The Official Status Thread
-
@Tsaukpaetra said in The Official Status Thread:
Status: Fucking wet and I don't know why...
Dry isn't that much fun.
-
@topspin said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Status: Fucking wet and I don't know why...
Dry isn't that much fun.
That's why God invented lube.
-
-
@error said in The Official Status Thread:
@pie_flavor said in The Official Status Thread:
@error said in The Official Status Thread:
Oh, and if somebody makes an SVG that takes parameters, I will turn it into a error_bot module for you.
Like remember the wtf-o-meter gif? I bet we could make one that's based on a percentage input, and probably even make the needle wiggle.
Of course you can already do this with discobot scripts.
I thought discobot scripts just printed ECC or not ECC?
The one I wrote does. You could just as easily write a script to parameterize an SVG.
-
Status: looking for a cheap, small apartment.
Found this one
That's it. That's the entire thing (minus two walls, a bathroom and an entrance).
I considered it, but I'm not sure there's even enough space for a decent desk and a chair.
-
@pie_flavor said in The Official Status Thread:
The one I wrote does. You could just as easily write a script to parameterize an SVG.
TBH it's both quicker and easier for me to write an error-bot module than to try and write a script for your bot.
-
@Tsaukpaetra said in The Official Status Thread:
@error said in The Official Status Thread:
somebody makes an SVG
Alright, I'm gonna have to stop you right there, bub...
Do you have something against SVG? It's a pretty decent format.
-
@error said in The Official Status Thread:
@pie_flavor said in The Official Status Thread:
The one I wrote does. You could just as easily write a script to parameterize an SVG.
TBH it's both quicker and easier for me to write an error-bot module than to try and write a script for your bot.
Only if you are lazy.
-
@error said in The Official Status Thread:
@topspin said in The Official Status Thread:
@error said in The Official Status Thread:
Minesweeper solver now uses a divide-and-conquer strategy, meaning an operation that previously took e.g. 240 steps now takes 4(210) steps.
Divide and conquer is meant to be recursive.
Not sure what you mean here, but it does recurse.
Also, Those 212 steps are sent to 8 different cores, so it ends up taking 212-log28=29 time.
The reduction you mentioned (from 240 steps to 4(210) steps) sounds like what you'd get if you divide just once. And the number is still fucking huge. (Not that I'd expect it to be reducible to a small number)
Let's compare it to sorting problem in naive O(n^2). You'd say, e.g., "it started with 1000^2 steps", but then you applied divide and conquer and arrived at "2*500^2 + 1000 steps". But that was just a single division / merge, whereas if applied recursively you end up with the well known O(n log n) algorithms, i.e. "~ 1000 log(1000) steps".Of course, I assume you know all this and did it correctly, it was just meant as a joke "you did it wrong" because the result is still so large (depending on what a "step" costs).
-
@anonymous234 said in The Official Status Thread:
Status: looking for a cheap, small apartment.
Found this one
That's it. That's the entire thing (minus two walls, a bathroom and an entrance).
I considered it, but I'm not sure there's even enough space for a decent desk and a chair.
Looks like a typical poor-student single room apartment of 20 m^2.
Wouldn't want to live that crowded.
-
@topspin said in The Official Status Thread:
The reduction you mentioned (from 240 steps to 4(210) steps) sounds like what you'd get if you divide just once. And the number is still fucking huge. (Not that I'd expect it to be reducible to a small number)
A reduction from 1,099,511,627,776 to 4,096 sounds trivial?
-
@error said in The Official Status Thread:
@topspin said in The Official Status Thread:
The reduction you mentioned (from 240 steps to 4(210) steps) sounds like what you'd get if you divide just once. And the number is still fucking huge. (Not that I'd expect it to be reducible to a small number)
A reduction from 1,099,511,627,776 to 4,096 sounds trivial?
As I said, I wasn't serious.
-
@error Let's just say I communicated the joke badly. Happens.
-
@topspin said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Status: Fucking wet and I don't know why...
Dry isn't that much fun.
This is the wet that's not fun.
-
@Tsaukpaetra said in The Official Status Thread:
@topspin said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Status: Fucking wet and I don't know why...
Dry isn't that much fun.
This is the wet that's not fun.
Don't kink shame watersports.
-
@error said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
@topspin said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Status: Fucking wet and I don't know why...
