Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 ... 4 5 [6] 7 8 ... 10

Author Topic: Dwarf Fortress 0.47.04 Released  (Read 256834 times)

therahedwig

  • Bay Watcher
    • View Profile
    • wolthera.info
Re: Dwarf Fortress 0.47.04 Released
« Reply #75 on: April 02, 2020, 06:46:18 am »
I... have been having this weird thing where my clothier strange mood ended in a legendary weaver, and my two leatherworker strange moods into legendary tanners... has anyone experienced something similar, or am I just going insane?
Logged
Stonesense Grim Dark 0.2 Alternate detailed and darker tiles for stonesense. Now with all ores!

delphonso

  • Bay Watcher
  • menaces with spikes of pine
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #76 on: April 02, 2020, 09:57:01 am »
All moods except Possessed will leave dwarves at Legendary in the skill they used in the artifact.
Logged

TomiTapio

  • Bay Watcher
  • OldGenesis since 2012
    • View Profile
    • My Flickr animal photos
Re: Dwarf Fortress 0.47.04 Released
« Reply #77 on: April 02, 2020, 10:29:12 am »
Quote from: therahedwig on April 02, 2020, 06:46:18 am
I... have been having this weird thing where my clothier strange mood ended in a legendary weaver, and my two leatherworker strange moods into legendary tanners... has anyone experienced something similar, or am I just going insane?
in 2011: didn't become legendary leatherworker, became legendary tanner instead:
http://www.bay12forums.com/smf/index.php?topic=86022.0
Logged
==OldGenesis mod== by Deon & TomiTapio. Five wood classes, four leather classes. Nine enemy civs. So much fine-tuning.
47.05e release: http://dffd.bay12games.com/who.php?id=1538
OldGenesis screenshots: https://twitter.com/hashtag/OldGenesis?src=hashtag_click&f=image
My Finnish language file: http://dffd.bay12games.com/file.php?id=14884

therahedwig

  • Bay Watcher
    • View Profile
    • wolthera.info
Re: Dwarf Fortress 0.47.04 Released
« Reply #78 on: April 02, 2020, 10:36:16 am »
Quote from: TomiTapio on April 02, 2020, 10:29:12 am
Quote from: therahedwig on April 02, 2020, 06:46:18 am
I... have been having this weird thing where my clothier strange mood ended in a legendary weaver, and my two leatherworker strange moods into legendary tanners... has anyone experienced something similar, or am I just going insane?
in 2011: didn't become legendary leatherworker, became legendary tanner instead:
http://www.bay12forums.com/smf/index.php?topic=86022.0
Ah! So it's just a quirk of the selection system. Good to know, I was afraid I was going to need to create a reproducible save.
Logged
Stonesense Grim Dark 0.2 Alternate detailed and darker tiles for stonesense. Now with all ores!

delphonso

  • Bay Watcher
  • menaces with spikes of pine
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #79 on: April 02, 2020, 10:56:48 pm »
My bad - missed the point.
Logged

Libertine Angel

  • Bay Watcher
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #80 on: April 06, 2020, 12:18:53 pm »
Got my first ever worldgen crash just now, hope I got the seed right; it was a Small world, Short history, Frequent mineral occurrence, everything else left as default.

Quote
Generating world using parameter set CREATE WORLD NOW = 3233334
 Seed: HjgIk8kEt4ALqf9UsUpv
 History Seed: Bvv0aWjKHK5zuKMNtZZg
 Name Seed: iMNEHVLyBK681aUCpfL3
 Creature Seed: bTL95d9oneKuYWJyCF6t
Logged

jecowa

  • Bay Watcher
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #81 on: April 08, 2020, 03:03:56 am »
I may have found a seed that crashes the game in 0.47.04. I was using the basic world generator. These were the last things saved to my gamelog.txt, so it looks like these are the seeds:
Code: [Select]
Generating world using parameter set CREATE WORLD NOW = 2433333
 Seed: g8gOqU2mY2c0Is4giUaG
 History Seed: 6CyucmKk4agmYmyWW2es
 Name Seed: YME64A2yyYSOK4ymoEUC
 Creature Seed: 2OwWAeEogSEkuagImycw
Nothing has arrived on the calendar.
The weather has cleared.

It crashed at around 420 out of 550 years. Here's the remains of the world if it helps:
https://www.mediafire.com/file/bbu1hcnk635uqsk/crashed_world_after_420_years_-_g8gOqU2mY2c0Is4giUaG.zip/file

The crashes are a lot more rare now. This is my first crash from generating over 50 worlds.
« Last Edit: April 08, 2020, 03:17:27 am by jecowa »
Logged

PatrikLundell

  • Bay Watcher
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #82 on: April 09, 2020, 04:09:30 am »
Toady, is it correct that you can replicate the parameters for a standard generated (crashing) world by:
1. Generating a standard world with the same basic parameters as the crashing one, where the numbers correspond to the position of each of the selections (starting with 1, not 0).
2. When the world generation is done, export the parameters using the 'p' binding, and then abort the world gen.
3. Copy the parameters exported from regionX-world_gen_param.txt into world_gen.txt.
4. Replace the seeds pasted as parts of the world gen parameters with those reported in the log (taking care to use the trouble ones, not the ones from the "host" world just generated).
5. Use advanced world gen and select generation of e.g. CREATE WORLD NOW = 2433333?

