| --- Log opened Tue Mar 23 00:00:42 2010 |
| 00:23 | <@Vornicus> | Yay forkbomb! |
| 00:27 | | Thaqui [Thaqui@27B34E.D54D49.F53FA1.6A113C] has joined #code |
| 00:29 | | celticminstrel [celticminstre@Nightstar-f8b608eb.cable.rogers.com] has joined #code |
| 01:25 | <@Vornicus> | clemin: if you have an error in your php, some webservers are set up to not show you the error. |
| 01:29 | < celticminstrel> | Through various fiddlings, I determined that it was most likely a syntax error. |
| 01:30 | < celticminstrel> | More recently I had the problem that $_POST was empty. |
| 02:27 | | Attilla [Attilla@FBC920.642D35.7B2B85.11BB86] has quit [Connection reset by peer] |
| 02:28 | | Serah [Z@26ECB6.A4B64C.298B52.D80DA0] has quit [[NS] Quit: The world is so funny sometimes; I often forget to laugh.] |
| 02:28 | | * Orthia pokes Vorn timidly. |
| 02:30 | | Serah [Z@26ECB6.A4B64C.298B52.D80DA0] has joined #code |
| 02:30 | <@Vornicus> | ? |
| 02:35 | < Orthia> | Vorn: My brain has decided to cause me a few problems this week. I admit, I should have seen it coming, but it is problematic all the same. Would you be willing to treat me like an idiot and help me with this LZW code? Trying to use python as the template is a wonderful idea, but trying to convert between the two languages is sort of not working well. |
| 02:38 | <@Vornicus> | Okay. |
| 02:39 | <@Vornicus> | So let's deal with encoding a string. |
| 02:40 | <@Vornicus> | FirWe,, not a string, a byte sequence. |
| 02:40 | <@Vornicus> | wow, that was good. |
| 02:40 | <@Vornicus> | Well* |
| 02:41 | < Orthia> | OK |
| 02:43 | <@Vornicus> | LZW depends on the existence of a dictionary of byte-sequence-chunks, and loading that dictionary with stuff that we've seen. Our first step is to load up that dictionary. |
| 02:43 | < Orthia> | Logical. |
| 02:43 | <@Vornicus> | Well, to load up that dictionary with all the length-1 byte sequences. |
| 02:44 | <@Vornicus> | By doing this we make it so we don't have to carry around our starting dictionary in our file - our encoder and decoder pair know what the dictionary starts with, and that's good enough. |
| 02:45 | <@Vornicus> | Now, our problem here is that our encoder only works in nonnegative numbers, and bytes happen to be signed. |
| 02:47 | <@Vornicus> | fortunately, we know the smallest byte we can use: -128, also called Byte.MIN_VALUE |
| 02:48 | < Orthia> | ... so instead of 0 to 256, we will be going from Byte.MIN_VALUE to Byte.MAX_VALUE |
| 02:48 | <@Vornicus> | So we're going to load up our dictionary - which is to say, a Tree that we made - with bytes from -128 to 127, and with codes from 1 to 256. code 0 is reserved for the stop code; we do kind of need that. |
| 02:49 | <@Vornicus> | (later we're going to want to add a clear code as well; it will probably end up being 1, but it's a minor adjustment we can do after we've done the core algorithm) |
| 02:52 | < Orthia> | hmm. I see. |
| 02:52 | < Orthia> | Is there a Byte.NEXT_VALUE equivalent? |
| 02:52 | < Orthia> | Or can you go Byte++? |
| 02:52 | <@Vornicus> | my_byte++ |
| 02:52 | <@Vornicus> | Easier way to do that |
| 02:53 | <@Vornicus> | er. |
| 02:53 | | Serah [Z@26ECB6.A4B64C.298B52.D80DA0] has quit [Ping timeout: 121 seconds] |
| 02:54 | < Orthia> | OK |
| 02:54 | <@Vornicus> | An easier way to do this though is to iterate from 0 to 255, and then add Byte.MIN_VALUE to the result, and then cast the whole shebang to a byte. |
| 02:55 | < celticminstrel> | Isn't the next_value stuff only really meaningful for floating-points, where it's not at all obvious what the next highest possible value is? |
| 02:55 | <@Vornicus> | and then for your codes just add 1 to the result of your iteration. |
| 02:57 | <@Vornicus> | celticminstrel: It doesn't appear that that function even exists. |
| 02:59 | < Orthia> | OK! |
| 03:00 | < celticminstrel> | I definitely recall the existence of functions that do something like that... |
| 03:00 | <@Vornicus> | So let's see that bit of code - the code that loads single bytes and their corresponding codes into the dictionary (which, by the way, is merely a Tree) |
