[Bugs] Moth Lamp

[simplex]

Yes, it's completely blank still. I'm running the latest commit on the development branch. :/

 

Don't Starve hates me today. >_<

:/

Well, I tried something else... See if it's still empty (this is getting so annoying, for some reason it's not working precisely on the systems it'd need to ).

[debugman18]

:/

Well, I tried something else... See if it's still empty (this is getting so annoying, for some reason it's not working precisely on the systems it'd need to ).

Still empty. I'll try on my laptop when I leave here.

[squeek]

trace_flow() crashes me too (Windows 7). trace.txt gets created but is empty. With Better Console enabled, my log.txt ends with:

> trace_flow()

Also, there's a new crash on startup due to line 108 in scripts/upandaway/main.lua. It references self.bottom_left_stuff which no longer exists in screens/mainscreen.lua (in the newly updated public preview version). I just commented out the mod credits button code in order to test trace_flow().

[squeek]

The crash in trace_flow() on Windows is due to the os.date() format containing %F. Simply calling os.date("%F") from console crashes me. Calling it without %F does not (os.date("%X") works as expected). Should just use %c, which gives a date like "10/17/13 22:22:52"

EDIT: And ClimbTo(1,1) works fine for me.

[simplex]

Also, there's a new crash on startup due to line 108 in scripts/upandaway/main.lua. It references self.bottom_left_stuff which no longer exists in screens/mainscreen.lua (in the newly updated public preview version). I just commented out the mod credits button code in order to test trace_flow().

 

Apparently bottom_left_stuff was just removed. Upon reading this, I opened mainscreen.lua (public preview version) to check and it was still there, so by the time of my last reply here yesterday it was still intact. Then I opened Steam, the game updated and it was gone. I just disabled the main screen modifications for now, since there are more pressing issues to attend.

 

The crash in trace_flow() on Windows is due to the os.date() format containing %F. Simply calling os.date("%F") from console crashes me. Calling it without %F does not (os.date("%X") works as expected). Should just use %c, which gives a date like "10/17/13 22:22:52"

EDIT: And ClimbTo(1,1) works fine for me.

Thanks for catching that. Apparently Windows does not support the %F format specification, which is a shame, since I like to always use ISO date formats. I replaced it with "%Y-%m-%d".

[debugman18]

@simplex

trace_flow() works as it should now. If you still want my trace.txt let me know and I'll upload it. However, it's probably irrelevant, because this time the game actually gave me an error.

 

...amApps/common/dont_starve/data/scripts/saveindex.lua:345: SaveIndex:GetSaveData: Load failed for file [cave_1_-1_1_staging] please consider deleting this save slot and trying again.LUA ERROR stack traceback:        =[C] in function 'assert'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/saveindex.lua(345,1)        =[C] in function 'GetPersistentString'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/saveindex.lua(344,1) in function 'GetSaveData'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/gamelogic.lua(901,1) in function 'DoLoadWorld'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/gamelogic.lua(948,1) in function 'LoadSlot'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/gamelogic.lua(1012,1) in function 'DoResetAction'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/gamelogic.lua(1047,1) in function 'complete_callback'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/upsell.lua(27,1) in function 'UpdateGamePurchasedState'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/gamelogic.lua(1052,1)        =[C] in function 'SetPersistentString'    ...        =[C] in function 'GetPersistentString'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/saveindex.lua(67,1) in function 'Load'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/gamelogic.lua(1068,1) in function 'callback'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/playerprofile.lua(378,1) in function 'Set'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/playerprofile.lua(278,1)        =[C] in function 'GetPersistentString'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/playerprofile.lua(276,1) in function 'Load'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/gamelogic.lua(1067,1) in main chunk        =[C] in function 'require'        C:/Program Files (x86)/Steam/SteamApps/common/dont_starve/data/scripts/mainfunctions.lua(598,1)

 

This was a fresh save.

[simplex]

@simplex

trace_flow() works as it should now. If you still want my trace.txt let me know and I'll upload it. However, it's probably irrelevant, because this time the game actually gave me an error.

Yes, trace.txt would only matter for a silent crash. Could you upload the cave_1_-1_1_staging savefile?

EDIT: But if you manage to get the silent crash again, please post trace.txt as well. However, it's entirely possible the last public preview update changed the nature of the (visible) error permanently.

[debugman18]

Yes, trace.txt would only matter for a silent crash. Could you upload the cave_1_-1_1_staging savefile?

There isn't a cave_1_-1_1_staging file, although the saveslot says it's Cave Depth -1. There is a regular one, but it's from an earlier save.

