Reproducible crash on v1.19.27
Moderator: Forum Moderators
Forum rules
Before reporting issues in this section, you must read the following topic:
Before reporting issues in this section, you must read the following topic:
- Spannerbag
- Posts: 968
- Joined: December 18th, 2016, 6:14 pm
- Location: Yes
Reproducible crash on v1.19.27
Apologies if this is already known.
On recruit page add search term that won't find anything.
Screen grabs showing setup prior to clicking "Recruit" button.
1.19.27 - windows 11 Short pause then game crashes (exits silently).
1.18.8 - windows 11 Recruit issues no unit recruited message and game continues.
Tried this on a couple of scenarios on both versions and behaviour stayed same.
Also noticed a change in recruitment during testing - traits do not seem to change for a specific unit when save-loading whereas, IIRC, in the past they did.
Not saying it's a problem, just mentioning in case it's related to this issue.
On recruit page add search term that won't find anything.
Screen grabs showing setup prior to clicking "Recruit" button.
1.19.27 - windows 11 Short pause then game crashes (exits silently).
1.18.8 - windows 11 Recruit issues no unit recruited message and game continues.
Tried this on a couple of scenarios on both versions and behaviour stayed same.
Also noticed a change in recruitment during testing - traits do not seem to change for a specific unit when save-loading whereas, IIRC, in the past they did.
Not saying it's a problem, just mentioning in case it's related to this issue.
Re: Reproducible crash on v1.19.27
I have a strange sense of déjà vu.
Do you have "Default RNG" selected? Or "Predictable RNG"?Spannerbag wrote: August 19th, 2026, 1:52 pm Also noticed a change in recruitment during testing - traits do not seem to change for a specific unit when save-loading whereas, IIRC, in the past they did.
- Spannerbag
- Posts: 968
- Joined: December 18th, 2016, 6:14 pm
- Location: Yes
Re: Reproducible crash on v1.19.27
Well I'm consistent.
Hope it won't be 3rd time lucky...
Didn't spot it until you mentioned the RNG options, but seems that this version uses "Reduced RNG" by default - or my cr*ppy touchpad changed it and I didn't notice.
Reverted to standard RNG and now recruit traits change when save-loading, so that's cleared up. Thanks!
Re: Reproducible crash on v1.19.27
It appears that "Default RNG" is no longer, in fact, the default.Spannerbag wrote: August 20th, 2026, 9:18 am Didn't spot it until you mentioned the RNG options, but seems that this version uses "Reduced RNG" by default - or my cr*ppy touchpad changed it and I didn't notice.
- Spannerbag
- Posts: 968
- Joined: December 18th, 2016, 6:14 pm
- Location: Yes
Re: Reproducible crash on v1.19.27
gnombat wrote: August 20th, 2026, 11:21 am It appears that "Default RNG" is no longer, in fact, the default.
Also seems that for SP you pick the RNG you want at the start of the campaign then that's it; there doesn't seem to be a way to change it.
Neither does the game report what RNG is active - but I didn't dig into everything and may well have missed something...
Update: just launched BfW and it came up with default RNG.
So I changed it but didn't play anything and exited, restarted and got default RNG again, so I think I must've somehow set it to reduced RNG...
Not tested this but am wondering what happens if I start a campaign on one RNG, play a few scenarios then restart it on a different one then return to earlier saves: would the earlier ones still have the original RNG or would the later setting override it?
Cheers!
Re: Reproducible crash on v1.19.27
No, they actually changed it.Spannerbag wrote: August 21st, 2026, 11:12 am Update: just launched BfW and it came up with default RNG.
So I changed it but didn't play anything and exited, restarted and got default RNG again, so I think I must've somehow set it to reduced RNG...![]()
5185dfd25b49e89a982d92b09d6a5e698d41f59b
If you delete your preferences file, then restart Wesnoth, you will see that the "Default RNG" is no longer the default RNG ... let me rephrase that ... the "Reduced RNG" is the default RNG, not the "Default RNG" ... let me try one more time ... the initially selected RNG is the "Reduced RNG".
(As an aside: apparently at some point this choice of RNG was renamed from "Biased RNG" to "Reduced RNG" to further enhance the confusion.)
The choice of RNG is stored in the save file, so I would presume that's what it uses.Spannerbag wrote: August 21st, 2026, 11:12 am Not tested this but am wondering what happens if I start a campaign on one RNG, play a few scenarios then restart it on a different one then return to earlier saves: would the earlier ones still have the original RNG or would the later setting override it?
- Spannerbag
- Posts: 968
- Joined: December 18th, 2016, 6:14 pm
- Location: Yes
Re: Reproducible crash on v1.19.27
I agree wholeheartedly with your arguments and sentiments.gnombat wrote: August 21st, 2026, 2:44 pm ...No, they actually changed it...
...If you delete your preferences file, then restart Wesnoth, you will see that the "Default RNG" is no longer the default RNG ... let me rephrase that ... the "Reduced RNG" is the default RNG, not the "Default RNG" ... let me try one more time ... the initially selected RNG is the "Reduced RNG".
(As an aside: apparently at some point this choice of RNG was renamed from "Biased RNG" to "Reduced RNG" to further enhance the confusion.)
Also even more consufed given this!
IMHO this is a non-trivial change to the sp gaming experience, breaking with established philosophy up to v1.18 and AFAIK there's no way to change RNG once a campaign has started?
I don't understand why the devs decided to do this given that in other matters (e.g. core units) they keep things the same across mp & sp.
It would've been appreciated if the devs had approached players and UMC creators for feedback first.
But perhaps, as sp appears to be the poor relation vs mp minor rant deleted), it was thought we wouldn't understand the implications sufficiently to provide coherent rational feedback.
