Bug Reports v4 -- POST BUGS HERE https://www.smogon.com/forums/ps-bug-report-form/

Status
Not open for further replies.
Probable minor bug with the implementation of the ability Mimicry: the determination of the Pokemon's "original type" can be inconsistent depending on whether Mimicry activated on switch-in or when terrain disappears.

https://replay.pokemonshowdown.com/gen8camomons-1192481738-el1rkyz3i9jzokowsj3oq6i1kh50evmpw

In this Camomons game, compare the effects of Normal and Psychic-type attacks on Stunfisk-Galar in turns 1-2 (damage resisted, due to Ground/Steel typing) with turns 8-9 (damage ignored, seemingly due to Ghost/Dark typing resulting from Curse and Foul Play being in the first two moveslots).

I don't know if there is a situation in which this divergent behaviour can manifest outside of OMs, but I think it's caused by the implementation of Mimicry itself; specifically, the "original type" being inconsistently sourced from either baseSpecies or species. But maybe this is actually an accurate simulation of some on-cart bug or inconsistency(?)

BUG STATUS: NOT A BUG; Matches on-cart behaviour. DaWoblefet edit: the associated research.
 
Last edited by a moderator:
In the teambuilder, Kantonian Raticate is listed as able to learn Shadow Claw in generation 7 and national dex. The Pokemon Showdown Pokedex and the Smogon Pokedex both say that it is able to learn it, but Bulbapedia, Pokemon Database, and Serebii all say it cannot learn this move.

BUG STATUS: FIXED by Marty
 
Last edited by a moderator:
there's a bug with the /statcalc command regarding EVs given, putting 85 EVs into one stat calcs as if 88 EVs were put

1600989192536.png

statcalc2.PNG


BUG STATUS: FIXED by PartMan
 
Last edited by a moderator:
We Wuz Nidokangz said:
Probable minor bug with the implementation of the ability Mimicry: the determination of the Pokemon's "original type" can be inconsistent depending on whether Mimicry activated on switch-in or when terrain disappears.

I don't know if there is a situation in which this divergent behaviour can manifest outside of OMs, but I think it's caused by the implementation of Mimicry itself; specifically, the "original type" being inconsistently sourced from either baseSpecies or species. But maybe this is actually an accurate simulation of some on-cart bug or inconsistency(?)
Click to expand...
This is indeed apparently how the on-cart inconsistency works - when a Pokémon gains Mimicry, it changes back to its real type, even if it obtained Mimicry through Transform... but if Terrain is set and removed, then it changes to its transformed type.
 
Aprox. speeds between pokemon not from level 100 will always be innaccurate. I was fighting a random battle where a quagsire had 60 more speed than what Showdown showed to me
opera_A6wNQuoI2f.png
opera_hsB8ZT633o.png



BUG STATUS: NOT A BUG, this was tier shift shenanigans (checking the replay using admin powers, I can see the format was gen8randombattle@@@Gen8TierShift) (DaWoblefet)
 
Last edited by a moderator:
Germinha said:
Aprox. speeds between pokemon not from level 100 will always be innaccurate. I was fighting a random battle where a quagsire had 60 more speed than what Showdown showed to me
View attachment 277799View attachment 277798
Click to expand...

this was probably a tier shift tournament, Quagsire being NU in gen5, it gains 30 base points in every stat bar HP in tier shift, reachins 65 base Speed. Base 65 at level 86 with 31 IVs, 84 EVs: 161.
 
Pujolly said:
there's a bug with the /statcalc command regarding EVs given, putting 85 EVs into one stat calcs as if 88 EVs were put

View attachment 277794
View attachment 277795
Click to expand...

It's not necessarily that it counted 88 evs instead of 84 evs.
Another explanation is that they forgot to floor the"ev / 4" part in the computation.
In the following computations, / is a floating point division, so 85/4 does not equal 21, but 21.25.
If you compute (2*45 + 31 + 85/4)*88/100+5, it gives 130.18, where (2*45 + 31 + 84/4)*88/100+5 gives 129.96

It's easy to test this hypothesis using different values.
For instance at lvl 87, the correct result is (2*45 + 31 + 84/4)*87/100+5 = 128 (after truncating)
With 85 ev, (2*45 + 31 + 85/4)*87/100+5 = 128.7575
With 88 ev, (2*45 + 31 + 88/4)*87/100+5 = 129.41
The statcalc command displays 128 with 85 ev.
It also displays 128 for 86 ev and 129 for 87 ev which agrees with my hypothesis.

It's still a bug of course.