 

 

Here's the first save, if it might help:

 

http://www.mediafire.com/view/pymza9na2prpruu/survival_1_staging

 

Edit: Annnnd now everything is working just fine. Made a fresh save, and did ClimbTo(1,1) without trace_flow() and it worked as it should. I'm confused as hell. If I see it again, I'll use trace_flow().

 

@Craig_Perry

Are you still experiencing this crash?

[Craig_Perry]

@Craig_Perry

Are you still experiencing this crash?

well I just tried to test it... but when the warning message popped up I click "I understand" and it crashed I've got the cause here:

 

 

...ods/UpAndAway-development/scripts/upandaway/main.lua:108: attempt to index field 'bottom_left_stuff' (a nil value)

[simplex]

well I just tried to test it... but when the warning message popped up I click "I understand" and it crashed I've got the cause here:

 

 

...ods/UpAndAway-development/scripts/upandaway/main.lua:108: attempt to index field 'bottom_left_stuff' (a nil value)

Seriously, I don't mean to be rude, but

• Read the thread. This was just discussed with squeek.
• Sync the mod code. This issue was solved about 7 hours ago.

[simplex]

@debugman18

It seems the save file is not being generated, then. Though a more informative error message from the game would help.

There isn't a file worldgen_log.txt being created in the mod folder, is there? (I'm guessing not, since it'd only catch errors triggered from within the mod itself)

[Craig_Perry]

Seriously, I don't mean to be rude, but

• Read the thread. This was just discussed with squeek.
• Sync the mod code. This issue was solved about 7 hours ago.

 

wait, I just synced it then re-extracted the zip file... what is going wrong here

 

I should've read further back apologies

[simplex]

wait, I just synced it then re-extracted the zip file... what is going wrong here

Re-extracted the zip file? So by sync you mean redownload the development.zip archive, right?

If so, my guess would be that GitHub doesn't necessarily always keep the zip in the latest version. It wouldn't be reasonable for it to rezip entire repos every time someone uses the link, so they probably keep a cached copy used for a period of time before updating it.

[Craig_Perry]

Re-extracted the zip file? So by sync you mean redownload the development.zip archive, right?

If so, my guess would be that GitHub doesn't necessarily always keep the zip in the latest version. It wouldn't be reasonable for it to rezip entire repos every time someone uses the link, so they probably keep a cached copy used for a period of time before updating it.

I'll try it again, never used github before, probably why Im messing up all the time

[simplex]

I'll try it again, never used github before, probably why Im messing up all the time

No problem, there's a learning curve to it. I never used the GitHub for Windows app, since I use git directly, so I feel I can't help you much on these details. But its documentation seems to be quite good.

[Craig_Perry]

ok, climbing a beanstalk doesnt crash anymore, but theres no beanstalk down

 

using trace_flow() makes it completely unplayable

[simplex]

using trace_flow() makes it completely unplayable

It does. It's meant to be used only just before a crash (not to mention trace.txt gets huge very quickly).

In a related note, I disabled the file reopening scheme I had put in when trying to work around the empty trace.txt. I don't think it's really necessary, and it makes trace_flow() run much slower. After working my memory some more, I believe the non-conformant behaviour in fflush() under Windows is just for input streams, not output, so this shouldn't be necessary.

Additionally, now all logs (worldgen_log.txt, entity_creation_profiling.txt and trace.txt, so far) go into a folder called log/, inside the mod folder. Now instead of explicitly ignoring every log file in .gitignore I'm just ignoring the whole contents of log/, which is cleaner and easier to maintain. The (empty*) log folder, however, is a part of the repository.

* There's just a .gitignore in it.

[Craig_Perry]

It does. It's meant to be used only just before a crash (not to mention trace.txt gets huge very quickly).

In a related note, I disabled the file reopening scheme I had put in when trying to work around the empty trace.txt. I don't think it's really necessary, and it makes trace_flow() run much slower. After working my memory some more, I believe the non-conformant behaviour in fflush() under Windows is just for input streams, not output, so this shouldn't be necessary.

Additionally, now all logs (worldgen_log.txt, entity_creation_profiling.txt and trace.txt, so far) go into a folder called log/, inside the mod folder. Now instead of explicitly ignoring every log file in .gitignore I'm just ignoring the whole contents of log/, which is cleaner and easier to maintain. The (empty) log folder, however, is a part of the repository.

ah I see, I'll be sure to use the trace when I can

 

on the log thing; sounds great, it'll make the log much faster to read

[simplex]

ah I see, I'll be sure to use the trace when I can

 

on the log thing; sounds great, it'll make the log much faster to read

No, no. It'll still make the game quite slow (not as much as now, but still unplayable), and as I said trace.txt gets massive very fast (take for example the one I posted, just the trace for calling ClimbTo(1,1) was ~80MB big). It should be used only to track down silent crashes which don't provide an error message (not on screen and not in log.txt). It's just a last resort kind of thing.

