and afterwards I tried to save under “Tyraxor1”, but the game crashed
  • I checked on my drive, and only “Tyraxor1.vlgm1” could be found

Let me know if you want me to send the file or you need anything else.

And some more stuff:

69#4 A console message I noticed while loading the Good to Go map: “Minor warning: lacking def info for 70 0 AHRANDOM.DEF”

69#5 Even on a road which should give movement penalties, the hero cannot move in spite of his movement bar saying he can still move 1 tile:

http://i4.photobucket.com/albums/y104/Zamolxis/VCMI%202009/090131-Movementnotleft.jpg [size=75](it may not be clear in the screenshot, but there is a bit of green bar still left next to hero’s icon)[/size]

What we had in Heroes 3 was actually quite the opposite: on roads, which give movement bonus, sometimes the movement bar was totally empty (meaning less than 1 movement tile left for the hero), but hero could still move due to the road bonus:

  • rarely on Dirt Road (75% mov.cost)
  • sometimes on Gravel Road (65% cost)
  • and often on Cobblestone Road (50% cost)

69#6 Now we have a new bug, with the game crashing when hero levels above 13 level. :slight_smile: - Always reproducible (at least I had it happening all 4 times when I got that far, on 2 different maps).

To easily reproduce both #5 & #6 you can use the “Just testing” map I’ve uploaded in the 0.64 thread.

I can confirm the crashes Zamolxis reported that happen when saving after picking up a treasure chest, and that the .vsgm1 file does not get created when the game crashes.

Best regards,
Steven.

Some bugs i found. Unfortunately, i wasn’t here for a looong time and i don’t now if they were already posted.

  1. When i hired Gelu, it had 10 pikemans (instead of shooters) in army.
  2. When i picked armor, it was still on the map (so i could copy it without limits).
  3. And the last - when i clicked on undiscovered terrain in place, when enemy hero was (i believe), the game crushed - unfortunately i didn’t make a screen.

It’s not important. It won’t affect gameplay in any way.

Are you sure such an arrow should appear? Let’s say a hero is on tile H (in a visitable object) and wants to go upwards through tiles N. Then we should see one arrow on the bottom N and an X on top N. As I see we get this in VCMI.

OOOO
OONO
OONO
OHOO
OOOO

We’ll see what causes this (either movement bar level calculation or pathfinder’s cost function).

I don’t have this problem on any map (even your “just testing” map). I tried three times and everything worked.

[edit]

It’s been already reported. We’ll fix that.

It’s because VCMI tried to display that hero’s infobox. When we disable checking by right mouse click what’s under FoW, the game won’t crash.

PS. Thanks for testing!

Thank you for answer, I download now vcmi_069.7z and it’s work.

@Tow dragon:

Check the attached map. Load it in both H3C & VCMI, go to the Magic Well, and then set a path with a destination above the Well. You’ll see what I mean.

Regarding the crash when hero is supposed to reach level 14, for me it happens all the time. I can try a system restart and maybe on other maps, but as said - in my case (and I have 0.69) - if I start opening all those chests in “Just Testing” to level up, 13 is the last level achieved before the crash. When I pick the experience again from a chest which (I guess) would give me L14, it crashes.
movement test.h3m (2.81 KB)

What is the normal movement percentage used for Gravel road? That’s the middle type of road (between Dirt and Cobblestone), right?

I got to L15 and L14 in two tries of the “Just testing” map. Other maps might take a bit longer.

Best regards,
Steven.

@Steven: sorry, it’s 65% for Gravel road. I corrected it in my original post as well.

It’s strange about my L14 crash. Must be because of all the game versions overwriting while testing the other bug. Last one installed is 0.69, but still, who knows what other file is causing the crash. I’ll do some more testing tomorrow (have to run to a wedding now). :stuck_out_tongue:

I’ve checked it and it really was (I’ve fixed it) a problem with VCMI.

Another bug appearing only on certain computers? :-/