| 03:01 | < celticminstrel> | ...though it's possible it was a proposal rather than something actually in the Java library. |
| 03:01 | <@Vornicus> | it's not in there, no |
| 03:02 | < Orthia> | http://pastebin.starforge.co.uk/179 |
| 03:06 | <@Vornicus> | Don't need to cast i alone there. |
| 03:06 | <@Vornicus> | also, this method is static. |
| 03:08 | <@Vornicus> | also: raw isn't a member variable but an argument variable, so remove yes; and encoded is a local variable so move it into the start of the method. |
| 03:08 | < Orthia> | 'so remove yes'? |
| 03:08 | <@Vornicus> | remove "raw" where you ask //remove? |
| 03:13 | < Orthia> | Ah! Right. |
| 03:13 | < Orthia> | I was reading too literally, and going "But where is yes to remove?" >_> |
| 03:13 | < Orthia> | Sorry, my brain is doing this massively today. |
| 03:13 | <@Vornicus> | and finally: starting i at 1 here is wrong. Start it at 0, use i + 1 in the add call to get the value. The bug there is that -128 never gets a code the way it works |
| 03:15 | < Orthia> | ...ok, thank you. I'd have never bloody figured that one for an age. |
| 03:16 | < Orthia> | OK. |
| 03:17 | < Orthia> | Now I need a method for a loop that takes a chunk of the data, adds it to the tree, and then records where it got added. Yes? |
| 03:17 | <@Vornicus> | Something like that, yes. |
| 03:19 | | Serah [Z@3A600C.A966FF.5BF32D.8E7ABA] has joined #code |
| 03:21 | < Orthia> | hm. Cannot use raw.read() as the check whether or not the loop should continue - reading actually takes a lump out of the file, so... hm. |
| 03:21 | < Orthia> | Maybe a do-while. |
| 03:21 | <@Vornicus> | we're going to do this using a two-level loop; this keeps us in bytes, for one thing, and reduces the number of reads we do, and keeps us sane. |
| 03:22 | <@Vornicus> | In certain other ways. |
| 03:22 | <@Vornicus> | what we're going to do is we're going to have an outer loop that reads 1 kilobyte of stuff from the file, then an inner loop that loops over the stuff in our array. |
| 03:23 | < Orthia> | OK |
| 03:24 | <@Vornicus> | so we need two locals for this: int bytes_read, and byte[] buffer. |
| 03:24 | <@Vornicus> | http://java.sun.com/javase/6/docs/api/java/io/InputStream.html#read(byte[]) <--- this is the method we'll use. |
| 03:24 | < Orthia> | Hm |
| 03:25 | < Orthia> | Oh, I think I see |
| 03:25 | < Orthia> | We take things out of read() and put them into buffer, so the LZW can keep encoding? |
| 03:26 | <@Vornicus> | Sort of, yes. |
| 03:26 | < Orthia> | OK? |
| 03:26 | <@Vornicus> | read(byte[]) fills our buffer, and returns the number of things it filled the buffer with. |
| 03:27 | <@Vornicus> | So we'll do this, process the stuff we filled the buffer with, and then read some more. |
| 03:28 | < Orthia> | Hm |
| 03:28 | < Orthia> | Is there a specific size array we want? |
| 03:28 | <@Vornicus> | Not /particularly/, but I'd use 1024. |
| 03:29 | < Orthia> | Would we want to use an ArrayList instead, so it's of variable size? |
| 03:30 | <@Vornicus> | No, an actual array. |
| 03:30 | < Orthia> | OK |
| 03:34 | <@Vornicus> | (because it will be of fixed size, we'll just be putting a possibly-variable amount of data in) |
| 03:34 | < Orthia> | http://pastebin.starforge.co.uk/180 |
| 03:35 | < Orthia> | read(byte[], b) is apparently expecting an ArguementList |
| 03:36 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has quit [[NS] Quit: Z?] |
| 03:36 | <@Vornicus> | What the crap dude |
| 03:37 | < Orthia> | What have I done now >_> |
| 03:38 | < celticminstrel> | Well yeah, byte[] is a type. You can't pass it to a function. |
| 03:38 | < Orthia> | Oh. Er. Oops. |
| 03:39 | < celticminstrel> | I'd suggest new byte[size], but I don't really know what you're doing there. |
| 03:39 | <@Vornicus> | pass buffer to the function; get bytes_read from it. Compare bytes_read to -1 for your outer loop's condition. |
| 03:40 | <@Vornicus> | then iterate over buffer (using bytes_read as the loop limit) for your inner loop. |
| 03:44 | < Orthia> | ...Hm. Wait, I get it - Bytes_read is a counter, isn't it? |
