The Sojournings of Grog (3.6.1 is out)

[Elvish_Hunter]

I added difficult levels also to A3. I kept our current settings for Easy, and used the old ones for Hard; I also made it so that the five Draugs surrounding the player's units are Revenants on Easy and Deathblades on Medium. While doing this, I tested the scenario with and without shroud and fog, and found two problems:
- first, the code for awarding bonus XP to recruited units doesn't seem to work well: I noticed that no bonus XP are given to side 4 Skeletons and Archer Skeletons (but to Chocobones are given). I still have to check why.
- second, I tried to add a recruitment_pattern to side 4 (fighter, fighter, archer, archer, scout). Despite we're stil using the 1.6 syntax, it is correctly recognized by the engine, at the point that I can see it using the inspect window; but then, it seems to be ignored, because I got three Chocobones (scout) and three Skeletons (fighter) recruited on side 4 first turn. And the stderr.txt doesn't says anything about it. Probably I'll have to file a bug report.
Apart from the issues above, and for two TODOs at the start, that scenario is complete. After fixing, or reporting, the bugs above, I'll be able to upload the new release.

[Eros]

And when you do I'll play test it

[Elvish_Hunter]

Version 2.10.7 is on the 1.9 server. Main changes:
- Added difficult levels and carryover from Prologue to A3 included. The campaign may be a bit easier on Easy due to carryover. I will mark it as 3.0.0 when I'll have completed adding levels to all scenarios.
- Custom [floating_text] tag removed, because is in core. The campaign will work only from Wesnoth 1.9.3 onwards.
- Fixed the bonus XP bug in A3 and other scenarios. This was due to our VARIABLE_OP that produced floating point numbers; giving non integer numbers to RANDOM macro blocked everything. Found by filling the code with [floating_text], solved by adding rounding operations.
- Added, but not yet used, a new Lua tag: [flash_screen]. It is meant as a replacement for {FLASH_WHITE () } etc. macros, it supports more colours (white, red, green, blue, cyan, magenta, yellow, black) and to use it no knowledge of Lua is needed. Syntax:

Code: Select all

[flash_screen]
    color=yellow
[/flash_screen]

Eros, do you agree if I replace the FLASH macros with this? (Imagine, for example, Elyssa in A3 saying "Melt in the hot suns Lich!", and then the screen flashes yellow.)
- Due to lack of time, I wasn't yet able to add Daravel's corrections and bugfixes
I think that it's all for now; Eros, now you can playtest the first scenarios until A3 and tell me what needs correction (I only had the time to test on Hard).

[shadowblack]

I have Wesnoth 1.9.3 and campaign version 2.10.7. I encountered serious bug with the replays for the first few scenarios, but I have no idea if they are bugs in Wesnoth itself, or in the campaign. The bug is that the traits of some units are incorrect in the replay. For example in Homecoming one of the three dwarves at start - the fighter, to be exact - has 41 HP in the replay, but in reality (as can be seen from the attached autosave) he has 38 HP.

Also if I press the Stop button to stop the replay while one of my units is moving during replay the game crashes with the message from the screenshot. Anyone else seen that happen?

wesnoth replay bug.JPG (29.53 KiB) Viewed 2859 times

Did I mention that the cactoses from The Long Way Home are NOT THERE in the replay?


Okay, enough with the bugs:
1) A suggestion for the Homecoming scenario: If the player took the shorter route (and thus did not kill the ants in the south) give him the task to clear the path south. It was really weird seeing Grog walk right through an ant at the end of the scenario, you know... Or just give the player a signpost he must reach instead of automoving Grog, Zhurg and Elyssa
2) On Hard the scenarios did not seem all that hard despite my atrocious luck (in fact the first one was a bit easier because instead of Walking Corpses there are Souless and so the Naga needed more time and more units to kill them than on lower difficulties). But that might be partially due to me having played the campaign before and having some idea what to expect and how to deal with it

Question: Could someone explain to me why I triggered only the scorpions in the first scenario? Not the humans and not the mudcrawlers. I looked at the code, but I still don't get it.

[Anonymissimus]

The C++ crash is not campaign-related. Can reproduce and will file a report.

[Elvish_Hunter]

The bug is that the traits of some units are incorrect in the replay. For example in Homecoming one of the three dwarves at start - the fighter, to be exact - has 41 HP in the replay, but in reality (as can be seen from the attached autosave) he has 38 HP.

I noticed that this happened for all those three Dwarves. We used the mainline GENERIC_UNIT macro, so, at least for now, I haven't any idea why happened. It looks like a mainline bug.

The C++ crash is not campaign-related. Can reproduce and will file a report.

Thanks.

Did I mention that the cactoses from The Long Way Home are NOT THERE in the replay?

I opened your replay, and I can see exactly five of them, at 8,19 13,19 15,17 22,17 and 30,26. The problem, on my side, was that it went in out-of-sync at turn 2, and I don't know why. You know, it's from the 1.4 series that replays had problems...

