Hi, again.Thx for all opinions. I had no Internet last days so I couldn’t reply earlier and/or add bugs to Mantis.
@ALL Sorry but I’m not testing VCMI from yesterday. I mean bugs/crashes which are from start of the game, in day 111. It couldn’t be missed if someone take the package and just play it or once load the game (already corrected). My point is that VCMI needs much more tester but it is not possible to encourage them to test VCMI if there are so “simple”/ obvious crashes/bugs. If you don’t have time before release, send it to me. I will test it.
Some details:
i.e. bugs.vcmi.eu/view.php?id=624
see the sceenshot there - its not fixed
easy crash in meeting window - just click OK/Enter
crashes after starting game if previous crashed happened
see added zip (sorry I have no time to add it to Mantis - tomorrow I’m leaving for vacation)
Don’t feel it too personally, I just want this project to see finished
See you. few bugs.zip (1.01 MB)
@Tow: These 2 indeed should be fixed now as well I think. I mean, I guess nobody will enjoy testing now w/o hero exchange window. And once the crash happens, we can’t open the application again at all.
I gave it a try and was able to reproduce from the first try so I guess it’s easily reproducible (and I see Boulie attached here the crash files).
Man am I sorry now I wasn’t around to test the dev version and the main release on the 1st. I’ll try not to let that happen again.
Meeting window -> my bad, trivial one-liner.
However I cannot reproduce the second issue (crash on starting after crash). All crashdumps point to font reading but I have no idea, how can this be releated with previous crashes or how that can crash now, especially since that code hasn’t been changed for a while. Are there any special steps (other that start game -> crash on closing meeting window -> close message -> start game again) I should take?
I’m not sure, I have not time to do more tests with this new upload, but I also cannot reproduce this crashes, so maybe it is already fixed, after you fixed “load crash”
Not for me. It was as simple as that: had the meeting window crash -> closed message -> tried to start game again to reproduce the meeting window crash, but couldn’t because it was crashing at start (3 times in a row). I’ll run more tests and see if I get anything else and get back to you -> I anyway have to install a fresh copy of H3 on my PC (yesterday it was just a rough test on my old laptop).
Edit: Could it have anything to do with the console message at start-up saying Error: Old file format! (file ./config/sysopts.bin ) ?
Edit2: I cannot reproduce it on my PC. I’ll have to run more tests to see if there were steps just before the 1st crash which have sth to do with it, or else I’ll go back to my old laptop and see if still reproducible (there I installed 0.83 over a pretty old 0.73).
Ah, that. Indeed I missed it but that’s actually not because I haven’t tried that but the issue occurred only when hero was bought or moved into an empty garrison (so move up -> buy combinations prevented bug from appearing).
Now should be fully fixed.
Is “prudence” an acceptable answer?
I believe I found and fixed the issue. When VCMI tried to load CREDITS.FNT it was actually parsing CREDITS.TXT. That leaded to out-of-bounds memory reading that was randomly crashing.
Should be fixed (workarounded) now.
So that all (meeting window + font loading + garrison operations + #539) I hereby announce to be a hotfix #2 for VCMI 0.83. I’ve integrated into 0.83 package.
Please redownload whole package or just the binaries ( download.vcmi.eu/vcmi_083_hotfix2.7z ).
keep going. I am waiting for stable version. I hope this project will have a better AI than Heroes of Might and Magic 3. I am visiting you from time to time and waiting. good job
Question: Would it be possible to have one of the next releases (0.84 or 0.85) focused on bug fixing?
The amount of issues has slowly grown to a notable number, and it’s getting hard to play without running into one or the other bug. As a tester, I feel my efforts had to slowly shift from “hunting bugs” to “avoiding bugs” (the many known, to get to those not known yet) and scanning through the many Mantis open records before reporting something, to make sure I don’t create a dupe.
We now have on Mantis:
27 crash + major bugs
180 other bugs
36 feature requests
We had a great bug fixing rate this year: 305 out of 352 reports (kudos to you guys:), however there’s a great deal of reports older than 1 year still open (185), for bugs which I kinda got used to. And even if they are not really annoying, I am wondering if they don’t perhaps work at a subconscious level, altering the motivation of choosing VCMI over other games when I have some time to play. I don’t know what other testers feel, but I miss a release like 0.75 was, which was focused on fixes, and which delivered a product which I very much liked (within the known boundries of so far implemented features).
So I think it would be great if before 0.9 you could:
fix most crash + major bugs
fix at least a third of the other bugs
implement at least a couple of the longest open feature requests
(all together, maybe try to bring the total number of open reports from almost 250 now, down to not more than 150 before launching 0.9, if possible including also the new reports which will still come by then)
Take this though only as suggestion/feedback. I’m not even sure if the other testers feel the same. I just noticed a drop in activity on the project the past months. On the coders side I guess it’s RL responsibilities and/or work on changes which are not as transparent to us as bug fixes. On the testers side on the other hand, it’s just my assumption that the reduced amount of fixes might have something to do with it. People have always time to play, so for gamers is more a matter of catching their motivation to test more, or go for other games (at least I for one get a boost of motivation to test more, whenever I see a bunch of bug fixes:)
But if the above doesn’t fit so much in your planning, giving the current resources, don’t let it weigh too much in your decisions. They’re just one tester’s thoughts at the moment and I hope they didn’t come at a bad time or anything.
