0.72 bugs and issues

[Zamolxis]

#75 : New crash when attacking a Green Dragon that used (Wait+)Defend in the previous round. I discovered this while playing the same battle which was subject to bugs #71 & 72. I don’t know if it’s related to #72, as behavior is quite different (no Angel freezing, or failed attack leading to the disappearance of the mouse cursor).

Here are the steps that led to it:

• Load game 2009-Jul-11-171057 (in the attachment below)
• Attack Angels with Sephinroth, Angels make the first move coming closer
• Lightning Bolt on Angels, Green Dragons Wait
• Double shoot with Grand Elves, who got morale and could double shoot again
• Shoot with Beholders, Harpies Wait, Attack with Dwarves from behind (Angels retaliate)
• Attack with Troglodytes, then with Harpies and Green Dragons
• Round 2: Angels attack Grand Elves (unless Elves got the morale extra shot previous round, Angels kill all Elves now)
• Lighting Bolt on Angels, Green Dragons Wait, Shoot with Beholders, Harpies Wait
• Attack with Dwarves (Angels retaliate), then with Troglodytes & Harpies
• Green Dragons defend (ending round 2)
• Round 3: If you’re lucky, Angels attack Green Dragons now, causing the crash

Of course, it may not always happen that Angels attack Green Dragons in round 3, especially if they manage to kill all Grand Elves in round 2, so you may need to replay the battle a few times, always applying Defend on Green Dragons at the end of each round starting with round 2. But if AI decides to strike the Green Dragons in this situation, then the crash is always reproducible (3 out of 3 attempts on my side) - see screenshot & attached logs.

i4.photobucket.com/albums/y104/Zamolxis/VCMI/VCMI%2007/th_090712-CrashwhenAngelsattackedDrago.jpg

EDIT: Probably an important note - It seems the bug is pretty much related to the Dragons. I tried Wait+Defend on Grand Elves in round 1, but the game didn’t crash when Angels attacked them in the round 2. That, while for Green Dragons is always reproducible as I said. On the other hand, I hope that’s not misleading information - maybe it’s not related to Dragons per se, but to attacking a 2-hex defending creature,… or something else. If you don’t manage to identify the cause, I’ll try to reproduce it in the future with other creature types (other 2-hex, other dragon types, etc).

#76 : And a small detail (not even sure if it’s a bug or meant as a feature), regarding the mini status window, the lower right box in the game interface, with brief information about the selected hero or updates (new day; size of resource picked; etc). In H3, in the beginning of a new day/week/month, we have a short specific animation. When the animation ends, the content of the box changes to info about the 1st hero (selected) in the hero list. In VCMI, the box remains frozen on the last frame of the new day animation. Moreover, even if we select another hero in the list during the Autosave, the box will still remain frozen on that last frame of the animation (see screenshot below, I selected Sephinroth during Autosave, and the info box did not change). Only after Autosave is completed, we can click on one of the heroes to have the box content changing.

i4.photobucket.com/albums/y104/Zamolxis/VCMI/VCMI%2007/th_090712-Statuswindownotchangedalsoif.jpg

Now, I don’t really mind that the box does not change immediately at the end of the new day animation (I always felt it was kind of short in H3). But I would expect that it changes at least at the end of the Autosave process, and/or at least when I click on another hero in the list, even if the game is still Autosaving. But again - this is just a detail, a small difference from H3 - I don’t think anybody would mind if it’s not fixed.
090712 - Crash when Angels attacked Dragons which Defended previous round.zip (104 KB)

[Zamolxis]

#77 : New Crash when enemy Champions tried to attack friendly Centaurs from below.

Steps to reproduce (if you’re lucky:):

• Load the Saved Game from #75 above and hit End Turn
• Select Alagar and attack the Champions at North-East
• Don’t use magic (it’ll kill them before you manage to reproduce the bug)
• Use Wait until it’s Iron Golems’ turn (if you’re lucky and AI moved Champions within their range)
• Attack Champions with Iron Golems, then with all your other creatures, surrounding them like in the screenshot
• Round 2: if Champion move 1 hex behind and attack the Centaur, you have your crash

This makes me more and more believe that #75 is not related only to the Dragons and maybe not even to the fact that they defended in the previous round. It could have something to do with the angle of attack. Perhaps even related to #73 (issues with correctly identifying the front hex of a 2-hex creature). The more I think of it, the more I realize many battle crashes lately (since 0.7 or even earlier) involve a 2-hex creature attacked from below. There may be an error related to one of the possible angles of attacking 2-hex creatures from below (maybe also from above, though now I don’t remember any such crash). What I do remember, is 2-hex creatures attacked/killed from the front or back hex, and that makes me think the issue maybe be with only 1 particular angle. I’ll keep trying to reproduce this during my attacks, but it seems crashes occur mainly during AI’s attack.