Dry isn't that much fun.
This is the wet that's not fun.
Don't kink shame watersports.
Even that's more fun than my current state....
-
Status: Shit shit shit shit fuck blast piss OMG
Not-my-code is suddenly fucked and I don't know why...
Edit: Does anyone know if trying to find chkstk.asm would be helpful in any way?
-
@Tsaukpaetra said in The Official Status Thread:
Does anyone know if trying to find chkstk.asm would be helpful in any way?
Based on filename (check stack?) it's probably the code detecting/raising the stack overflow.
-
@error said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Does anyone know if trying to find chkstk.asm would be helpful in any way?
Based on filename (check stack?) it's probably the code detecting/raising the stack overflow.
Indeed! But I can't see anything about the stack that's overflowed (or I don't know how). I don't really know how to debug this problem, it's never really happened unless doing recursion, and that's definitely not happening here...
-
@Vixen Change that middle plus to two dashes and you'll have a flag
-
@Tsaukpaetra said in The Official Status Thread:
I don't know why...
If I early-return from the main
while(dontDie)
loop, it seems fine, but naturally dies after one loop (undesirable).If I remove that
return 0;
, instacrash at the top of the function.Methinks I need to revisit C++ 101 to figure out when things get allocated, and when they go to the stack instead of heap...
-
@hungrier said in The Official Status Thread:
@Vixen Change that middle plus to two dashes and you'll have a flag
true.... true.... i could do that.....
-
@Tsaukpaetra said in The Official Status Thread:
If I early-return from the main
while(dontDie)
loop, it seems fine, but naturally dies after one loop (undesirable).If I comment out the
while
itself, crash partway.Even better!
-
status that was the most painful set of dungeon/raid runs I've done in a long time. And I can't even complain in game, because the people in the last one were in my FC (guild). As a tank, I healed and dpsed (with one other DPS) the last boss from 30% on our third try. In the raid, I tanked (as DPS this time) a mob because our tank dc'd. Ugh.
-
@Benjamin-Hall said in The Official Status Thread:
I tanked (as DPS this time)
Sounds like a Red Mage to me!
-
-
@TwelveBaud program it, I dare you!
-
@error said in The Official Status Thread:
@Vixen said in The Official Status Thread:
# Status #
-annoyed grunt-
I wrote a quick-and-dirty SVG URL maker for you: https://errorx666.github.io/svg-preview/
You can open the above image in a new tab, view source, copy, then paste that into the webapp.
Maybe I'll make it so you can paste the data URL in the output box in the webapp, but
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 220 100"> <style> text { dominant-baseline: hanging; text-anchor: middle; font: 13px Tahoma,Arial,sans-serif; } </style> <line x1="50%" y1="20" x2="50%" y2="100%" stroke="#000"/> <rect x="10%" y="20" width="80%" height="50" stroke="#000" fill="#fff"/> <text x="50%" y="30">Clothing Optional</text> <text x="50%" y="50">Beyond This Point</text> </svg>
Geez, signatures are limited to 255 characters? I can barely fit a single line SVG data URL in that.
Edit: oh, and no markup? Signatures suck now.
-
@error said in The Official Status Thread:
Signatures suck now.
I though it was an attempt to reduced the attack surface for signature guy.
-
@Tsaukpaetra said in The Official Status Thread:
@error said in The Official Status Thread:
Signatures suck now.
I though it was an attempt to reduced the attack surface for signature guy.
That's what I just said.
-
-
Status: Am I going to regret signing up on a instance named
witches.live
, aren't I?
-
@Tsaukpaetra
Lernoud & Hauspie, duh
-
@Luhmann said in The Official Status Thread:
@Tsaukpaetra
Lernoud & Hauspie, duhClassic, but this definitely wasn't that.
-
@Tsaukpaetra said in The Official Status Thread:
But I can't see anything about the stack that's overflowed (or I don't know how).
Shouldn't you see the stack the moment you break into the debugger when the overflow occurs? Check the Call Stack window.
-
@topspin said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
But I can't see anything about the stack that's overflowed (or I don't know how).
Shouldn't you see the stack the moment you break into the debugger when the overflow occurs? Check the Call Stack window.
