The Official Status Thread
-
-
I moved it this morning to see if the cooties situation improves.
My apologies, then, to @ben_lubar for blaming him. My is now directed at you.
-
He suggested it.
-
Hm, I'm not sure I have enough rotten, moldy s to throw at both of you.
-
-
Status: I just got an email that a FOSS project of which I am a user implemented a feature I suggested. Now I just have to figure out what I had in mind when I suggested it >4.5 years ago. :)
-
Status****strong text haunted house done for another year. Exhausted. Will strike set tomorrow.
Today was supposed to have no rain. But right at 6pm, it rained (going right until 10pm). On the dot. It's rained every Halloween for the past 5 years. Stupid weather.
At least it meant the people who did come to go through the haunt were the ones who REALLY wanted to be there. Mwhahahahahah!
-
I can't guarantee that the objects are a DAG because doubly-linked lists and trees with parent pointers are possible in Cool. I can, however, guarantee that the stack is a DAG (more specifically, a singly-linked list with extra edges that point out into the heap). Nothing ever gets a pointer to the stack itself.
Are we talking references or pointers for the reverse links? The difference (in my interpretation of my question) is that references maintain some sense of ownership of their target, whereas pointers don't. A child tree node might maintain a pointer to their parent, but wouldn't have a reference. (If you prefer, call pointers “weak references”.)
Ring buffers implemented with linked lists are a harder case. Those necessarily introduce a true loop in the graph.
-
Status: My computer had been up long enough that it really needed to be rebooted, but bluescreening during the night while I was asleep isn't exactly the preferred method of accomplishing that.
-
STATUS: My least favorite part of starting a new OSS project - thinking of a name. So far, I've had stupid, obscure, duplicate and boring, in various combinations. Hope I beat the odds this time.
-
STATUS: My least favorite part of starting a new OSS project - thinking of a name. So far, I've had stupid, obscure, duplicate and boring, in various combinations. Hope I beat the odds this time.
Brain: How about the noun associated with product + "tastic"?
Common sense: How about fuck you.
-
Yeah crabtastic might be a sub optimal choice
-
Yeah crabtastic might be a sub optimal choice
Sorry I was away buying my xxxtastic.com domain name.
What?
I didn't say the common sense won.
-
-
That was still available? How did the p0rn industry miss this one?
WTF, it actually is available!
And it's on sale, only $2!
-
What are you waiting for? It's a bargain!
-
You know, I've seen a loop like that before, and it actually did what its name would imply in a sane world; it sped up the operation of the code following it. For a while in Second Life, running ; a shitton of times would get the server to allocate an inordinate amount of processing time to that script, effectively supercharging your script at the expense of... everything else. I had one little monstrosity that did that with 350 scripts at once. Brought the server to a near standstill, using 95-99% of its resources for a single function call repeated hundreds of thousands of times per second.
-
-
Yeah, but that was on a different album.
-
Still had Enema in the title.
-
Oh! Yeah, that's a good point.
-
Status: My wife wants to watch 'Se7en' tonight since it is Halloween. My only thought is, "I got my first blowjob while watching that movie".
/me reads this right after listening to 'Family Reunion' (which is usually paired with 'The Blow Job Song'), tries to resist asking, 'speaking of which, how is your mother these days', unable to do so
-
http://puu.sh/l5x3m/284c1df833.jpg
Hmm... what happens if I click. Oh I actually get @undefined's profile
-
It's like what Java has. There's no way to get something that isn't a pointer to an object.
-
What a shocker. Brafox hangs out on Second Life. WHODATHUNKIT
-
Status: You know how you sometimes push tasks you don't like until you've reached a deadline and you can't avoid said task anymore?
Well, that just happened to my Astronomy exam. And good grief, did the pupils figuratively murder the good Kepler and his laws. I mean, not only did they largely botch using the provided formulas to calculate stuff like escape velocity, the orbital radius of a geostationary satellite or the force of gravity on the moon, no they also all managed to make a mess of stuff they were supposed to remember by rote. Stuff I had expressly told them to learn.
I mean, it's not as if it was much that I expected:
a) an idea of which planet goes where in the solar system and what it roughly looks like (e.g. Saturn has rings, Mars is quite small compared to Earth,...) and
b) the three Kepler's Laws and
c) where the seasons come from (the tilt of the Earth's axis) and why summer is of a different length than winter (the result of Kepler's 2nd law).I'm now wondering what I did wrong or if I indeed went wrong somewhere. I sometimes think that it's part of this trend of what we call "bulimia learning" - you binge learn for one or two days and then regurgitate it back up.
That simply does not work with physics.
It also does not help that they're intimidated by any formula which contains more than two variables. Plus, astronomy deals with quite large distances so we also need to use scientific notation of numbers.
Argh.
-
I'm now wondering what I did wrong
Did you do interim tests? Just short ones, but nonetheless demonstrating that you really were expecting them to learn the stuff?
-
@Rhywden said:
I'm now wondering what I did wrong
Did you do interim tests? Just short ones, but nonetheless demonstrating that you really were expecting them to learn the stuff?
No, I did not. I gave them homework, however. But you're right, I'll probably step up the test frequency.
Still, tests or not, does not excuse them for taking my "Learn this" not seriously, especially when we went through that stuff repeatedly by different methods.
-
But you're right, I'll probably step up the test frequency.
It's what my teachers used to do at school. It seemed to work.
does not excuse them for taking my "Learn this" not seriously
Against wilful stupidity (i.e., deliberately ignoring your advice) there's only a limited amount that you can do. More frequent testing might let you hammer home that not learning what you tell them to will have consequences before it actually does…
-
@Rhywden said:
But you're right, I'll probably step up the test frequency.
It's what my teachers used to do at school. It seemed to work.