A suggestion for the Homecoming scenario: If the player took the shorter route (and thus did not kill the ants in the south) give him the task to clear the path south. It was really weird seeing Grog walk right through an ant at the end of the scenario, you know... Or just give the player a signpost he must reach instead of automoving Grog, Zhurg and Elyssa

* waiting for Eros' opinion *
In the meantime, I can tell you that both solutions will require increasing the turn limit, and this will make the next scenario easier due to carryover - unless I increase also the enemy's gold in A2-2 and remove the LIMIT_CONTEMPORANEOUS_RECRUITS macros to enable unlimited recruitment of Berserkers and Runesmiths - but, at that point, such scenario will become unbalanced and a sort-of lottery against these units.
Will be a good idea if I auto-kill all remaining Ants and Spiders on victory?

On Hard the scenarios did not seem all that hard despite my atrocious luck (in fact the first one was a bit easier because instead of Walking Corpses there are Souless and so the Naga needed more time and more units to kill them than on lower difficulties). But that might be partially due to me having played the campaign before and having some idea what to expect and how to deal with it

There are also other reasons:
- I wanted to have an hard, but winnable difficult (did you ever managed to win NR's Nightmare without using debug? I not).
- on Hard, you usually face more and higher-leveled units. This means more XP for you, that in turn may mean that next scenarios could be easier on Hard rather than on Easy.
- you may be a better player than me

Question: Could someone explain to me why I triggered only the scorpions in the first scenario? Not the humans and not the mudcrawlers. I looked at the code, but I still don't get it.

The humans are triggered only by taking one of their villages on grass.

[shadowblack]

I opened your replay, and I can see exactly five of them, at 8,19 13,19 15,17 22,17 and 30,26. The problem, on my side, was that it went in out-of-sync at turn 2, and I don't know why. You know, it's from the 1.4 series that replays had problems...

It went Out-of-sync because SOME of the coctoses are missing. See the following screenshots: the first was made after I loaded my turn 1 autosave, the second was made after I loaded the replay and had it play onl;y the moives of my units. Notice how the cactoses from the first shot are missing from the second even though there’s no way I could have killed them?

The autosave is in the archive I attached in the previous post, so you can check this out for yourself.

Will be a good idea if I auto-kill all remaining Ants and Spiders on victory?

Yes. In fact, now that I think about it this is the best solution.

The humans are triggered only by taking one of their villages on grass.

I took ALL of their villages on grass. No humans appeared. Maybe it is because I captured the first village with a merman? Or because of where I disembarked? If it is neither, then there’s a bug.

And since I didn’t have to fight the humans the first scenario was easier than it would have been otherwise.

[Elvish_Hunter]

The autosave is in the archive I attached in the previous post, so you can check this out for yourself.

Checked. I opened your replay with Notepad++. And I noticed that the Cactose starting positions aren't registered anywhere: this means that SCATTER_UNITS scatters them in different locations. In a few words: SCATTER_UNITS seems to corrupt replays. Can you attach also you start-of-scenario savegame of Long Way Home, so I can see if I can file a report?

I took ALL of their villages on grass. No humans appeared. Maybe it is because I captured the first village with a merman? Or because of where I disembarked? If it is neither, then there’s a bug.

So, I just looked the code and can be more precise: there are two branches, beach landing and pier landing. The second one has the humans, and is triggered only by landing near the wooden docks. By landing elsewhere, you triggered the beach landing branch, and correctly no human appeared. Only problem is that, in this case, should appear Mudcrawlers instead of Scorpions. Clearly it's a bug and I'll correct it as soon as I can. Thanks for your reports

[shadowblack]

Checked. I opened your replay with Notepad++. And I noticed that the Cactose starting positions aren't registered anywhere: this means that SCATTER_UNITS scatters them in different locations. In a few words: SCATTER_UNITS seems to corrupt replays. Can you attach also you start-of-scenario savegame of Long Way Home, so I can see if I can file a report?

So, I just looked the code and can be more precise: there are two branches, beach landing and pier landing. The second one has the humans, and is triggered only by landing near the wooden docks. By landing elsewhere, you triggered the beach landing branch, and correctly no human appeared. Only problem is that, in this case, should appear Mudcrawlers instead of Scorpions. Clearly it's a bug and I'll correct it as soon as I can. Thanks for your reports

You're welcome.

Save game attached.

[Daravel]

I chose 'forests' because Cactoses are variants of the Wose whom are variants of trees that form forests
It's the collective name for the habitat of the race (school of fish, flock of owls, herd of cows etc...).
Other words you could use: Grove, wood, copse, cluster, pod, swarm, drove, gang, nest - there are loads to be found via google. Depends on the image you want to give of the Cactose

[Elvish_Hunter]

You're welcome.

Save game attached.

Thanks, I'll check it.