It's not the Call Stack that's overflowed though, it's the actual allocated stack space. The Call Stack is three entries long (unless it's dove into the sqlclient dll, then it's about 9 lines long).
Edit: It's 14 lines long:
@Tsaukpaetra said in The Official Status Thread:
-
@e4tmyl33t I actually was RDM for that run.
-
@Tsaukpaetra said in The Official Status Thread:
@topspin said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
But I can't see anything about the stack that's overflowed (or I don't know how).
Shouldn't you see the stack the moment you break into the debugger when the overflow occurs? Check the Call Stack window.
It's not the Call Stack that's overflowed though, it's the actual allocated stack space. The Call Stack is three entries long (unless it's dove into the sqlclient dll, then it's about 9 lines long).
You're on a Von Neumann architecture, there's only one stack for data and return addresses. If the call stack isn't deep, it usually doesn't overflow unless a single function allocates a shit ton on in a single frame. Which it usually doesn't because then it would always overflow.
Check the local variables (or whatever it's called) window then, and if there's nothing weird in the current frame, go through the 3-9 calls and check the variables there. (Haven't used VS in a while, but I assume if you double click an entry in the call stack window it jumps to the corresponding position and displays the local variables of that frame instead). There's got to be something that uses too much memory, or else there's something really weird going on.
-
@topspin said in The Official Status Thread:
Check the local variables (or whatever it's called) window then, and if there's nothing weird in the current frame,
The current frame is before the function started. If memory serves there wasn't anything allocated in Locals but I'll double-check tomorrow.
@topspin said in The Official Status Thread:
go through the 3-9 calls and check the variables there.
Can't. I don't have the source code for Microsoft's SQL Native Client. It's a really dark black box... And the frames I do see don't have anything untoward going on (and didn't before the Engine upgrade).
I'll iterate again, this was working fine before not-my-code was updated elsewhere.
I set a clean compile (which takes some 6 hours or so) and we'll see if something shitty was cached or whatever.
@topspin said in The Official Status Thread:
or else there's something really weird going on.
Well, when previously working-great untouched code suddenly goes batshit insane, really weird is the first place I'd think things were going on.
-
@Tsaukpaetra said in The Official Status Thread:
@topspin said in The Official Status Thread:
Check the local variables (or whatever it's called) window then, and if there's nothing weird in the current frame,
The current frame is before the function started. If memory serves there wasn't anything allocated in Locals but I'll double-check tomorrow.
I see, I assume that means the current function wants to allocate a lot of things on the stack, calls into chkstk to see if that's possible, then fails. The current function is most likely the culprit, but could of course be unlucky in that the one before used "almost all" the stack, but left just enough to call in another to break that.
@topspin said in The Official Status Thread:
go through the 3-9 calls and check the variables there.
Can't. I don't have the source code for Microsoft's SQL Native Client. It's a really dark black box... And the frames I do see don't have anything untoward going on (and didn't before the Engine upgrade).
That doesn't make things easier.
-
@topspin said in The Official Status Thread:
used "almost all" the stack
I'm pondering if the thread stack size default was changed or something. Epic would do something like that just for fun.
-
@Vixen It turned the sign into a flag. nice.
-
Status: Oops.
-
Status
That's better......
-
@Vixen said in The Official Status Thread:
Status
That's better......
Remind me to send you a windows firefox screenshot when I get home.
-
Also even on yours one of the | doesn't line up.
-
@PleegWat said in The Official Status Thread:
@Vixen said in The Official Status Thread:
Status
That's better......
Remind me to send you a windows firefox screenshot when I get home.
let me guess.... it displays that differently becazuse of course font rendering is different on Firefox.......
...... looks okay to me so it't can just be firefox......
-
@Vixen said in The Official Status Thread:
let me guess.... it displays that differently becazuse of course font rendering is different on Firefox.......
Too many variables when trying to line things up using variable-width fonts.
-
@PleegWat said in The Official Status Thread:
@Vixen said in The Official Status Thread:
let me guess.... it displays that differently becazuse of course font rendering is different on Firefox.......
Too many variables when trying to line things up using variable-width fonts.
i wouldn't be trying to if i could use formatting in my signature
or an image....
-
@Vixen Send a proposal to Unicode and use that whenever it's added to the standard
U+DFF69 CLOTHING OPTIONAL BEYOND THIS POINT SIGN