| 03:45 | <@Vornicus> | Sort of - it tells us how many bytes we got in our latest read() attempt. |
| 03:51 | <@Vornicus> | If you have anything you should show me. |
| 03:52 | < Orthia> | hrn, sec, still nutting it out |
| 03:54 | < Orthia> | http://pastebin.starforge.co.uk/181 |
| 03:54 | < Orthia> | I know it is wrong, but here's what I'm poking at. |
| 03:55 | <@Vornicus> | Yeah, not quite |
| 03:55 | <@Vornicus> | Notice the method that I linked you to - it takes a parameter b which is an array of bytes. |
| 03:56 | < Orthia> | hrm. OK. If it reads from that, then ... hn, OK |
| 03:56 | <@Vornicus> | What this method /does/ is it fills the array with stuff. Remember that arrays are thrown around by reference: by passing that array in, it modifies the array so the outside world can see it. |
| 03:56 | < Orthia> | ... Ah-hah |
| 03:57 | < Orthia> | I had forgotten that aspect of arrays, it was "But why am I throwing the bit I want /out/ into the function?" |
| 03:58 | <@Vornicus> | then bytes_read tells you how full that array is. |
| 03:59 | < Orthia> | So bytes_read is essentially the same as byte[].Length(); in another world? |
| 03:59 | < Orthia> | (Yes I know it doesn't work that way) |
| 04:00 | <@Vornicus> | vaguely, yes |
| 04:00 | <@Vornicus> | It's essentially how much of the array is /valid/ - going beyond that point gets you into garbage land. |
| 04:01 | < Orthia> | hn |
| 04:03 | <@Vornicus> | (use it like you would use buffer.length if we weren't sometimes partially filling it.) |
| 04:05 | < Orthia> | So |
| 04:05 | < Orthia> | buffer[??] = raw.read(buffer); |
| 04:05 | <@Vornicus> | bytes_read = raw.read(buffer) |
| 04:06 | < Orthia> | ...ah-/hah/ |
| 04:06 | < Orthia> | Because .read will output the number of bits it chomped, and it will try to fill the array to full if at all possible - correct? |
| 04:06 | | * Orthia was just figuring it out. |
| 04:06 | < celticminstrel> | And buffer is passed by reference, of course... |
| 04:07 | < Orthia> | Yeah. I kept forgetting that bit. |
| 04:07 | <@Vornicus> | Precisely |
| 04:07 | < Orthia> | So, ok |
| 04:07 | < Orthia> | That fills our buffer and tells me how much data is in the buffer. |
| 04:08 | <@Vornicus> | Yep. |
| 04:08 | < Orthia> | We then start looping... iterating through the buffer up to bufferSize, hitting trie.add on each bit? |
| 04:08 | | celticminstrel [celticminstre@Nightstar-f8b608eb.cable.rogers.com] has quit [[NS] Quit: *hums* Can't stay now!] |
| 04:08 | <@Vornicus> | Not on each bit, we'll deal with that after we've got our byte looper working. |
| 04:09 | < Orthia> | OK then, what does it need to do next? |
| 04:09 | <@Vornicus> | just say //process for your actual per-byte process for now. |
| 04:09 | <@Vornicus> | Show me your code. |
| 04:11 | < Orthia> | http://pastebin.starforge.co.uk/182 |
| 04:13 | | Serah [Z@3A600C.A966FF.5BF32D.8E7ABA] has quit [Ping timeout: 121 seconds] |
| 04:14 | < Orthia> | On the right track at all? |
| 04:14 | <@Vornicus> | Pretty much right. |
| 04:15 | <@Vornicus> | 30 should be != -1 instead of > 0; 26 you don't need to cast i alone. |
| 04:15 | < Orthia> | Awesome. |
| 04:16 | <@Vornicus> | index should be 257, not 256. |
| 04:18 | < Orthia> | OK |
| 04:18 | < Orthia> | Those two make me think I should be using index in the for loop to cut down on breakability, but we will leave that another day |
| 04:18 | < Orthia> | Also, er. For loop halts at i < 256. Index now starts at 257. Fencepost error? |
| 04:19 | <@Vornicus> | nope. |
| 04:19 | <@Vornicus> | remember that the indexes we put into the trie were i + 1 |
| 04:20 | <@Vornicus> | So we'll have indices from 1 to 256 in the trie, 0 reserved, and index waiting at 257. |
| 04:21 | < Orthia> | Oh. Funky. |
| 04:21 | <@Vornicus> | And we will /not/ be using index in the for loop directly - you'll see why in a bit. |
| 04:21 | < Orthia> | Yessir! |
| 04:21 | < Orthia> | Done all that. Want to see, or want to get to the next bit, which I'm pretty sure is the Elderitch Horror? |
| 04:23 | <@Vornicus> | The next bit is lines 10-16 in http://pastebin.starforge.co.uk/156; we'll have to fiddle with it a bit to get it into java, but it's mostly pretty straightforward. |