1601023863314.png

1601024172689.png

1601024190688.png
 
Ytreza said:
It's not necessarily that it counted 88 evs instead of 84 evs.
Another explanation is that they forgot to floor the"ev / 4" part in the computation.
In the following computations, / is a floating point division, so 85/4 does not equal 21, but 21.25.
If you compute (2*45 + 31 + 85/4)*88/100+5, it gives 130.18, where (2*45 + 31 + 84/4)*88/100+5 gives 129.96

It's easy to test this hypothesis using different values.
For instance at lvl 87, the correct result is (2*45 + 31 + 84/4)*87/100+5 = 128 (after truncating)
With 85 ev, (2*45 + 31 + 85/4)*87/100+5 = 128.7575
With 88 ev, (2*45 + 31 + 88/4)*87/100+5 = 129.41
The statcalc command displays 128 with 85 ev.
It also displays 128 for 86 ev and 129 for 87 ev which agrees with my hypothesis.

It's still a bug of course.

View attachment 277855
View attachment 277856
View attachment 277857
Click to expand...

it makes a lot of sense and you explained it really well, thanks :]
 
I've known about this for a while but never decided to actually do anything about it.
/ms (pokemon) in past generations searches for learn sets in /nms, and it should probably be searching the learn set from the given generation.

For example, /ms3 jirachi returns Charm as one of the moves, but it isn't actually able to learn that move until Gen 8.
msbug.PNG

BUG STATUS: CONFIRMED - CHAT (SERVER) (DaWoblefet)
 
Last edited by a moderator:
How magnezone trapped me it was not a steel type it was a ditto. I've faced this problem before too.
Here is the replay https://replay.pokemonshowdown.com/gen8ou-1193468705-i9zff6ojifjzp6x3x73ka2kbibxlr98pw

BUG STATUS: CONFIRMED - MECHANICS (CLIENT), it displays that the Pokemon might be trapped but it's not actually trapped. Since the player should know their own Pokemon's typing, the Magnet Pull text should not display since it's only checking for Magnet Pull vs. Steel-types. This generally speaking always displays with Transform when it shouldn't. (DaWoblefet)
 
Last edited by a moderator:
So i found a bug when typeless pokemon with wondergard can still be hit by everything.

Normaly only status move or wheater is suppose to touch it.

Before someone ask yes this is possible.
Use mimic with Shedninja to have the move entrainement. Then use Entrainement on a Fire type and use burn out with that fire type.

BUG STATUS: NEEDS MORE INFO, I cannot reproduce this (note that I skipped the Entrainment step but Entrainment should have nothing to do with it). Can you provide a replay? Here's one of it working: https://replay.pokemonshowdown.com/gen8doublescustomgame-1197449841
 
Last edited by a moderator:
site has started causing all kinds of problems as of yesterday. I was told to google this and it had nothing and it's gotten worse so I need this to be known. once in a while, the site shows I joined a battle but I have to click the battle then refresh to even do anything then it's fixed. the big issue now is the site has started to get to where I need to click next turn just to make the first move in the battle. yes I'm one of the players in it. if this doesn't fix it I need to refresh to make a move. today it forced the refresh and I couldn't even see the battle until I did /report and clicked the battle again.

BUG STATUS: NEEDS MORE INFO, this is really vague and hard to reproduce. Can you provide video behavior? (DaWoblefet)
 
Last edited by a moderator:
I found a bug with the modlog.

It seems the modlog was recently updated with keyword arguments for easier filtering, but this does not always work as expected. For example, in the OU room /modlog staff=uberpyro is expected to bring up all punishments carried out by me. However, it is missing a punishment from earlier today. [10:09pm] (overused) HOURMUTE: [onestinkyboi]: by uberpyro shows up with /modlog and /modlog action=hourmute, but does not show up with the expected /modlog staff=uberpyro. In fact, it seems that all punishments from 9/29 and later do not show up in the modlog with the /modlog staff=STAFF argument, but they do show up when searched for in other ways. Additionally, when searching for punishments/commands done on me with /modlog uberpyro (again in the OU room), there is some weirdness like randomly including some of the bans done by me as well as including dates on which no punishments are shown. I am concerned since the modlog is an important record for the accountability of mods and tracking repeat offenses of certain users.

BUG STATUS: FIXED by Annika
 
Last edited by a moderator:
Ruft said:
Why is Shiny Zeraora with Knock Off legal? From my understanding, Zeraora can only learn Knock Off through USM Move Tutor and its Shiny version is only obtainable in Sword & Shield.
Click to expand...
This still needs fixing, though I understand if it has low priority.
 
During a battle (this one) the music suddenly stopped, I started only hearing pokemon cries.