See also if the attached logs are of any help.
090712 - Crash when Champions attacked Centaurs.zip (3.83 KB)

[Tow]

Second part of fixes. Build will be released this evening.
Unfortunately it will break save format, so your last savegames won’t be used probably
However crashes with two-hex creatures are known and not so hard to reproduce.
(the size and attack angle are the problem)

---

#53 - We still have redundant quotation marks in artifact description (a fix is mentioned in the changelog, but either it refers to sth else, or it didn’t make it to the 0.72 package):

Fixed.

#54 (similar to #46 and hardly a bug) - As seen in the screenshot above, the artifact name is displayed on 4 rows, because the number of maximum characters per row is smaller as compared to H3:

I’ve increased the limit, should be better now.

but new anim for tavern is probably displayed as in wrong format (let be bug #55)
(not whole animation box used, animation is very gray, animation looks like semi-random dots)

EDIT: I have forgotten about another bug
there it is
#56 hero can begin exchange with himself sometimes (if not always) when clicking space

Not a 0.72 bugs!

#57 The screen opens in VCMI with the games apparently sorted: first by Number of Players and second by Map Version. In H3, the screen opens with games sorted: first by Map Version and second by Scenario Name.

Implemented.

#58 Sorting by Scenario Name does not work at all.

Seems to be already fixed.

#60 The map save date & time text is more detailed than H3 - which is okay - but perhaps because of the fonts used, it seems a bit too long in terms of occupied space:
The real “bug” here however is the square character

Square character has been removed. I can also print date in H3 format but I’m not sure if it’s better. For me it’s confusing, because I’m used to YYYY-MM-DD format when H3 uses MM-DD-YYYY.
By the way I fixed also bug that converted time to GMT (it was once reported). Now local time will be used

#61 If a resource pile is at “0” and we have an event attempting to reduce that pile, not only should the pile not go under “0” (bug fixed in one of the previous versions), but also the event message window should not display that resource anymore.

Fixed.

#62 The Load interface opens with none of the Saved Games preselected, but actually with the game that would be first in the list in the New Game screen. We can even press on Load and the new game starts.

Fixed.

#63 Lean To says it’s “long abandoned” at first visit already.

Not reproducible
If it occurs again, please try to get save.

#66 This is probably meant as a feature,

Feature
(or rather minor improvement)

#67 This however is a valid issue I believe: H3 used to randomly give -1/+1/+2/+3 Luck penalty/bonus. If I’m not mistaken, the chances of getting 1 of the 4 possibilities were pretty equal. In VCMI however, after 10-15 tries, I got once +3, once +2, but in the rest, a lot of +1 or +0 occurrences.

Really strange, since code for the bonus is:

gbonus.bonus.type = HeroBonus::LUCK;
gbonus.bonus.val = rand()%5 - 1;

I’ll take a closer look at this later.

#69 - I see Luck is implemented since 0.71, but after playing like 50 battles, I have the feeling I haven’t seen any horseshoe spinning above my creatures, even on heroes with Expert Luck. Is the animation not implemented yet, or was I just unlucky?

Fixed a bug that prevented showing luck animation

#70 - After searching through a broken Wagon, I was very happy to find the ‘%s’. Now I’ll just have to take care I’ll put it to good use.

Fixed.

There are two other reported bugs for which I was looking for a resolution: #26 (experience needed for level-ups)

Not fixed, though I’ve improved current values, so first 15 levels should match.
Full fix later.

#47 (maybe fixed together with #27?)

#47 is not so simple. I’ve added a workaround that should prevent crashes. That’s all I can do for upcoming build.

[Zamolxis]

#63 Lean To says it’s “long abandoned” at first visit already.

Not reproducible
If it occurs again, please try to get save.

Yep, it’s a real pity I didn’t upload a Saved Game when I had the chance. It happened with both Lean-to’s next to the Tower on the All for One map, so I thought it’s a general bug, easily reproducible. I tried a gain the map, but didn’t reoccur.

This is a good example maybe of why I created suggestion suggestion 45 in the Missing features thread. I think the resources for Lean-to might be generated at map start (?), and if that is the case, I could have used a save at game start (which would not get overwritten by End Turn saves later on).