| 04:23 | < Orthia> | Curiously enough, I'm almost looking forward to it! |
| 04:25 | <@Vornicus> | Before we can actually /do/ this though we need to know where we currently are in the trie; I called this "current_codes", you'll probably want to call it "current_node" or something. It's a Tree, and it starts initialized to trie. |
| 04:27 | < Orthia> | ...OK |
| 04:27 | < Orthia> | Is this in addition to or in replacement of my 'trie'? |
| 04:28 | <@Vornicus> | In addition to. We need to keep track of the root of the trie, because we'll head back there every time we run off the end of the existing trie. |
| 04:28 | <@Vornicus> | (which happens a lot - it's the only time we /add/ to the trie, too) |
| 04:30 | < Orthia> | In this case, if we have two trees running around, do I want to rename trie? |
| 04:30 | <@Vornicus> | No. |
| 04:30 | <@Vornicus> | well, okay, trie_root |
| 04:31 | <@Vornicus> | maybe, if you want. |
| 04:32 | < Orthia> | Righto. |
| 04:32 | <@Vornicus> | Now that we have that set up we can tell the inner loop what to do. |
| 04:32 | < Orthia> | Tree currentNode = new Tree(); ? |
| 04:33 | < Orthia> | Also, is this inside the loop or a function variable? |
| 04:37 | <@Vornicus> | no, tree currentNode = trie_root |
| 04:37 | <@Vornicus> | and it's a function-wide variable. |
| 04:38 | < Orthia> | Yessir. |
| 04:39 | <@Vornicus> | we use current_node to search down the trie as we move through the buffer. Which brings us, finally, to what we're actually doing here. |
| 04:39 | < Orthia> | Aha! OK. |
| 04:44 | <@Vornicus> | We need to search in current_node for the current byte. |
| 04:46 | < Orthia> | .seek, right? |
| 04:47 | < Orthia> | No, that's not right. |
| 04:47 | <@Vornicus> | no, that is right. |
| 04:48 | < Orthia> | http://pastebin.starforge.co.uk/183 - Do I still want to be adding crap to rootTrie ? |
| 04:48 | < Orthia> | Or trieRoot or whatever I called it. |
| 04:48 | <@Vornicus> | trieRoot? no, that stuff is Wrong |
| 04:50 | < Orthia> | Delete entirely, start fresh? |
| 04:50 | <@Vornicus> | Those two inner lines should be deleted, yes. |
| 04:51 | < Orthia> | Replaced with: currentNode.seek(buffer[i]); |
| 04:51 | <@Vornicus> | All right, we need to assign the result of that to something; dunno what to call it, but it's a TreeNode, and we're going to extract some things from it. |
| 04:51 | < Orthia> | Hm |
| 04:52 | < Orthia> | So wait, currentNode should probably be currentTree |
| 04:52 | < Orthia> | Then we can have currentNode within the tree? |
| 04:52 | <@Vornicus> | ....sure, works for me. |
| 04:53 | < Orthia> | And currentNode is a TreeNode we have called from within the currentTree, correct? |
| 04:53 | < Orthia> | - Are we /able/ to call individual Nodes? |
| 04:53 | | * Orthia checks. |
| 04:54 | < Orthia> | Oh, hey. |
| 04:54 | < Orthia> | seek returns the TreeNode anyway. Handy. |
| 04:55 | < Orthia> | currentNode = currentTree.seek(buffer[i]); |
| 04:55 | <@Vornicus> | We're able to talk about individual nodes; seek does return it. |
| 04:55 | < Orthia> | (God but I love Eclipse and 'refactor' |
| 04:57 | <@Vornicus> | okay. Now, we need to check whether curentNode is null or not. I'd start with if (currentNode != null), it's the more common case. |
| 04:58 | < Orthia> | Righto |
| 04:59 | < Orthia> | And then inside, if it is not Null, we want to start doing Stupid Pet Tricks. |
| 05:00 | <@Vornicus> | these tricks aren't that stupid |
| 05:00 | < Orthia> | righto |
| 05:00 | <@Vornicus> | We need to set current_tree to current_node.getChild() or whatever it's called |
| 05:01 | <@Vornicus> | and we need an int current_index (which is to say, the index that this would encode to, not the index that we'd next add to the tree) to store current_node.getIndex or whatever it's called |
| 05:01 | < Orthia> | ...oh. Oh, that's clever. |
| 05:02 | <@Vornicus> | This way when we get to the /next/ character, if that one takes us off the end of the tree, we can encode the previous sequence of characters. |