I suspect this is a bug in PS! given what I found in my browser's console:

Code: 
Uncaught (in promise) DOMException: The media resource indicated by the src attribute or assigned media provider object was not suitable. battle-sound.js:134
    playSound https://play.pokemonshowdown.com/js/battle-sound.js?16512e2e:134
    playEffect https://play.pokemonshowdown.com/js/battle-sound.js?16512e2e:126
    animSummon https://play.pokemonshowdown.com/data/graphics.js?55507df6:2026
    animSummon https://play.pokemonshowdown.com/data/graphics.js?55507df6:1381
    switchIn https://play.pokemonshowdown.com/js/battle.js?3556fc3f:801
    runMajor https://play.pokemonshowdown.com/js/battle.js?3556fc3f:3288
    run https://play.pokemonshowdown.com/js/battle.js?3556fc3f:3403
    nextActivity https://play.pokemonshowdown.com/js/battle.js?3556fc3f:3500
    add https://play.pokemonshowdown.com/js/battle.js?3556fc3f:3072
    add https://play.pokemonshowdown.com/js/client-battle.js?c66fa815:220
    receive https://play.pokemonshowdown.com/js/client-battle.js?c66fa815:107
    receive https://play.pokemonshowdown.com/js/client.js?efa29945:931
    onmessage https://play.pokemonshowdown.com/js/client.js?efa29945:764
    dispatchEvent https://play.pokemonshowdown.com/js/lib/sockjs-1.4.0-nwjsfix.min.js:2
    _transportMessage https://play.pokemonshowdown.com/js/lib/sockjs-1.4.0-nwjsfix.min.js:2
    forEach self-hosted:206
    _transportMessage https://play.pokemonshowdown.com/js/lib/sockjs-1.4.0-nwjsfix.min.js:2
    <anonymous> self-hosted:1161
    emit https://play.pokemonshowdown.com/js/lib/sockjs-1.4.0-nwjsfix.min.js:2
    onmessage https://play.pokemonshowdown.com/js/lib/sockjs-1.4.0-nwjsfix.min.js:2

Firefox 81.0+linuxmint1+tricia, Linux Mint 19.3 Cinnamon 64b

BUG STATUS: NEEDS MORE INFO, I'm unable to reproduce this specific console log (DaWoblefet). Whatever happened is a bug, though. I can reproduce a similar effect of music disappearing by pressing "Last turn", however.
 
Last edited by a moderator:
Several users have independently reported problems with custom challenges throwing anomalous errors. I can't actually replicate this myself, but I can give examples and some details about what I think is happening.

Here is an example of a challenge code that works for me and I believe is intended to be supported, but a user has reported throwing an error:-

/challenge USERNAME, [Gen 8] Mix and Mega @@@STABmons Move Legality, *Acupressure, *Belly Drum, *Bolt Beak, *Boomburst, *Double Iron Bash, *Extreme Speed, *Fishious Rend, *Shell Smash, *Shift Gear, *Spore, *V-create, *Wicked Blow, +Eternatus, +Lunala, *Blissey, *Chansey, *Eternatus, *Lunala, *Magearna

When the user tried to challenge using this code, Showdown apparently showed the following error:-

Unrecognized rule " *Acupressure"

When I looked into this, the direct cause of the failure is understandable, because Showdown seems to just look at the first character in the alteration to determine what operator, if any, is being intended to apply, otherwise treating it as a rule (dex.ts: validateRule). But here, it is looking at a leading space instead of the '*', causing it to be miscategorized as a rule and not understood.

However, I don't understand how it the space was left in in the first place. If I try to challenge after executing this same same command, no error occurs and in the preview before executing it, it looks as if the client(?) has already trimmed out the spaces:-

Screen Shot 2020-10-03 at 21.13.36.png


So I think this aspect is some kind of bug with the client. I asked the user who ran into the error using this challenge code about their browser, which was apparently Chrome. I usually use Firefox (OS X) and tried making the challenge on Chrome, but I still couldn't replicate the error. (Unfortunately, I wasn't able to find out what operating system the user who reported this specific case was using.)

After searching the client source a bit, it looks like the area in client-chat.js around "'challenge'" manages this, but it seems as if it is properly trimming spaces from the clauses...is it possible that this method might not be supported properly in some environments? From what I've seen, this problem seems to be somewhat common for users creating custom challenges. If I get more reports of this, I will try to find out out more information about the users' environments and add it to this post.

I also had the thought that, even if the intention is that the client is responsible for cleaning up these commands, it might be good to do it on the server as well (e.g. in dex.ts, trimming "rule" before validateFormat is called(?))

Workaround: it seems that you can avoid this problem by manually deleting any spaces around commas in a challenge code, although this might be a bit annoying when making challenge codes by adapting existing longer tour codes.

BUG STATUS: NEEDS MORE INFO, I also am unable to replicate this (DaWoblefet)

EDIT: Okay, I was able to contact the user mentioned above and get more specific information about their environment:-
OS: Windows 10
Browser: Chrome Version 85.0.4183.121
 
Last edited:
Status
Not open for further replies.
Back
Top