Vcmi 0.71 bug seeking thread

It isn’t problem with the pathfinder. Ha! :stuck_out_tongue:

The difference is caused by the visiting the Scholar. In H3 it takes two move points - one to pick him from the map and second to move on the tile he occupied. VCMI takes only one point (moving) because Scholar is not handled yet. (It’ll be supported in the next release).

The same issue may apply to the monster armies, I’m not sure how it’s working exactly in original H3 (I’ll check this).

Same as #2, all crashes on revisiting objects with spacebar are the same issue.
[Crash on non-first day when revisiting objects that show pop-up (or deactivate adventure map interface in another way)]

The real issue is that Mutare Drake should not show in that tavern. [Garrison is as specified in H3 config file]
I’ve improved getting allowed heroes list, so it should be fixed anyway - Mutare Drake will be correctly recognized as not allowed hero (unless specified otherwise in the map).

Indeed. I made some tests and I don’t see a clear formula. The values I observed:
1 stack: 5%
2 stacks: 62%
3 stacks: 33%
(40 trials)

Additionally, if hero starts with a War Machine it counts as the second stack (but will be given always).

Will be implemented soon :slight_smile:


Rather we should apologize for making so many bugs :wink:
The more reports we receive the better, since all that bugs would come out sooner or later anyway. So - it’s better to have them fixed sooner.

Many thanks for your help! :smiley:

I’ve set up SourceForge tracker. Tracker from trac tools seems to be more advanced, but I don’t know how to allow everybody to submit new tickets. Anyway, you can freely post bugreports there instead of in this forum. If we we consider this way of handling bug reports more convenient, we’ll ask everybody to report all bugs there instead of in this forum.

I took a look at this and it seems that it’s the terrain numbers that is wrong. for example
"8. lava" should be moved to nr 9. and “7. grass/pines” should be number 8. this is why we always come across lava obstacles when fighting on grass.
Also guessing then that magic plains should be moved from nr. 9 to nr. 7. Couldn’t test this tho since magic plains (and the other terrains like: magic clouds or rocklands that you put on top of usual terrains) didn’t seem to be implemented as a battle texture yet.

there is more to this since fighting on rough terrain allowed ship-obstacles to be placed, assuming then that rough is either number 24 or 25 which atm is ship backgrounds.

Ok this post got kinda confusing :stuck_out_tongue: But I guess my question comes down to this: Is there a better way to see what terrain should be on what number than like start fighting on all different kinds of terrains on a edited map and compare the obstacles appearing with the allowed terrain numbers in obstacle.txt?

I’ve copied that list from ERM documentation - Battle Atributes (BA). I’ll check if I’ve implemented ir correctly, but I can’t see why lava should be moved to 9.

Using the list form obstacles.txt should be enough. VCMI uses the same list.

ok it’s probably just something wrong with how it’s implemented. Since ATM all lava obstacles correctly have only nr.8 Lava as allowed terrain. But lava obstacles is anyway showing up on grass terrain. this got me to the conclusion that nr 8 must be the one allowing grass terrain and not lava. And with a little testing I found that it’s nr 9 that is allowing lava terrain ATM.

This should probably not be changed in obstacles.txt tho and instead be changed where this is handled.

Your observation is right.
In the obstacles.txt terrains types uses indices beginning with 1. Later, in the code handling adding obstacles, same indices are used in the array indexed from 0. I think this is causing the problem.
If I’m right, it’ll be trivial to fix :slight_smile:

ERM help is also available online at h3maparchives.celestialheavens.c … /erm_help/

Also something I came across playing around with obstacles:

#22 Obstacle can get placed outside of grid.

Edit: And ye im gonna start report bugs at sourceforge from now on :).

I fear that creating two separate places for reporting and discussing bugs is a bad idea. Ending with two different buglists will only increase confusion. If there are no widely available information about them - even worse. Having some bugs on bugtracker and rest on the board won’t be helpful.