| 05:04 | <@Vornicus> | once you've put that in, show me your code. |
| 05:08 | < Orthia> | http://pastebin.starforge.co.uk/184 |
| 05:08 | < Orthia> | That took like four minutes to bloody upload. |
| 05:08 | <@Vornicus> | Yeesh |
| 05:09 | <@Vornicus> | anyway that's not wrong unless those are the wrong words, that's perfectly sensible. |
| 05:09 | <@Vornicus> | And that's /all we do/ if we're still living in the trie. Time fore... |
| 05:09 | <@Vornicus> | ELSE |
| 05:09 | <@Vornicus> | AND THE ALTERNATIVE CONDITIONS |
| 05:10 | <@Vornicus> | ...is possibly the shittiest band name ever. |
| 05:10 | < Orthia> | My one concern is thus: We are returning the index of the current node, and the /child/ of that node, not the childs index. |
| 05:10 | <@Vornicus> | Right. the child is a tree, it doesn't have an index. |
| 05:11 | <@Vornicus> | And we want the index of the current node: this way, the next time, if we have run off the end of the tree, everything /before/ we ran off the end gets encoded. |
| 05:12 | <@Vornicus> | Did that make sense? |
| 05:12 | < Orthia> | Aha, it does. I was just checking it was doing what we thought it was doing. |
| 05:13 | <@Vornicus> | OKay, anyway, on to our else. |
| 05:15 | < Orthia> | Right! |
| 05:15 | <@Vornicus> | If we get the else, then current_node is null; this means that we've run off the end of the tree! |
| 05:16 | <@Vornicus> | Which means we are ready to encode. |
| 05:18 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has joined #code |
| 05:20 | <@Vornicus> | so, 1. we add current_index (which is to say, the code index for the /previous sequence of bytes/) to encoded; 2. we add the current byte and index to current_tree; 3. we increment index (because we've used this index to make a new code; 4. start again from the trie_root and follow the path for the current byte. |
| 05:21 | <@Vornicus> | 1 2 and 3 are 1-liners; 4 is the three lines we used to get and apply current_node, except we use trie_root instead of current_tree |
| 05:24 | < Orthia> | hrn |
| 05:25 | <@Vornicus> | 2 is the actual trick to LZW - we're adding the previous sequence /plus/ this latest character to the dictionary of codes we know. |
| 05:26 | < Orthia> | OK, query: In the else, currentIndex is not set to anything previously. Yes? |
| 05:26 | <@Vornicus> | current_index is set to something /by the previous run through the loop/ |
| 05:26 | < Orthia> | Aaaah. |
| 05:26 | <@Vornicus> | as is current_tree |
| 05:26 | < Orthia> | It is erroring out of fear that it may not have been initalised. |
| 05:27 | < Orthia> | Oh, duh. I never set it to /anything/. |
| 05:27 | <@Vornicus> | Initialize it at the top of the function. 0, I guess. |
| 05:27 | < Orthia> | 0 is our stop code~ |
| 05:27 | <@Vornicus> | YOu're right! |
| 05:28 | <@Vornicus> | but here's the thing: the first byte will /always/ go through the if. |
| 05:28 | < Orthia> | Or is this a matter of 'we are confident we will have run through once previously', and if we /haven't/, it's time to stop anyway? |
| 05:28 | <@Vornicus> | It is indeed. |
| 05:30 | < Orthia> | 2: We're adding in the current buffer byte, and the currentIndex? |
| 05:30 | <@Vornicus> | and index, not current_index |
| 05:30 | <@Vornicus> | index tells us the next thing we add to the trie; current_index tells us the next thing we'll potentially add to encoded. |
| 05:31 | < Orthia> | Oh, yes. |
| 05:31 | <@Vornicus> | These might not be the best names. |
| 05:31 | < Orthia> | We will fix them in a second then~ |
| 05:32 | < Orthia> | Up to 4. |
| 05:32 | <@Vornicus> | 4, well, here, let me show you the difference in python. |
| 05:32 | < Orthia> | Hrn. What 4 lines? |
| 05:33 | <@Vornicus> | current_index, current_codes = current_codes[c] |
| 05:33 | <@Vornicus> | That's in the if. |
| 05:33 | <@Vornicus> | current_index, current_codes = codes_root[c] |
| 05:33 | <@Vornicus> | and that's in the else. |
| 05:34 | < Orthia> | Oh, right |
| 05:34 | < Orthia> | We want to start iterating from the root again instead of continuing down the branches, yes? |
| 05:34 | <@Vornicus> | yep. |
