Bug - Not A Bug Autoadventuring not working when remote URL is unavailable

[Captain Kirk]

So I am trying to autoadventure in the Giant's Castle with no Goals.
The first time I try to do this I see:

Validating adventure sequence...
Requests complete.

(no adventures are used although, I've set it to 50 adventures). The pane is green.

So I try it a second time:

Validating adventure sequence...
(and it hangs out in gray mode).

I already restarted mafia and tried again and the exact same thing happened.

Any ideas on what I should do?

Update: It finally started to work, after a huge delay.. I know that a lot of my ash
scripts check things on the kolmafia forums... could that be the problem?

Update2: Tried to switch familiars.. from a He-Boulder to a Hobo monkey, got a huge delay and eventually this message showed up
(in a red pane):
_meatpermp => 4.090909
_meatpermp => 4.090909
_meatperhp => 0.7391199
_version_BalesUniversalRecovery => 3.891
Division by zero (Universal_recovery.ash, line 1547)
Your Hobo Monkey can't wear a quadroculars
Your Hobo Monkey can't wear a quadroculars

update 3:
I usually run the relay script "mad-tea-party" and usually it comes up instantaneously, but today it just
seemed to hang. Then after several minutes it finally appeared in the relay browser.
After the delay, I saw:
map_mad-tea-party.txt => , checked 20110812
You have the latest mad-tea-party.txt. Will not check again today.

 

[fronobulax]

Autoadventuring worked just fine for me with r9711 today. I did, however, notice many things that used scripts took longer than I expected. On one character it took over 3 minutes (compared to a more typical 60 seconds) to process my Login scripts and breakfast. Whenever a character first triggered UR, it took longer than usual, but everything ultimately succeeded. There is another report about excessively long delays due to timeouts when scripts go to a web site to check a file and the site is not answering.

 

[morgad]

It took about an hour before BCA (+ smartstasis + Universal recovery + counterchecker + zlib + ???) finished its' first combat for me
hope Zarqon sorts his server problem soon

Dave

 

[fronobulax]

It took about an hour before BCA (+ smartstasis + Universal recovery + counterchecker + zlib + ???) finished its' first combat for me
hope Zarqon sorts his server problem soon

Dave

Wow. If the time out is 5 minutes and the combined scripts look for 12 files, there's your hour, but still...

 

[Winterbay]

yeah, script timeouts could be handled slightly better possibly. Potentially Ibcould set a timeout limit somewhere. Or a message could be printed if no response is received after X seconds.
having to wait for a 5min timeout for each file when starting BCC, running into a fight with Batbrain gets troublesome without any kind of update on what is happening. And aborting doesn't abort the request either so...

 

[lostcalpolydude]

Thread title changed to hopefully be more accurate.

 

[fronobulax]

Thread title changed to hopefully be more accurate.

If we are all agreed that the problem is related to timing out on url visits in scripts then I believe this is probably not a bug or at least waiting for more information. There is evidence presented for successful auto-adventuring and every case where it failed there is also evidence presented that some script was running and had not completed when the failure occurred. So I would claim that the problem is that the operator did not wait for all scripts to finish before attempting to auto-adventure. That is still something that could be addressed but probably as a feature request or as part of something to deal with timeouts.

 

[Winterbay]

The problem is that you are not given any feedback on the timeout issue, and it's not just before autoadventuring, but also for scripts running during autoadventure (betweenbattke, recovery and consult scripts to name a few). I do however agree that it should probably be changed in some way to a feature request regarding timeouts instead.

 

[lostcalpolydude]

I'm not saying I changed it to a good title, I just didn't want the title to be completely inaccurate (based on the OP's edit, since they couldn't have known that to begin with and they couldn't change it later). Feel free to change it to make more sense. I'm unlikely to see this issue because the level of automation in those scripts is scary to me.

 

[fronobulax]

@lost - my apology if the comment on the title change caused any offense. Yours was a reasonable choice and my focus was really on the change in direction and not anything else.

Somewhere veracity posted her opinion on the use of the "Improve handling of semi-random lag spikes". My recollection is that she said one effect of that checkbox was to shorten timeouts. She observed that responses could come in after timeouts and there were many ways that mafia could get confused if that happened so she personally never used the feature. Her arguments there would seem to apply to url fetching in scripts so my expectation would be that the solution to this would be blocking other actions while a fetch was in progress or informing the user that a fetch was in progress.

 

[Winterbay]

I think informing the user about the wait is probably the best quick solution at least.

 

[Tom Sawyer]

Not sure if this related to my own problem in that the button greys out, or the relay browser just cycles.

I however again created a stand alone folder, logged in and other than checking run moods, maintain health, maintain mana during manual adventure nothing else is in play. No scripts, no moods.

After about 30 or so turns I started once more seeing in Firefox connecting as it just cycles. You can use the mini-browser to play a turn which WILL allow you then to click on firefox and play that turn from there. But you then again have to use the mini-browser option.