Before deploying bugtracker there are several questions we have to answer. Do we still want receive bugreports on forum? Should the buglist or information about solved issues be available on forum? And so on…

We need to discuss and work out solutions before making bugtracker available. Migration to bugtracker should be neat and properly announced. Haste is wrong in such cases.

So please, stick with this thread for reporting bugs for a little longer.

Via “Permissions” tab in the administration panel… (you have to give yourself access to it using SF admin page)
It’s hard to say which tracker (from SF or Trac) is better, they both have its own pros and cons. I had made some preparations for using the Trac one (categories with their owners, issue types) but I see you made generally the same for the SF tracker…


Fixed. The random numbers generator wasn’t properly initialized in the objects thread handling objects.
Setting and remembering specific resource bonus at the start of each week would complicate the implementation significantly. (These info would have been sent from server to all clients, added to the savegame format, etc.) It can be done but in my opinion it’s now not effort-worthy, especially if the difference can matter only if player is abusing save/load options to get info about bonuses in objects.

Strange… if you encounter it again, please send on my mail savegame and logfiles.

You’re right, best thing would probably be to start with a tracker when a new version is released. And notice everyone clearly where to report.

As for these things I think when we start to use the bug tracker all bugs should be reported there. That way we can get a clear view of what bugs is already reported. We can also see what is solved and so on. so I don’t think that bug lists / solved lists will be needed on the forum. (The bugs in the Unresolved List we have now should IMO be put in the tracker by someone so all bugs we know of is there and so we don’t get any unnecessary duplicate reports).

We will probably still need to have a forum thread for every version tho. Like a discussion thread because there will always come up questions about bugs/features that isn’t that easy to just put up on a ticket. These threads could be useful for you also if you want testers to focus on specific things, like new features or want more information on a bug (in that case refering to the bug in the bug tracker).

Hehe… you got me this time. :slight_smile:

But I’m actually happy I’m finally wrong about spotting a pathfinder issue. I’ve kinda got bored as well of reporting one after every release. :stuck_out_tongue:

You mean Garrison is the name of the in-game Mutare Drake in the game files? (I’m curious what is Mutare’s then…) :smiley:

Anyway, I agree that she shouldn’t have showed up in that Tavern because she’s a campaign hero only (introduced in AB I think). But I thought she’s first of all a Dungeon hero, so which should have Dungeon troops. The fact that she was in the Tavern was less suspicious for me, but I guess that’s because of all the user made maps I’ve played, where often campaign heroes are allowed at Tavern (and I guess we should foresee this case as well, for the future VCMI map editor). :stuck_out_tongue_winking_eye:

Personally I don’t have any experience with bug trackers. I only used a bug report tool when I was testing for Ubisoft, but I assume that’s sth different. I didn’t like it anyway (it was much more fun on the forum, where one gets to see and comment on the bugs reported by others :->).

That doesn’t mean I’ll mind using the bug tracker. It does sound like sth that can be of great help for the devs, now that we seem to find like 50-100 bugs at every release. After I’ll get the hang of it, I’ll prolly love it and prefer to the forum. But a bug tracker will always leave out the new/occasional testers, for which bug threads should always be there (and the let’s say “regulars” could help logging all bugs to the tracker).

[size=75]So, to go back to the bugs: :stuck_out_tongue: [/size]

#23 When I close VCMI using the window close button, I get the typical Windows error: “VCMI_client.exe has encountered a problem and needs to close. We are sorry for the inconvenience.”

Doesn’t bother much, as it’s only when I exit the application. I’m not sure though if I had this the previous days as well, but now I have it every single time I close VCMI. I’ll let you know tomorrow (after a PC restart) if I still have it. I’m thinking it could be a tiny conflict with Winamp: I noticed that first time I opened VCMI today it took like 30 secs to get to the Main Menu screen. And except for Winamp, there was no important application running in the background that may cause a conflict in resources.

For the record, closing VCMI from the in-game dedicated button (Quit to Desktop) does not generate the error.

I’m attaching the Client Logs for one and the other types of application exit - maybe it helps.

I’m attaching the
VCMI_Client_log after (1.29 KB) (1.38 KB)

Fireballs used by Gogs aren’t flying in straight line - instead they are doing something like curve.