| 05:34 | <@Vornicus> | because we ran off the end of the tree. |
| 05:36 | <@Vornicus> | this comes out to three lines in Java; note that the three lines you already have are broken up by an if. |
| 05:37 | < Orthia> | Hrn. How do I reset currentNode? |
| 05:37 | <@Vornicus> | How do you set current_node originally? |
| 05:38 | < Orthia> | gottit |
| 05:38 | <@Vornicus> | Show me the code. |
| 05:39 | < Orthia> | http://pastebin.starforge.co.uk/185 |
| 05:39 | <@Vornicus> | If it's right, we're three lines from victory. |
| 05:39 | < Orthia> | Is that correct? |
| 05:39 | < Orthia> | It looks like you could factorise a couple chunks of it, but we'll leave it for the moment |
| 05:39 | <@Vornicus> | Not quite |
| 05:39 | < Orthia> | OK, where'd I go wrong |
| 05:39 | < Orthia> | (And after this we will have to finish I think ,but thank you thank you thank you, sir) |
| 05:39 | <@Vornicus> | delete 43 - we won't set currentTree yet |
| 05:40 | <@Vornicus> | then in what's currently called 44 use trieRoot instead of currentTree |
| 05:40 | <@Vornicus> | and then after that set currentTree to currentNode.getChild(); |
| 05:42 | < Orthia> | right |
| 05:42 | <@Vornicus> | All right. |
| 05:42 | <@Vornicus> | And that is everything we do in the loop. |
| 05:42 | <@Vornicus> | Now there's two things left to do. |
| 05:42 | <@Vornicus> | 1. we need to encode the last few bits. |
| 05:43 | <@Vornicus> | 2. we need to send the stop code. |
| 05:43 | < Orthia> | And we need to actually add these wonderful indexes to the string. |
| 05:43 | <@Vornicus> | well. 3. we need to return the encoded. |
| 05:43 | <@Vornicus> | packing is another matter. We won't do that in this function. |
| 05:44 | < Orthia> | No we won't. |
| 05:44 | <@Vornicus> | so first off: when we hit the end of the loop, we need to add the index for the last few bytes to our list of indices. |
| 05:44 | < Orthia> | Right |
| 05:45 | <@Vornicus> | This should be trivial. |
| 05:45 | < Orthia> | else {...} encoded.add(currentIndex); |
| 05:45 | < Orthia> | ? |
| 05:46 | <@Vornicus> | not outside the else. |
| 05:46 | <@Vornicus> | outside /the entire loop/ |
| 05:46 | <@Vornicus> | indeed, even outside the /outer/ loop. |
| 05:46 | <@Vornicus> | This happens when there is absolutely nothing left at all to look at. |
| 05:47 | < Orthia> | So we only add things to the encoded string when we are about to return it? |
| 05:47 | < Orthia> | 'cuz once you're outta the loop, yer done. |
| 05:48 | <@Vornicus> | Well - we add things to the encoded string when we run off the tree, and when we run out stuff entirely, which is when we're about to return it. |
| 05:48 | < Orthia> | Oh I see, the else does it too. Okay. Carry on. |
| 05:48 | < Orthia> | Next step? |
| 05:49 | | * Orthia hates to be impatient, but is getting poked to cook dinner, so will be ending this after this, he suspects~ |
| 05:49 | <@Vornicus> | Next step. |
| 05:49 | <@Vornicus> | Stop code! |
| 05:51 | < Orthia> | if(currentIndex = 0) rootTree = new Tree(); ? |
| 05:51 | <@Vornicus> | ...what |
| 05:51 | < Orthia> | OK, I think we might want to leave this till later then, 'cuz my brain is fried. |
| 05:51 | <@Vornicus> | This is the /stop code/ |
| 05:51 | < Orthia> | Also I am being waved at by my guests~ |
| 05:51 | <@Vornicus> | all we're doing is throwing it on to the encoding. |
| 05:51 | < Orthia> | Oh, right. |
| 05:52 | <@Vornicus> | The clear code requires some rather interesting reengineering, and we're not doing it now. |
| 05:52 | < Orthia> | OK |
| 05:52 | < Orthia> | encoded.add(0); |
| 05:52 | <@Vornicus> | Bingo. |
| 05:52 | <@Vornicus> | And then we return, and VICTORY |
| 05:53 | < Orthia> | It's done? |
| 05:53 | < Orthia> | Sweet. |
| 05:53 | < Orthia> | Thank you thank you thank you, mister Vorny |
| 05:53 | < Orthia> | I will now leave you to sleep, and myself go flomp into lounge to get food, for I am meant to be getting food~ |
| 05:54 | <@Vornicus> | laters |
| 05:54 | < Orthia> | http://pastebin.starforge.co.uk/186 - ze nights work. Thank you! |
| 05:56 | <@Vornicus> | next up is decode, then there's pack and unpack, and then we get to suss out clear codes. |