As I said, good idea. Hmmh, maybe a short 5-minutes test every week? I could also re-use some of the questions in the next exam... the opportunities!
-
Hmmh, maybe a short 5-minutes test every week?
No idea. I was one of those irritating types that used to aim for 100% on every test out of sheer boredom…
-
@Rhywden said:
Hmmh, maybe a short 5-minutes test every week?
No idea. I was one of those irritating types that used to aim for 100% on every test out of sheer boredom…
I'll see how it goes. They'll complain, of course, but the "parts of the tests will be in the next exam" should mitigate that a bit.
There's only the question about the current exam - shall I have them write it again or use it as a shot across the bow? Oh well, will take it up with the principal tomorrow, have to talk to him anyway.
-
Status : Removing the "666 unread items in this thread" notification from this thread by posting to it.
-
Sorry I was away buying my xxxtastic.com domain name.
Huh I just tried the most obvious with your pattern and it seems to be available too! fap
ntastic, hurry up before someone else here buys it.
-
Status: Waiting for the people playing the Mafia game to stop talking and start lynching!
-
MOD ABUSE!
-
Isn't @TwelveBaud having a little time this weekend anyway?
-
@ben_lubar: Did you ever get a chance to look at either of the Richard Jones books on GC? I would expect the older one to be in most college libraries, at the very least, and if you can read the newer one, it has a lot of good details on optimization, especially regarding generational and adaptive methods.
While I doubt that you'd need adaptive methods, it does sound like a generational approach would make sense for this project, especially if there is a significant amount of static and/or long-duration objects.
I've been looking into adaptive GC, as I want to be able to efficiently handle several different use cases, even within a single program - I intend to allow different libraries to handle their own GC, especially those for logic, relational, and constraint programming (which have unique constraints and optimizations), though most would presumably use whatever the default is, or the method selected by the program as a whole. I am seriously considering syncing GC with paging in Kether, allowing groups of pages or even individual pages to handle their own GC, and conversely, letting the pager invoke the garbage collectors to reduce the need for page-outs.
-
Isn't @TwelveBaud having a little time this weekend anyway?
Is he? I don't know. Looking at his profile...Last seen 3 minutes ago. I just have patience issues. Also, lots of thoughts, but don't want to share anything while the game is still going on.
-
I think he mentioned something in-thread about there being no deadline because he's spending a weekend at his fiancé's.
-
The Afterlife is great :3
-
Aren't you rotting in Hell or something?
-
We have cannolli!
-
And really good spagetts.
-
Status: rewrote string interning with comments because it was crashing and it's easier to start from scratch. Output of gdb:
Program received signal SIGSEGV, Segmentation fault. IO.symbol () at basic.s:164 164 movl offset_of_Symbol.name(%eax), %eax (gdb) info reg eax 0x6 6
How the fuck did it get 6 there?
-
Is SL's furry community really that unknown that you need to provide a google search to help prove your point? Shit, I guess we need to spread the word more.
-
.globl IO.symbol IO.symbol: enter $12, $0 // there's no state in IO, so release `this` movl 12(%ebp), %eax cmpl $0, gc_offset(%eax) jle 1f decl gc_offset(%eax) 1: // 8(%ebp) = the string we're looking for // -4(%ebp) = the address of the "next" field of the previous symbol // -8(%ebp) = the address of the symbol we're currently looking at // -12(%ebp) = our current hash value movl $symbol, -4(%ebp) movl symbol, %eax movl %eax, -8(%ebp) movl $0, -12(%ebp) // make sure we actually have a string movl 8(%ebp), %eax test %eax, %eax jz runtime.null_panic 2: // add 1 to our hash incl -12(%ebp) // are we at the end? movl -8(%ebp), %eax test %eax, %eax jz 3f movl offset_of_Symbol.name(%eax), %eax push %eax movl 8(%ebp), %eax push %eax call String.equals cmpl $boolean_true, %eax je 4f // go to the next one leal offset_of_Symbol.next(%eax), %eax movl %eax, -4(%ebp) movl (%eax), %eax movl %eax, -8(%ebp) jmp 2b 3: // box the hash movl $(size_of_Int + 4), %eax movl $tag_of_Int, %ebx call gc_alloc movl $gc_tag_root, gc_offset(%eax) movl -12(%ebp), %ebx movl %ebx, offset_of_Int.value(%eax) movl %eax, -12(%ebp) // mark the string and its length as a root movl 8(%ebp), %eax movl $gc_tag_root, gc_offset(%eax) movl offset_of_String.length(%eax), %eax movl $gc_tag_root, gc_offset(%eax) // make a new symbol object movl $size_of_Symbol, %eax movl $tag_of_Symbol, %ebx call gc_alloc movl $gc_tag_root, gc_offset(%eax) movl 8(%ebp), %ebx movl %ebx, offset_of_Symbol.name(%eax) movl -12(%ebp), %ebx movl %ebx, offset_of_Symbol.hash(%eax) // save it at the end of the list movl -4(%ebp), %ebx movl %eax, (%ebx) leave ret $8 4: // let the string be garbage collected movl 8(%ebp), %eax cmpl $0, gc_offset(%eax) jle 5f decl gc_offset(%eax) 5: // grab the symbol we found movl -8(%ebp), %eax leave ret $8
-
-
-
Status: I wrote a heap consistency tester thingy. Apparently I have an object with tag -1 (garbage) that is also marked as a root for the garbage collector. And then:
(gdb) bt #0 gc_check () at gc.s:291 #1 0x080488bb in gc_collect () at gc.s:165 Backtrace stopped: frame did not save the PC
Apparently, gdb can't understand more than one stack frame. Does the x86
enter
instruction not save the PC the way it wants to?