2433333 would correspond to Smaller/Long/Medium/Medium/Medium/Medium/Sparse

Using the method described, I failed to replicate jecowa's crash on Win64 using 2 attempts with DFHack enabled and 2 without (the reason for the number 2 being that there might be a difference between the first generation attempt and the second one). Thus, it looks like that crash might be a pesky one that's not trivial to repeat.

Also, if the replication method is correct, I'd suggest reporters verify that they can replicate the crashes they see before posting the generation details.

Edit: I tried Libertine Angel's parameters using the same technique and failed to replicate the crash. Win64, two attempts, without DFHack.
« Last Edit: April 09, 2020, 04:16:31 am by PatrikLundell »
Logged

Shonai_Dweller

  • Bay Watcher
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #83 on: April 09, 2020, 04:30:34 am »
Seeds are in gamelog.txt. No need to export anything.
But, yeah, some seeds which crash just don't repeat the same worldgen when you try again for some reason.
Only seeds which repeatedly crash should be posted.
Logged

PatrikLundell

  • Bay Watcher
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #84 on: April 09, 2020, 07:35:28 am »
@Shonai_Dweller: I don't think you understood what I was saying (but you did get what the aim of the exercise is)

Yes. The seeds are in the log (that's where step 4 says you should take them). However, for us non-Toadys, there is no trivial way to use those seeds to actually generate the world to verify that the seeds actually result in repeatable crashes.
The steps are intended to build up a portfolio of advanced world gen sets where we can actually plug in the standard world gen seeds resulting in crashes so we can verify the crash results. Of course, I may have missed something obvious, in which case I'd like to know how to repeat world gen from standard seed sets without the extra work (assuming that even works...).

What I'm after is to get the community to do the grunt work of weeding out all those seeds that do not result in repeatable crashes, so Toady don't have to waste as much time on failures. Obviously, we'd like to get cause o the crashes that are not reliably repeatable as well, so if a set is found to only crash a significant portion of the time, those sets would be extra valuable. I'd say 1 out of 10 would be a bargain in that market, but it would take some dedication to identify those.
Logged

Shonai_Dweller

  • Bay Watcher
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #85 on: April 09, 2020, 07:49:13 am »
Ah, I see. Standard Worldgen seeds.
Yes, testing that method seems to replicate standard worldgen worlds in advanced worldgen exactly.  (World name, location and owner of several key sites, number of histfigs, number of dead histfigs and number of events. All that matches, probably don't need to check further).
« Last Edit: April 09, 2020, 08:12:15 am by Shonai_Dweller »
Logged

jecowa

  • Bay Watcher
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #86 on: April 09, 2020, 08:12:50 am »
Quote from: PatrikLundell on April 09, 2020, 04:09:30 am
Toady, is it correct that you can replicate the parameters for a standard generated (crashing) world by:
1. Generating a standard world with the same basic parameters as the crashing one, where the numbers correspond to the position of each of the selections (starting with 1, not 0).
2. When the world generation is done, export the parameters using the 'p' binding, and then abort the world gen.
3. Copy the parameters exported from regionX-world_gen_param.txt into world_gen.txt.
4. Replace the seeds pasted as parts of the world gen parameters with those reported in the log (taking care to use the trouble ones, not the ones from the "host" world just generated).
5. Use advanced world gen and select generation of e.g. CREATE WORLD NOW = 2433333?

2433333 would correspond to Smaller/Long/Medium/Medium/Medium/Medium/Sparse

Yes, that's right. I tested it a 3 or 4 times with a working seed and got the same thing.

Quote from: PatrikLundell on April 09, 2020, 04:09:30 am
Using the method described, I failed to replicate jecowa's crash on Win64 using 2 attempts with DFHack enabled and 2 without (the reason for the number 2 being that there might be a difference between the first generation attempt and the second one). Thus, it looks like that crash might be a pesky one that's not trivial to repeat.

Also, if the replication method is correct, I'd suggest reporters verify that they can replicate the crashes they see before posting the generation details.

Edit: I tried Libertine Angel's parameters using the same technique and failed to replicate the crash. Win64, two attempts, without DFHack.

I just tried a third time. It crashed on year 424. This is the long-form seed:
Spoiler: big seed (click to show/hide)

Then I tried another time but with DFHack disabled and it gave me this after the crash:
Spoiler: error message (click to show/hide)
Logged

Shonai_Dweller

  • Bay Watcher
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #87 on: April 09, 2020, 08:29:24 am »
Ah interesting.

I tried jecowa's seed twice, and while it didn't crash either time, it actually produced a slightly different world the second time.
Number of histfigs and events are slightly out. A couple more tries and it'll probably crash.