| 05:57 | <@Vornicus> | decode is relatively straightforward but there's a fiddly exception; pack and unpack I haven't even tried to do yet. |
| 06:00 | <@Vornicus> | clear codes, there is a problem that gets kind of glossed over in the assignment. |
| 06:02 | <@Vornicus> | well, two problems. 1. clearing at any loss in efficiency at all means that any time we end up dropping a first- or second- level index it'd clear, so better is to wait for the efficiency to drop 5%. 2. the unpacker doesn't have enough information to know when you clear unless you pass a clear code itself. |
| 06:04 | | AnnoDomini [annodomini@Nightstar-688fac12.adsl.tpnet.pl] has joined #code |
| 06:04 | | mode/#code [+o AnnoDomini] by Reiver |
| 06:10 | <@Vornicus> | (or some other amount. I picked 5% because it seems a convenient amount.) |
| 06:47 | | Vornicus is now known as Vornicus-Latens |
| 07:10 | | You're now known as TheWatcher |
| 08:13 | | You're now known as TheWatcher[afk] |
| 09:00 | | Rhamphoryncus [rhamph@Nightstar-8931f88f.abhsia.telus.net] has quit [Client exited] |
| 09:16 | | Orthia [orthianz@Nightstar-04ecdbee.xnet.co.nz] has quit [Client closed the connection] |
| 09:32 | | Orthia [orthianz@Nightstar-04ecdbee.xnet.co.nz] has joined #code |
| 09:38 | | Thaqui [Thaqui@27B34E.D54D49.F53FA1.6A113C] has quit [Client closed the connection] |
| 10:10 | | You're now known as TheWatcher |
| 10:21 | | Beaver [guest@583787.B5B398.0F5279.CF6C9A] has joined #code |
| 10:21 | | Beaver [guest@583787.B5B398.0F5279.CF6C9A] has quit [G-Lined: Spamming/Advertising] |
| 10:25 | | Attilla [Attilla@FBC920.642D35.7B2B85.11BB86] has joined #code |
| 10:25 | | mode/#code [+o Attilla] by Reiver |
| 10:26 | | AnnoDomini [annodomini@Nightstar-688fac12.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 10:28 | | AnnoDomini [annodomini@Nightstar-af2cc94e.adsl.tpnet.pl] has joined #code |
| 10:28 | | mode/#code [+o AnnoDomini] by Reiver |
| 11:13 | | Orth [orthianz@Nightstar-ca0ee0f8.xnet.co.nz] has joined #code |
| 11:14 | | REGIS_mark_V [WhyNot@NetworkAdministrator.Nightstar.Net] has quit [Ping timeout: 121 seconds] |
| 11:14 | | REGIS_mark_V [WhyNot@NetworkAdministrator.Nightstar.Net] has joined #code |
| 11:15 | | Orthia [orthianz@Nightstar-04ecdbee.xnet.co.nz] has quit [Ping timeout: 121 seconds] |
| 11:30 | | celticminstrel [celticminstre@Nightstar-f8b608eb.cable.rogers.com] has joined #code |
| 11:30 | | celticminstrel [celticminstre@Nightstar-f8b608eb.cable.rogers.com] has quit [Client exited] |
| 12:48 | | Serah [Z@3A600C.A966FF.5BF32D.8E7ABA] has joined #code |
| 13:34 | | REGIS_mark_V is now known as PinkFreud |
| 13:48 | | You're now known as TheWatcher[d00m] |
| 13:54 | | Orthia [orthianz@Nightstar-832e0407.xnet.co.nz] has joined #code |
| 13:56 | | Orth [orthianz@Nightstar-ca0ee0f8.xnet.co.nz] has quit [Ping timeout: 121 seconds] |
| 14:07 | | celticminstrel [celticminstre@1AB00B.855209.A256BB.B16D09] has joined #code |
| 14:28 | | Serah [Z@3A600C.A966FF.5BF32D.8E7ABA] has quit [Ping timeout: 121 seconds] |
| 15:06 | | You're now known as TheWatcher[afk] |
| 15:09 | | Serah [Z@26ECB6.A4B64C.298B52.D80DA0] has joined #code |
| 15:51 | | celticminstrel [celticminstre@1AB00B.855209.A256BB.B16D09] has quit [Client exited] |
| 15:54 | | * SmithKurosaki hugs |
| 15:55 | | * Vornicus-Latens hugs Smith |
| 15:55 | < SmithKurosaki> | lol |
| 15:55 | < SmithKurosaki> | So /ame sends it to all chans you are in, interesting |
| 15:56 | < gnolam> | You just typoed it, then, or what? :) |
| 15:57 | < SmithKurosaki> | Yea |
| 15:57 | < SmithKurosaki> | I wanted to type /me, but typed /ame and now its in all my chans in all my networks |
| 16:48 | | Syloqs-AFH [Syloq@NetworkAdministrator.Nightstar.Net] has quit [Connection reset by peer] |
| 16:52 | | Syloqs_AFH [Syloq@NetworkAdministrator.Nightstar.Net] has joined #code |
| 16:54 | | Syloqs_AFH is now known as Syloqs-AFH |
| 17:10 | | Attilla [Attilla@FBC920.642D35.7B2B85.11BB86] has quit [Ping timeout: 121 seconds] |
| 17:13 | | Attilla [Attilla@FBC920.DDABA2.4E6AE9.B3BF85] has joined #code |