If said feedback was requested than I, for one, missed it completely...
I think I'll see if there's any way to preselect the default RNG at [campaign] level (tho' I suspect RNG selection is invisible to WML).
I could go on...
Really don't understand the rationale behind this.
Thanks for the info, much appreciated.
Cheers!
Re: Reproducible crash on v1.19.27
There was some discussion of it on Discord/IRC:Spannerbag wrote: August 22nd, 2026, 12:00 pm If said feedback was requested than I, for one, missed it completely...
https://irclogs.wesnoth.org/2026/06/%23 ... -06-19.log
https://irclogs.wesnoth.org/2026/06/%23 ... -06-20.log
https://irclogs.wesnoth.org/2026/06/%23 ... -06-21.log
https://irclogs.wesnoth.org/2026/06/%23 ... -06-22.log
https://irclogs.wesnoth.org/2026/06/%23 ... -06-25.log
https://irclogs.wesnoth.org/2026/07/%23 ... -07-20.log
https://irclogs.wesnoth.org/2026/07/%23 ... -07-21.log
https://irclogs.wesnoth.org/2026/07/%23 ... -07-22.log
- Spannerbag
- Posts: 968
- Joined: December 18th, 2016, 6:14 pm
- Location: Yes
Re: Reproducible crash on v1.19.27
Thanks for taking the trouble to compose and post these logs, much appreciated.
Sadly didn't have time to read in detail all your posts (might get chance during coming week).
However I understand, to some extent, the pressures that underpinned the discussion and decision, but still feel it would've been courteous and considerate to at least post something here.
My campaigns are generally intermediate difficulty (mainly 'cos I'm not a good or patient enough player to be considered "expert" ... experienced yes but nowhere near a top tier player).
So to keep things "intermediate" I buffer vs poor RNG, by gold provision, map design (kinda), ai tweaks etc. plus cloaked cth manipulation occasionally.
Using a more deterministic RNG will muck this up, especially (as has been noted in the discussions) specials like poison and slow are buffed.
I also feel that it would be useful to provide a way to change the RNG selected at campaign start later in the game.
FWIW I feel that if new/novice players are put off by poor RNG could I humbly suggest that maybe campaign designers are being a tad too clever?
I've given some passing thoughts to writing a campaign for novice players but have never had the time to properly think the idea through.
If players really want a more predictable gameplay experience then fair enough, let them have that.
However I really don't feel that defaulting to a more predictable RNG is the right approach; just design campaigns with more focus on the likely capability of the player. This does not mean a scenario must be strategically easy, just that combat is likely to be more survivable. To win the player must still make correct (or at least avoid stupid) decisions.
Thanks again for the post - never used discord (prefer this forum because - I imagine - time zones don't matter so much).
Cheers!
Sadly didn't have time to read in detail all your posts (might get chance during coming week).
However I understand, to some extent, the pressures that underpinned the discussion and decision, but still feel it would've been courteous and considerate to at least post something here.
My campaigns are generally intermediate difficulty (mainly 'cos I'm not a good or patient enough player to be considered "expert" ... experienced yes but nowhere near a top tier player).
So to keep things "intermediate" I buffer vs poor RNG, by gold provision, map design (kinda), ai tweaks etc. plus cloaked cth manipulation occasionally.
Using a more deterministic RNG will muck this up, especially (as has been noted in the discussions) specials like poison and slow are buffed.
I also feel that it would be useful to provide a way to change the RNG selected at campaign start later in the game.
FWIW I feel that if new/novice players are put off by poor RNG could I humbly suggest that maybe campaign designers are being a tad too clever?
I've given some passing thoughts to writing a campaign for novice players but have never had the time to properly think the idea through.
If players really want a more predictable gameplay experience then fair enough, let them have that.
However I really don't feel that defaulting to a more predictable RNG is the right approach; just design campaigns with more focus on the likely capability of the player. This does not mean a scenario must be strategically easy, just that combat is likely to be more survivable. To win the player must still make correct (or at least avoid stupid) decisions.
Thanks again for the post - never used discord (prefer this forum because - I imagine - time zones don't matter so much).
Cheers!
Re: Reproducible crash on v1.19.27
I also prefer this forum to Discord, but I'll just note that Discord allows you to reply to other people's messages, and the Discord user interface makes it pretty clear what message you are replying to, even if the original message was made a long time ago. (So in that sense Discord bears some resemblance to a forum rather than simply being a real-time chat application.) You could even reply to one of the messages in the above logs (from June and July) and it will work fairly well. If you want to provide feedback, that might give better results than posting here in this forum, since the developers seem to mostly use Discord these days and rarely (or never) use this forum.Spannerbag wrote: August 23rd, 2026, 12:47 pm Thanks again for the post - never used discord (prefer this forum because - I imagine - time zones don't matter so much).![]()
- Spannerbag
- Posts: 968
- Joined: December 18th, 2016, 6:14 pm
- Location: Yes
Re: Reproducible crash on v1.19.27
Mmm, I've noticed a few "regulars" seem to have stopped posting. Maybe they're now only active on discord ... or they've moved on.gnombat wrote: August 23rd, 2026, 4:37 pm ...If you want to provide feedback ... developers seem to mostly use Discord these days and rarely (or never) use this forum.
Oh well, I'll give discord a try sometime...
Cheers!
Re: Reproducible crash on v1.19.27
Was this ever logged in GitHub? A Discord conversation means nothing as it's ephemeral and some may not even get access to it.
I'm surprised to see that this is still an issue with current master at time of writing. So this hasn't been resolved.
I'm surprised to see that this is still an issue with current master at time of writing. So this hasn't been resolved.
Soli Deo Gloria