--edit
Yeah, each time you close Dwarf Fortress, reopen and run this seed it makes a sllghtly different world (judging from number of events, histfigs and dead). That doesn't happen with other seeds. They usually replicate exactly. I haven't crashed yet, but I assume it's only a matter of time.
(If you just push a to abort and rerun the generation of this seed right away, it'll actually generate the exact same world again).

--editedit
Hmm. Actually I can't replicate a completely different seed generated with the rest of jecowa's worldgen settings either.
Checked the procedure again just to be sure, and seeds from small 550 year worlds created in Advanced worldgen repeat every time. What's different about these standard Worldgen settings?

-Finaledit
Tried making 33x33 size world in Adv worldgen (same as jecowa's world) as that seems to be the only difference. And seeds don't repeat. Huh.


--Science-edit
550 year worlds in Advanced Worldgen.
Pocket Island, Smaller Island and Small Island - standard settings but with megabeast cut-off at -1 (off) and max year 550.
Procedure: Create world, copy seed, try to recreate world 5 times (close down and restart Dwarf Fortress each generation)
Criteria for successful replication: Hist figs, dead and Events number.
Windows 10, vanilla DF 47.04


Results (world name and general site placement replicated every time):
Pocket Word (17x17): 5 successful replicated worlds
Pocket world test 2 (since pockets are so quick): 5 successful replicated worlds.
Smaller World (33x33): 5 worlds, 5 different sets of hist fig/events numbers. (Also noted changes to ages Forest-Titan/Hydra/Goblins)
Small world (65x65): 5 worlds, 5 different sets of hist figs/events numbers.

I then tested Smaller and Small worlds with shorter histories of 125 years (new seeds):
Smaller World (33x33): 5 successful replicated worlds
Small World (65x65): 5 successful replicated worlds

Conclusion: Seeds aren't working very well, worlds start off the same then diverge. So, unless it's a short history or very small world, it probably won't replicate, so might not be a good source of crashes.

Note: Even in long histories, if you just push a to abort and start a new worldgen without shutting down DF, world will replicate exactly.
Note 2: I also tried Large Island at 125 years. Replicated 5 times. So it's not losing track just through the sheer number of events occurring (there's more events in this than in the smaller 550 year worlds), but something specific that only happens over a long period of time causing worldgen history to drift.
« Last Edit: April 09, 2020, 09:00:21 pm by Shonai_Dweller »
Logged

jecowa

  • Bay Watcher
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #88 on: April 09, 2020, 11:03:59 pm »
Here's a screenshot of what it looks like the moment it crashes:
Spoiler: crashed world-gen map screenshot (click to show/hide)

I'm curious if yours looks the same.

I changed the world gen settings so it ends at 424 instead of 550, and it didn't crash. So I guess the problem happens in 425. Then I tried to generate it a second time to see if it matched.

The histories of both were identical:
Spoiler: history (click to show/hide)

And the world sites and pops were identical.

And the maps were identical:
Spoiler: map (click to show/hide)

The file sizes of the two save folders were different though. (14,179,937 and 14,183,845)

I opened the save folder and used FC to check both the art_image dat files in the save, and they were identical. Then checked the first 50 and last 50 years of region_snapshot dat files, and they were identical. But checked all the unit dat files, and only 35% were identical:
Spoiler:  unit dat files (click to show/hide)

The world.dat files were also different.
Logged

Shonai_Dweller

  • Bay Watcher
    • View Profile
Re: Dwarf Fortress 0.47.04 Released
« Reply #89 on: April 09, 2020, 11:51:18 pm »
Quote from: jecowa on April 09, 2020, 11:03:59 pm
Here's a screenshot of what it looks like the moment it crashes:
Spoiler: crashed world-gen map screenshot (click to show/hide)

I'm curious if yours looks the same.

I changed the world gen settings so it ends at 424 instead of 550, and it didn't crash. So I guess the problem happens in 425. Then I tried to generate it a second time to see if it matched.

The histories of both were identical:
Spoiler: history (click to show/hide)

And the world sites and pops were identical.

And the maps were identical:
Spoiler: map (click to show/hide)

The file sizes of the two save folders were different though. (14,179,937 and 14,183,845)

I opened the save folder and used FC to check both the art_image dat files in the save, and they were identical. Then checked the first 50 and last 50 years of region_snapshot dat files, and they were identical. But checked all the unit dat files, and only 35% were identical:
Spoiler:  unit dat files (click to show/hide)

The world.dat files were also different.
Did you shut down and restart Dwarf Fortress before retrying the seed?

-edit
Nope, tried 5 times at 420 years with your seed and a slightly different world generated each time.
(Remember to shut down and restart Dwarf Fortress each time).

It's very similar, ending in the age of Desert Titan and the sites are usually in the same place. But histfigs, dead and events are different each time, so not the same.
« Last Edit: April 10, 2020, 12:02:56 am by Shonai_Dweller »
Logged
Pages: 1 ... 4 5 [6] 7 8 ... 10
« previous next »