| 17:13 | | mode/#code [+o Attilla] by Reiver |
| 17:21 | | Rhamphoryncus [rhamph@Nightstar-8931f88f.abhsia.telus.net] has joined #code |
| 17:47 | | celticminstrel [celticminstre@Nightstar-f8b608eb.cable.rogers.com] has joined #code |
| 17:58 | | You're now known as TheWatcher |
| 18:07 | | Serah [Z@26ECB6.A4B64C.298B52.D80DA0] has quit [Ping timeout: 121 seconds] |
| 18:27 | | Vornicus-Latens is now known as Vornicus |
| 18:31 | | Serah [Z@26ECB6.A4B64C.298B52.D80DA0] has joined #code |
| 21:45 | | McMartin [mcmartin@Nightstar-6397d65b.pltn13.sbcglobal.net] has quit [Client closed the connection] |
| 21:50 | < Rhamphoryncus> | heh awesome. I just found some code that would be simplified by swapping the class on an instance, as a way of changing modes |
| 21:51 | < Rhamphoryncus> | An article on prototyping (as an alternative to classes) was promoting that approach. Of course they claimed it was a special ability of prototyping |
| 21:53 | | McMartin [mcmartin@Nightstar-3c130f1a.pltn13.sbcglobal.net] has joined #code |
| 21:53 | | mode/#code [+o McMartin] by Reiver |
| 22:11 | <@ToxicFrog> | "Swapping the class on an instance"? |
| 22:29 | < Rhamphoryncus> | class Foo(object): pass; class Bar(object): pass; x = Foo(); x.__class__ = Bar |
| 22:29 | < Rhamphoryncus> | replacing the type pointer |
| 22:32 | <@McMartin> | Morphing is bad, bad. |
| 22:32 | | * McMartin whaps Rhamphoryncus with a rolled-up newspaper |
| 22:32 | < Rhamphoryncus> | oh I agree |
| 22:32 | < Rhamphoryncus> | That's why I'm so entertained |
| 22:33 | < Rhamphoryncus> | These aren't public classes, so there's no chance of any further subclassing. This is the only modal state involved. It would work perfectly |
| 22:35 | < Rhamphoryncus> | http://pastebin.starforge.co.uk/188 |
| 22:40 | < Rhamphoryncus> | http://pastebin.starforge.co.uk/189 |
| 22:43 | < Rhamphoryncus> | oops forgot to remove the mode assignment in __init__ |
| 22:45 | < Rhamphoryncus> | and come to think of it, temp_watches is supposed to be a common attribute.. not that I recall what it's for |
| 22:46 | < Rhamphoryncus> | The whole morphing thing wouldn't be so bad if you had multiple delegates you swap, rather than just a single __class__ attribute to swap |
| 22:47 | <@McMartin> | If you have a reassignable delegate object, that isn't morphing |
| 22:48 | < Rhamphoryncus> | If the delegate object is empty and the compiler optimizes it down to just a type pointer... |
| 22:49 | < Rhamphoryncus> | Actually, not empty, but rather a union |
| 22:49 | <@McMartin> | If it's empty, it's not a delegate~ |
| 22:49 | < Rhamphoryncus> | An inlined delegate |
| 22:50 | <@McMartin> | Any member in C++ is "inlined". |
| 22:50 | <@McMartin> | The whole point is that you're looking up the type pointer and getting the method to call. |
| 22:50 | <@McMartin> | If this is *purely* for RTTI, you're doing something *else* terribly wrong. |
| 22:50 | < Rhamphoryncus> | No, the point is that the compiler is doing it for me :) Hypothetically |
| 22:51 | < Rhamphoryncus> | There's no language that does it in such a fashion |
| 22:52 | <@McMartin> | Violating type safety in C++ will let you forcibly reassign the vtable pointer |
| 22:56 | < Rhamphoryncus> | http://pastebin.starforge.co.uk/190 |
| 22:57 | | AnnoDomini [annodomini@Nightstar-af2cc94e.adsl.tpnet.pl] has quit [[NS] Quit: Sleep.] |
| 22:58 | < Rhamphoryncus> | The advantage is you can group the methods for a mode togoether, rather than having a tree of if-statements in every method |
| 22:59 | < Rhamphoryncus> | Disadvantages include not being extensible, but I don't think the tree of if-statements ie extensible either |
| 23:00 | < Rhamphoryncus> | it is a niche tool, albeit perfect for this job. Does it generalize to other finite state machines? |
| 23:06 | < Rhamphoryncus> | eww hey, there's a semicolon in my soup! |
| 23:20 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has quit [[NS] Quit: Z?] |
| 23:39 | | You're now known as TheWatcher[T-2] |
| 23:42 | | You're now known as TheWatcher[zZzZ] |
| --- Log closed Wed Mar 24 00:00:43 2010 |