There are two other reported bugs for which I was looking for a resolution: #26 (experience needed for level-ups)

Not fixed, though I’ve improved current values, so first 15 levels should match.
Full fix later.

That should be good enough for now. Until L13 formula was variating too much anyway, so I guess you just set hard values (?) and left the implementation of the formula for higher levels for later. Anyway, many thanks for all the fixes!

[Tow]

29 - Crash when trying to enter Monolith Two Way on Deluge map. It’s always reproducible - use the same saved game, and go to the Monolith south of your castle (I’m also attaching the logs, not sure if useful).

Fixed.

[Tow]

#9 interface leak
no infowin on enemies. The infowin is not shown,
and vcmi yells “no infowin for hero xxx” or “no infowin for town yyy”

Fixed (implemented).

#13 timed event bug
timed event which should be everyday from day 1 occurs in vcmi only once (at day 1)

Fixed.

#26 - VCMI Hero Level-Up formula does not correspond to the H3 one*:

Fixed. Fully

#40 In Town Screen, if we have more than 3 towns, clicking on the 1st or 3rd down in the list also scrolls the list, instead of just selecting it. Perhaps is meant as a feature, but personally I find it pretty confusing (so far).

What you could however implement, is Arrow keys for scrolling through the town list.

For clarification, here is the H3 behavior:

• Clicking on another town in the list (of 3 visible) selects/opens it, but does not recenter the list on it (it remains 1st, 2nd or 3rd of the initial visible list, unless we click on the scroll arrows).
• Keyboard Arrows scrolls through the list of towns, maintaining the position of the selected town. So if 3rd was selected, pressing arrow down selects the next one, which then becomes the 3rd visible.

Implemented.

#46 - When gaining the skill from a Scholar (and probably same from Witch Hut), “Intelligence” word does not fit on one row:

Should be fixed.

[majaczek]

the “ooh” bug (which i said about before) was easily reproducible in my game and I think it is/was related which creatures are doing in battle (probably the creature with no sound “steals” sound from last sound played)
sorry for not synchronised message as I’m not able to frequently acces the internet recently

[Zamolxis]

Regarding the “ooh” bug (though Steven may actually be able to give us a better English spelling of that sound :p) :

#8 sound interface issue - when moving units on battle sometimes there’s “ooh” sound even if nothing is attacked - mostly when just standing near opponent

I haven’t reproduced it. Either has been fixed or I had (bad?) luck. Please check in the next build if this problem is still present and if possible tell how it’s reproducible.

I believe it’s the same with what Steven reported previously as 0.71b#25:

0.71b#25 - When in battle, there is the same pain sound effect whenever something happens, such as an attack, or casting a spell, and sometimes when a unit is finished moving.

This is easily reproducible if you load the game attached below. Or any other - if I’m not mistaken, it’s actually happening only on loaded games (which would explain why Tow couldn’t reproduce it by just trying some battles on a new game). But just in case that’s not the reason, simply load the attached game, then pick any of the 3 heroes and attack any of the adventure monsters within their range. You’ll have the pain sound at the end of (almost?) any creature action during battles.
090725 - wrong sound in battle.zip (75.1 KB)

[Tow_dragon]

#35 - The basic value for the Ballista damage is incorrect. We have 2-3 in VCMI, while in H3 it is 6-9.

I don’t know what could cause this… the dmg in VCMI should be even higher than in WoG because VCMI does not take into account distance penalty (yet).

]#36 - The Magic Arrow damage seems too high. I say “seems” based on the number of enemy creatures killed by it, as the battle log does not give the actual damage. EDIT: Valid for other destructive spells as well - possibly they’re all overpowered.

To better tell how far from the correct value the spell damage is, it would be useful to implement spell damage information for the spell book & battle log, but I’m gonna report this in the Missing features thread.

To implement damage calculation I’ve used this site : heroesofmightandmagic.com/he … sair.shtml . Do these formulas seem correct?

Probably bug, maybe not - on Arrogance3 - battle with enemy hero on the coast begining on different battle arenas (attacking from top - battle on “dirt”; from right - battle on “boat”). In HoMM battle arena depends on position of defending army.

It works for me correctly (VCMI almost 0.73).

Regarding #17, #71, #72 and similar - these bugs seem to have common cause. I’ll try to fix them for VCMI 0.74.

#77 : New Crash when enemy Champions tried to attack friendly Centaurs from below.

Err… you managed to attack friendly stack? Interesting, I hope it will be fixed togather with those bugs mentioned above.

