This build is basically 0.89c plus two days of fixing.
[size=150]DOWNLOAD LINK download.vcmi.eu/vcmi_089d.7z [/size]
Resolved reports: bugs.vcmi.eu/changelog_page.php?version_id=47
Development log: sourceforge.net/apps/trac/vcmi/timeline
Manual: dropbox.com/s/vry1h23dc4rti … manual.pdf
Previous development build (0.89c): [forum.vcmi.eu/t/vcmi-0-89c-development-version-r2928/528/1)
Apart from bugfixes there is one new feature from me in this release: an initial version of my new battle AI. This package contains both the new “BattleAI” and used so far “StupidAI”. By default still the StupidAI is used. However it is possible change the battle ai the VCMI will use by typing into console:
So the allowed commands are now:
The new battle AI may:
- cast ofensive spells
- cast many of the timed effect enchantments
- wait for better opportunity
- move a little smarter than StupidAI
Still it’s not exactly tested and I’m not sure if it will always finish the battle (so I’m afraid of infinite AI-AI battle). Please give it a try and tell me, if it caused you any issues.
The 0.9 release is planned on Monday. Let me know if there are any urgent issues. Please check, if the package is correct.
Few minor comments on the Manual (more for the sake of bumping this topic over 0.89c):
- We also support resolution 1280x800 now
- To change res the text says “Use interactive ”settings” menu” > maybe it can be replaced with “Use dedicated setting from System Options”
- page 6 is almost empty, chapter 4 starting after a page break, which is a bit inconsistent with the rest of the chapters which didn’t require page break before.
But for the rest, no real issue that I can see. Haven’t browsed the manual in a while, so I was happy to see that it continues to improve, growing into a very nice, professional-looking document. Nice job WM!
@Tow: the package seems correct. At least I used a clean install over a basic version of WoG, and it works with no issues.
And a question for you or Ivan: if WoG data files are now handled as mod, does it mean VCMI doesn’t need WoG anymore (and a H3C install would suffice)?
No. The only change is location of WoG files. VCMI still needs some WoG files like ZELP.TXT (complete list would be quite big).
At some point all WoG features can be made optional but I am not sure about WoG files - they’re a bit more complex to remove.
Tow, will 0.90 come with BattleAI as well or this is just for testing\preview purposes?
Not sure if related to your changes or other recent fixes, but have a look at #1114. It’s pretty much a blocking crash because we’re not in control of the AI so players may not be able to get far if we crash on AI’s turn. So preferably should be fixed before 0.9.
Both AIs will be in the package. I’ll rather leave the StupidAI as the default one, BattleAI still doesn’t feel trustworthy.
It is fixed.
Still I see however one more related crash that needs more investigations (though doesn’t seem to be a new one).
Perhaps you already intend to do so, but maybe it’s best to make a mention in the 0.9 announcement of what users need to change to switch from StupidAI to BattleAI, in case they want to test and report bugs about the latter.
As a side note: For the less familiar users, this parallel period may create some confusion on Mantis, if they create bug reports for issues specific to StupidAI only and then a dev will close them as “no change required” and a comment about BattleAI (which may not even become default in next release). So I’d suggest instead - if we have such cases - to just mark them as Acknowledged for starters, and then move them to Resolved the day BattleAI is made default in an official release. [this is just psychology: does not help the motivation of the tester to see their report was useless]
I’m going to keep testing VCMI whenever I’ve got some spare time, I won’t be reporting any bugs this week because I’m going on a trip.
I wanted to report a bug about wrong spell sprite, the weakness spell looks like the rust that the rust dragons cast when attacking. But I’m okay with this, it might be artistic freedom that you think that weakness should look like that.
Also, there is still this bug with the AI when the game crashes when some AI player is playing. I believe that many people posted about this bug so I don’t want to make a commotion on the bugtracker.
I’ve managed to complete a standard H3 scenario - Good to Go - in 0.89d, with no crash and only few encounters with minor bugs. It was enjoyable to play with the feeling that almost all features are in there, and to finally have an alive AI opponent. I could see the AI was still basic in some cases, making mistakes here and there, but I was also surprised by a couple of intelligent moves, like sending two heroes towards my town via different routes, making it challenging for me to defend against both, or recruiting 8 heroes when I was approaching its town, and spreading them as such to force me to consume my movement in fighting them instead of further approaching the town. I’m not sure if the choice to recruit the maximum amount of heroes daily was coincidental, or related to me just defeating their strongest hero, but it looked really smart. I was just 1 day away, but thanks to this strategy it became 3 days, just enough to pass the week and get a strong army again. Unfortunately this was followed by the mistake of leaving the town with that army, instead of staying behind to protect it with the help of the towers. Actually AI did that mistake for all 3 opponents: never left a defending hero in town, to benefit from the protection of the fully built Castle. So there’s one thing I couldn’t test - the sieges. Hopefully there aren’t many bugs there.
So 0.89d is finally a pretty stable release. I’d say that whatever bug fixes will still be done before 0.9, should be with extra care to not break anything.
I guess this is end of the road for the 0.8x releases. VCMI is ready for 0.9x