[Craig_Perry]

No, no. It'll still make the game quite slow (not as much as now, but still unplayable), and as I said trace.txt gets massive very fast (take for example the one I posted, just the trace for calling ClimbTo(1,1) was ~80MB big). It should be used only to track down silent crashes which don't provide an error message (not on screen and not in log.txt). It's just a last resort kind of thing.

ah, gotcha I was about to do a trace for evrything ah that'd have been real silly, I'll go back and see what I can find

 

EDIT: got a bug, the golden eggs are now "missing name" and the textures and function are not there

[debugman18]

So I think this should be used for bug reporting, as it should be.

 

So here's a bug that just occured, having to do with skyflies:

Stale Component Reference: GUID 109874, @scripts/entityscript.lua:766scripts/entityscript.lua:769: Something is wrong: inst.Transform:GetWorldPosition() stale component reference?Entity information:    entity: [109838 - skyflies]    variable name: self    valid: true    in limbo: false    asleep: false    time alive (seconds): 16.03    position: (-310.77, 0.00, -257.75)    distance to player: 23.90LUA ERROR stack traceback:        =[C] in function 'assert'        scripts/entityscript.lua(769,1) in function 'GetDistanceSqToInst'        ../mods/UpAndAway/code/brains/skyflybrain.lua(15,1)        =(tail call) ?        =(tail call) ?        scripts/behaviours/follow.lua(39,1) in function 'Visit'        scripts/behaviourtree.lua(558,1) in function 'Visit'        scripts/behaviourtree.lua(22,1) in function 'Update'        scripts/brain.lua(212,1) in function 'OnUpdate'        scripts/brain.lua(135,1) in function 'Update'        scripts/update.lua(125,1)        =(tail call) ?        =[C] in function 'xpcall'        ../mods/BetterCrashes/modmain.lua(165,1)LUA ERROR stack traceback:        =[C] in function 'error'        ../mods/BetterCrashes/modmain.lua(154,1)        =(tail call) ?scripts/frontend.lua(712,1) SCRIPT ERROR! Showing error screen  

[simplex]

So I think this should be used for bug reporting, as it should be.

 

So here's a bug that just occured, having to do with skyflies:

Were you playing as Wendy? Had Abigail recently despawned?

[debugman18]

Were you playing as Wendy? Had Abigail recently despawned?

I was playing as Wendy, yes, but I had no Abigail. I had tried to summon Abigail, but to no avail.

[simplex]

I was playing as Wendy, yes, but I had no Abigail. I had tried to summon Abigail, but to no avail.

What happened when you tried to summon her? Did something strange made her not appear when she should?

The error was caused by a skyfly following an entity which became invalid (the actual crash happened due to a inst.Transform:GetWorldPosition() returning nil), and skyflies only follow entities with the "character" tag (is that really correct or did you mean the "player" tag?). The only entities in U&A which have this tag are owl, duckraptor and Moose, but I don't see how any of these could've become invalid without getting killed (and the skyfly's brain tests for that). The entities in vanilla with the tag are abigail, bunnyman, chester, mandrake, merm, perd, pigman, player, rocky, smallbird and walrus, and I don't see any fitting the scenario other than abigail.

Anyway, this should be fixed by having skyflybrain test for the validity of its leader, which I just put in and pushed to the repo. But it'd be nice to know what was the cause.

[debugman18]

What happened when you tried to summon her? Did something strange made her not appear when she should?

The error was caused by a skyfly following an entity which became invalid (the actual crash happened due to a inst.Transform:GetWorldPosition() returning nil), and skyflies only follow entities with the "character" tag (is that really correct or did you mean the "player" tag?). The only entities in U&A which have this tag are owl, duckraptor and Moose, but I don't see how any of these could've become invalid without getting killed (and the skyfly's brain tests for that). The entities in vanilla with the tag are abigail, bunnyman, chester, mandrake, merm, perd, pigman, player, rocky, smallbird and walrus, and I don't see any fitting the scenario other than abigail.

Anyway, this should be fixed by having skyflybrain test for the validity of its leader, which I just put in and pushed to the repo. But it'd be nice to know what was the cause.

She didn't appear, unless I missed it by not paying attention. I'm not sure what caused it, as it was only something I did on a whim during testing.

 

I don't recall if I meant character or player, but I think player makes more sense.