Otherwise keep up the good work, which IMO remains probably the best quality Heroes project of the past 5 years [size=59](Ubi’s releases included :p)[/size]
Well, it looks like we’re doing it the other way round - working towards new mechanics to get this game more complete. A little AI and features like Stack Experience or Spell Scroll and stack/commander artifacts. However, currently entire mechanics engine is undergoing deep changes, so I need to wait until Tow is done with it.
Ivan was also doing something with animations and general graphics engine, not sure where is he now.
Most of remaining issues are either irrelevant, too complicated or related to battles => Tow Dragon appears busy. Lack of time always hurts.
I wish I had more time for fixing bugs in VCMI… as Warmonger says, currently Tow is doing major changes in the engine. It will fix some bugs and create a lot of new ones but the engine will be more feature-complete. Additionally, VCMI has been accepted as my and Tow’s students project at our University (not as a whole, just AI-related parts) so most of our activity around VCMI will be focused on it (AI handling) during the following month.
As for battle bugs - I’m afraid most of battle interface need a rewrite. For a few years we were trying to add as many new features as possible and finally the monster of the need of refactoring has caught us. I hope I’ll have some free time for VCMI in February. I’ve recently taken Mathematics as a second major at my University and the amount of free time I have has significantly decreased.
GZ Tow and Tow Dragon
you get it accepted as university project
so you can merge needed and wanted
now you can spend more time on VCMI giving that you can say you working on university project
as for battle interface - could you provide more features during refactoring?
I mean it would be great if new version of BattleInterface would be ready for changes which are now not possible
the one i think now is battle screen scrolling and scroll-following the moving creature - it is needed for low resolutions
but the thought is general - I mean after refactoring the battle Interface should be easier to alter, and doit once great, so you wouldn’t needthe second refactoring
Basically that’s the point of what I want to do. But I"m don’t think scrolling is a priority. Tow is going to make scrolling a part of GUI system and doing another, unrelated scrolling implementation would be wrong.
Primarily I’m going to focus on:
different speed of animations for different creatures according to original config file data
better ordering of displayed creatures
more advanced and probably text file-customizable stack animation beginning conditions
The last one is actually more laborious than difficult and I’d like to make just a general framework and have somebody fill it with necessary conditions.
Given the circumstances, the timing of my above request was not the best then, but what we might get instead is even better so no bummer on my side.
Good luck then with the AI sub-project! And let us know if there’s anything we could help with. [size=75] I’ve already posted some ideas for a better battle AI in the General Discussion section - see links - but it’s incomplete work, as I didn’t expect working on battle AI to start so soon, so I don’t know if it is of any help.[/size]
I’m now in the middle of major rewrite in the engine, so I fear I won’t have time for many fixes for 0.84. But when my changes are finished, bugs will get their focus.
I certainly do want to have VCMI pretty stable and without major bugs before undertaking work on 0.9 that’ll incur more major changes.
It’s true, I and Tow Dragon are quite busy nowadays. Now I spend all my VCMI time on bonus system, because I’m the only one capable of rewriting it and that feature is now a blocker for other features. That’s why recently there are no bugfixes from me.
To precise: the subject of the project is not to write an AI, we will create a framework allowing testing and comparing various battle AIs. That’ll include possibility of running battle only game (may be interesting as special MP mode) and automatic “tournaments” between set of AIs. But even if everything goes right, they will be written by the next year students (so not soon ).
VCMI is an open project, any help will be welcomed
For a long time the biggest missing part of game is an AI, especially its “adventure” (non-battle) part. If you could pick it up, it would be great. Even fixing (or rewriting) it to undertake basic game actions would be an significant improvement.
You can ask all your questions on this forums (feel free to create thread in development board), so information will be available to other developers as well.
Well I want to write my bachelor thesis about it…
I need to convince a prof to allow it.
So I have I few questions:
At which University are you?
Whats the name of your prof?
How did you convince your prof thats a good idea to allow you study work at VCMI?
Can your prof maybe convince my prof that writing an AI is a good topic for an bachelor thesis?
Where are you guys from?
Do you have skype? I realy would like to talk to you. Or call you.
I realy want to find work at a game studio after finishing study. Writing an AI would be a great reference.
We can talk more private if you like.
Hey, guys! What’s going on? Everything seems a little bit too peaceful, so I’m worried you guys have stopped developing VCMI. Can we expect any major releases any time soon?
If you want to know if we are doing something, then look at our source code repository: sourceforge.net/apps/trac/vcmi/timeline . Recently there is a lot of commits, new features and bugfixes.
Since last release Tow has redesigned an important piece of VCMI, what makes further development significantly easier. Unfortunately, many already implemented features have been broken. We have to fix the most obvious bugs first, then make a few devreleases before we can make a general release. We can’t do these things before February 1st so the next version (0.84) will come out on March 1st.
Sorry but I’m not testing VCMI from yesterday. I mean bugs/crashes which are from start of the game, in day 111. It couldn’t be missed if someone take the package and just play it or once load the game (already corrected). My point is that VCMI needs much more tester but it is not possible to encourage them to test VCMI if there are so “simple”/ obvious crashes/bugs. If you don’t have time before release, send it to me. I will test it.
).