Tested on Arrogance3. In HoMM III, the number of creatures at the beggining is their base upkeep - for example, 1 Behemot. In VCMI, with Grail and Castle, you can buy 3 instead.

#26 Reproducible crash: on the map “A Warm and Familiar Place”, going down right from your castle, pass the scholar, sulfur mine and windmill. Pick up then the closest Mercury. If you try to access the Alchemist Lab after that, the game crashes (every time):

#27 Unreproducible crash: I had a crash similar with the one above, when trying to access a Seer’s Hut (from a similar angle as I had above, and also when hero was 2 movement tiles away from destination). A retry on the same map and from the same angle failed. I’ll try using the same alignment, heroes, and maybe even order of movements (if I remember well) to reproduce it. I’ll get back on this if I succeed.
UPDATE: I tried a couple of times to reproduce it (same map, hero), but it doesn’t occur again.

#28 Garrisoning, then removing from Garrison of a hero, does not move him/her at the bottom of the hero list.

#29 Just a feature actually: I see Redwood Observatory functionality has been implemented. The same functionality (same radius) should then be associated with the underground Pillar of Fire as well.

#30 Heroes are still walking over each other on pre-defined paths sometimes (hardly reproducible - had it happening only twice in the last hour)

#31 Nothing happens if I pick up a scroll: the bottom right status box window does not show anything and no scroll appears on the hero…

#32 We cannot cancel a selected spell in any way: not by pressing ESC, not by right-clicking and also not by selecting a different spell from the book instead

#33 We can open the spell book at any time, even after we’ve cast the spell for the turn. The button should be grayed out and unaccessible until the beginning of the next turn.

#34 After playing al lot with opening the spell book numerous times during one battle, and also trying to cancel a spell in different ways, during the next battle I couldn’t cast any spell anymore. I was opening the spell book, selecting a spell, but then nothing was happening when I was clicking the target creature. Note that I’ve had the bug for 1 battle only. The battle after (so 3rd let’s say) could again be carried out also by using spells.

#35 Game crashed again during hero movement. And again (for the 3rd time!) it happened two tiles away from the pre-set destination on his parth:

s[/s] This is most probably related to bug #18. The visible obstacles are ignored, while “invisible obstacles” seem to prevent the creatures to access some hexes which are completely free:

Edited because #36 was indeed same as #18. #36 used for another related bug reported by Aidis below.

No, by “Garrison” I meant her starting army. Sorry for that mental abbreviation.





It’s probably caused by the 1024x768 resolution and walking nearby bottom end of the map.
Border is to small and it apparently causes the crashes. I hope it will be fixed soon.

Apart from that I’ve added support for Schools of Magic/War that were requested in another thread. Additionally I’ve added autosave which wasn’t requested but I thought it will be definitely good to have it, especially when VCMI is still rather unstable :wink:

While trying to reproduce this bug, I also ran into a crash:

#37 I gave one hero a pre-defined path (destination was the town gate, about 10-15 movement arrows away). Then I moved 2 other heroes in 2 different places on that path. Then I hit End turn. In the beginning of the next turn, I tried to move the 1st hero. He did start his movement, but got stuck just before the first “obstacle hero” on his path, also causing VCMI to crash.

I believe you’re right. I tried to reproduce bug #26 on 800x600 and couldn’t (while in 1024x768 it always crashes).

Autosave is indeed a very good idea. VCMI is lately always crashing before I even complete the first week (with so many heroes exploring different areas in the same time). So it’s a bit frustrating I never manage to get far on any map. :stuck_out_tongue: Plus autosave might help in better reproducing some situations leading to crashes.

Thanks for all the other bug fixes and implemented features. :slight_smile:

I think it’s because of wrong blockmaps in obstacles.txt… I still hope someone else will fix it.

I’ll try to implement it soon.

Yes, it’s because of too narrow border.

#38 The Corpse is not accessible from above.

#39 The Info text should not only give the name of the object (i.e.:“Corpse”), but also the (Not visited)/(Visited) text next to it.