[Tow]

I don’t know what could cause this… the dmg in VCMI should be even higher than in WoG because VCMI does not take into account distance penalty (yet).

The base value is 2-3 and it’s coded in H3 .txt config files. However H3 uses special formula for calculating ballista damage range: base_value * (hero_attack + 1).

It works for me correctly (VCMI almost 0.73).

IIRC I’ve fixed it somewhen.

#23 - LEARNING SECONDARY SKILL - not increase value of gained exp.

Not a bug - it was never said to be implemented (cf spreadsheets.google.com/ccc?key= … pLe4raNAWA )

#45 Music does not change when switching between towns.

Fixed.

[Zamolxis]

#35 - The basic value for the Ballista damage is incorrect. We have 2-3 in VCMI, while in H3 it is 6-9.

I don’t know what could cause this… the dmg in VCMI should be even higher than in WoG because VCMI does not take into account distance penalty (yet).

The base value is 2-3 and it’s coded in H3 .txt config files. However H3 uses special formula for calculating ballista damage range: base_value * (hero_attack + 1).

True. However that should be the damage displayed in the War Machine card, as well as the damage inflicted. For example, this is how the Ballista stats of a L1 hero with Attack 2 look in battle (H3):

http://i4.photobucket.com/albums/y104/Zamolxis/VCMI/VCMI%2007/090726-H3Ballistastats.jpg - and that is also the inflicted damage if the enemy is in full range

So in VCMI both the Stats card and the inflicted damage need to be updated, because VCMI displays & inflicts only 2-3 damage always.

//

#36 - The Magic Arrow damage seems too high. I say “seems” based on the number of enemy creatures killed by it, as the battle log does not give the actual damage. EDIT: Valid for other destructive spells as well - possibly they’re all overpowered.

To better tell how far from the correct value the spell damage is, it would be useful to implement spell damage information for the spell book & battle log, but I’m gonna report this in the Missing features thread.

To implement damage calculation I’ve used this site : heroesofmightandmagic.com/he … sair.shtml . Do these formulas seem correct?

The Magic Arrow formula seems correct (I didn’t check the others), but it’s not how it’s implemented in VCMI. According to the formula in the link a hero with Power 1 should inflict 20 damage at basic level. In VCMI the same hero inflicts 60 damage. And that is valid for all damage spells implemented in VCMI: they seem to be 3 times stronger than they should.

By the way, I would take the values in that link with a grain of salt. Whoever uploaded them also got fooled by the way the Ballista damage is calculated, as they give 8-12 as default value (which is probably what they had on the screen when extracting the data, simply because perhaps it was a L2 hero with +2 Attack

//

#77 : New Crash when enemy Champions tried to attack friendly Centaurs from below.

Err… you managed to attack friendly stack? Interesting, I hope it will be fixed togather with those bugs mentioned above.

Not me, the “AI”. But - as you said - it was probably caused by the same reason we’ve had all the other bugs regarding 2-hex creatures. Looking forward to 0.73.

[Tow_dragon]

#35 - The basic value for the Ballista damage is incorrect. We have 2-3 in VCMI, while in H3 it is 6-9.

Fixed (the damage should be 2 * (hero attack+1) - 3 * (hero attack+1) )

#36 - The Magic Arrow damage seems too high. I say “seems” based on the number of enemy creatures killed by it, as the battle log does not give the actual damage. EDIT: Valid for other destructive spells as well - possibly they’re all overpowered.

To better tell how far from the correct value the spell damage is, it would be useful to implement spell damage information for the spell book & battle log, but I’m gonna report this in the Missing features thread.

Fixed. Stupid bug caused all spells to have armageddon’s damage.

#52 - Flagged mines are still shown as neutral (gray) in the mini-map:

Fixed

[Tow_dragon]

All crashes regarding attacking two-hex creatures are (hopefully) fixed. Obstacles should now fit the battlefield too.

[Tow]

34__advmap originally when a hero hasn’t movementpoints to go diagonally(i.e. off the road) but has for a step by road, an attempt to go off the road decreases movementpoints without moving, in vcmi there is no mp loss but the path Xmark dissappears (it’s better to leave it like this, but if following h3 u should fix )))

Will be left as it is for now. Current behaviour prevents us from accidentally losing possible movement.

#49 - Large stack numbers don’t fit very well in the boxes:

Fixed.

#59 Sorting by other categories works, but only ascending. A second click on the column header should sort in descending order.

Fixed.