@shadowblack and Daravel: I corrected the bugs that you found in A0, by converting several [event] [if] to [event] [filter_condition], and by splitting the bottle event in two parts, to avoid having Scorpions picking the bottle. I tested it a lot using the debug console, can you kindly test it as well? If it works well also for you, I'll upload very soon 2.10.8.

A0-Those_Sand_Strewn_Beaches.cfg

(29.89 KiB) Downloaded 196 times

I chose 'forests' because Cactoses are variants of the Wose whom are variants of trees that form forests

Thanks for your clarification. Then I'll keep it.

@Eros: I had some ideas for the A4 and A5 levels:
- A4: raise Bats' gold, decrease Zurg's and Grog's gold, raise swimming Walking Corpses on Medium and both Walking Corpses and Soulless on Hard, increase the amount of turns to resist before opening tunnels.
- A5: Nagas. On Hard, the L2 Assassin could be useful, but as you know its sprite is just a non animated frankenstein, and that's why we removed it. I also have another solution: Vendanna made the sprite (animated) of a female Naga that has a magical attack, like the Mermaid Initiate. (link) I think that it may be useful also for B7, where it can be added to the Nagas' recruit list: opinion?

[shadowblack]

I just tried the revised A0 on Hard. Replay attached (out-of-sync/corrupted because in the replay the nagas have different traits than in the actual game; I assume it is the same problem as with the traits of the dwarves we talked about earlier).

I have also attached a couple of saves, as I want to make sure this is the intended behavior:
1) Beach1 - if I move a unit in the desert I trigger the mudrawlers, as expected
2) Beach 2 - a few turns later. If I capture one of the villages I trigger the humans, and if I move a unit into the desert I trigger the scorpions (regardless of whether I triggered the humans before that). I did both, and that proved a very good idea, as scorpions and humans fight each other as well as me

Is all of the above working as intended?

[Elvish_Hunter]

Is all of the above working as intended?

Yes, and thanks to this I just uploaded 2.10.8 on the 1.9 server. Only changes are the two bugfixes (the other is the bottle event), and also all Spiders and Ants are killed on victory in A2-1.

Replay attached (out-of-sync/corrupted because in the replay the nagas have different traits than in the actual game; I assume it is the same problem as with the traits of the dwarves we talked about earlier).

And also those use GENERIC_UNIT. Now I'll need to find some time to make a test case and file two bug reports for both SCATTER_UNITS and GENERIC_UNIT in replays.

[shadowblack]

Replay attached (out-of-sync/corrupted because in the replay the nagas have different traits than in the actual game; I assume it is the same problem as with the traits of the dwarves we talked about earlier).

And also those use GENERIC_UNIT. Now I'll need to find some time to make a test case and file two bug reports for both SCATTER_UNITS and GENERIC_UNIT in replays.

Let me save you a bit of time (or maybe give you more work):
I just played (a bit of) A Tale of Two Brothers (on the lower difficulty to save time). In the first scenario (replay attached, named AToTB-Отърваване_от_Ма..._повторение.gz in case you're wondering) everything was fine, since it uses the NAMED_LOYAL_UNIT macro. In the second, however, the problem occurs on turn 1 - with two strikes I killed the nearest Scout since it did NOT have the "strong" trait. But in the replay it DID have the "strong" trait and so the replay is out of sync. However that scenario does not use the GENERIC_UNIT macro, as far as I can see. I've attached my start of scenario save as well as the autosaves for turn 1 and turn 2 - watch the replay from the latter and see for yourself. You think it's the same issue? Or should I make a separate bug report about it?

[Elvish_Hunter]

watch the replay from the latter and see for yourself. You think it's the same issue? Or should I make a separate bug report about it?

First: at the moment of this answer I didn't yet watched your replay.

You think it's the same issue? Or should I make a separate bug report about it?

I checked the code for The Chase: it has several

Code: Select all

        [unit]
            side=2
            type=Elvish Scout
            x=15
            y=20
        [/unit]

And the wiki says, about the unit tag:

Spoiler:

random_traits: "no" will prevent random trait generation for units. You should only need to set this for placed nonleaders in multiplayer games or if you want to give a unit less traits than it would normally get for its unit type. When generating traits for a unit, first traits the unit has already been given are excluded. Then "musthave" traits (undead, mechanical) for the unit type are given. Then for leaders (canrecruit=yes) traits that are not available to "any" (currently that's all of them to avoid a multiplayer OOS issue, but later will be restricted based on multiplayer play balance issues) are removed from consideration. Then traits are added randomly until the maximum allowed for the unit type is reached or there are no more available traits. Random traits can now be used in MP games but only when spawned in an event, so not for leaders and other units in the [side] definition.

Now, those [unit] are all inside the [side] tag. And this happens also for my GENERIC_UNIT in A0. But the Dwarves in A2-1 are generated with GENERIC_UNIT inside the prestart event. Considering that these aren't multiplayer games, IMO the replays shouldn't be corrputed at all by random traits. So, please file a bug report, of course using the Two Brothers files. Should I find something to add to your report, I'll do it.