search of the:   for: 
code logs -> 2008 -> Mon, 10 Mar 2008< code.20080309.log - code.20080311.log >
--- Log opened Mon Mar 10 00:00:14 2008
00:25
<@AnnoDomini>
Finally finished that thing. Now only to complete my rounds around the tubes and I can go to sleep... and start work on the other things when I wake up. :/
00:48 ReivUni [~82d94c4d@Nightstar-29731.dsl.in-addr.zen.co.uk] has joined #Code
00:49 You're now known as TheWatcher[T-2]
00:53 You're now known as TheWatcher[zZzZ]
00:56 AnnoDomini [AnnoDomini@83.21.28.ns-26444] has quit [Quit: ACHTUNG! ALLES LOOKENSPEEPERS! Das Internet is nicht fuer gefingerclicken und giffengrabben. Ist easy droppenpacket der routers und overloaden der backbone mit der spammen und der me-tooen. Ist nicht fuer gewerken bei das dumpkopfen. Das mausklicken sichtseeren keepen das bandwit-spewin hans in das pockets muss; relaxen und watchen das cursorblinken.]
01:52 gnolam [lenin@85.8.5.ns-20483] has quit [Quit: Z?]
02:33
< ReivUni>
...Okay, there's /got/ to be a class that handles parsing commandline inputs in Java somewhere, right?
02:33
< ReivUni>
And/or I'm overthinking this.
02:35
< ReivUni>
Input/Output: Your program should run from the command-line (i.e. no GUI) with the command xsort and accept the following command-line arguments: -o outputfile -- where outputfile is the name of the file into which the final, sorted data is to be stored. If this argument is omitted then results are displayed to standard output. -t tempdir -- where tempdir is the name of the directory where temporary files are created. If this ar
02:36
< ReivUni>
This is pretty basic stuff; and we haven't been instructed to do it by hand (We have in other parts which are the bit they're actually assessing us on) - so why am I failing utterly to figure out a straightforward method of doing it? >.>
02:40 Thaqui [~Thaqui@Nightstar-123.jetstream.xtra.co.nz] has joined #code
02:40 mode/#code [+o Thaqui] by ChanServ
02:51
< ReivUni>
...ooookay, random syntax errors. Lovely. ??
02:52 * ReivUni throttles it, goes to class. Will have more goes later tonight.
02:53 ReivUni [~82d94c4d@Nightstar-29731.dsl.in-addr.zen.co.uk] has quit [Quit: CGI:IRC (EOF)]
04:51 GeekSoldier|bed [~Rob@Nightstar-9089.dip.t-dialin.net] has quit [Ping Timeout]
05:38
<@Reiver>
Okay, I'm doing something wrong with it.
05:39
<@Reiver>
Should I be trying to hold onto the /last/ (i-1) string, so that I know what to do with /this/ one?
05:42
<@ToxicFrog>
Generally, you look at string i to determine what to do with string i+1
05:42
< Vornicus>
What is it we're doing anyway?
05:42
<@ToxicFrog>
One would think this would be something like Map args = Getopts.parse(argv), but java doesn't seem to have getopts
05:47
<@Reiver>
hrmn
05:48
<@Reiver>
Issue is I need to know what to do with commands that don't have an -x in front of them
05:49
<@Reiver>
Hence my guess that if I read an -x, I then keep that bit of information so I know what to do with the string that comes after it
05:49
<@Reiver>
And if I /haven't/ read an -x, then I take it as a raw command.
05:50
<@ToxicFrog>
The way I do it is generally something like:
05:50
<@ToxicFrog>
for arg in next_arg do -- where next_arg is an iterator returning the next argument from the list
05:50
<@ToxicFrog>
 if arg == "-x" then
05:51
<@ToxicFrog>
 handle_flag_x(next_arg()) -- if it's "-x foo"
05:51
<@ToxicFrog>
 handle_flag_x() -- if it's meant to be just "-x"
05:51
<@ToxicFrog>
 ... code for other flags goes here ...
05:51
<@ToxicFrog>
 else
05:51
<@ToxicFrog>
 handle_raw_argument(arg)
05:51
<@ToxicFrog>
 end
05:51
<@ToxicFrog>
end
05:51
<@Reiver>
hm
05:52
<@ToxicFrog>
It'll look somewhat different in Java, but the principle is the same; if you get a flag that needs an additional argument, you grab the next argument (in a manner that ensures it'll be skipped by the main argument parsing loop) and handle the combination of flag+argument
05:52
<@ToxicFrog>
If you get a plain argument, or a flag that doesn't need one, you just handle it directly.
05:53
<@ToxicFrog>
This is easier since you don't need to handle constructs like "-xzy" => "-x -y -z" or "-ffoo" => "-f foo"
05:54
<@ToxicFrog>
And now I really do sleep
05:54 * Reiver eyes that
06:00
<@jerith>
ToxicFrog: Java does have getopt.
06:00
<@jerith>
It *really* needs something like optparse, though.
06:01
<@Reiver>
getopt?
06:02
<@jerith>
The system originally written to parse options in C.
06:03
<@jerith>
Been ported everywhere.
06:03
<@Reiver>
I see
06:03
<@Reiver>
So if I just use that, I can save having to hand-code the blasted thing?~
06:03
<@jerith>
Ruby and Python have optparse, which is /far/ superior.
06:03
<@jerith>
Not entirely.
06:04
<@jerith>
But you get a /lot/ of stuff done for you.
06:04
<@jerith>
Things like "-h" text generation, handling of optional and mandatory parameters, etc.
06:05
<@jerith>
I've got some code that uses it at work, so I can sanitise and send you an example it you want.
06:06
<@Reiver>
...Is it simple enough to use? I have a total of three optional arguements, here.
06:06
<@jerith>
Yes.
06:06
<@jerith>
In fact, the docs should give you all you need.
06:06
<@jerith>
Do you have my IM details?
06:07
<@Reiver>
aye, IIRC
06:07
<@Reiver>
For gtalk, yes?
06:07
<@jerith>
Cool. I'm off to work now and I prefer to to IRC from there.
06:07
<@jerith>
firxen@, yes.
06:07
<@Reiver>
Okay. Bye!
06:08
<@jerith>
Well, not *right* now. Still have to pack my bag. :-)
06:08
<@jerith>
This /is/ Java, right?
06:09
<@Reiver>
Java, yes.
06:09
<@jerith>
Ah. Then getopt, horrible as it is, is the lesser of the available evils.
06:10
<@jerith>
I wonder why Java doesn't hae an optparse implementation...
06:10
<@jerith>
+v
06:10
<@jerith>
Perhaps I'll write one.
06:12
<@jerith>
*Now* I'm off to work. Cheers all.
07:01 Vornicus is now known as Vornicus-Latens
08:00 AnnoDomini [AnnoDomini@83.21.28.ns-26444] has joined #Code
08:00 mode/#code [+o AnnoDomini] by ChanServ
08:01 GeekSoldier|bed [~Rob@Nightstar-8928.dip.t-dialin.net] has joined #code
08:05 GeekSoldier|bed is now known as GeekSoldier
08:43 You're now known as TheWatcher
09:54 * TheWatcher eyes XML::Simple, curses its lack of sort function option
10:29 Thaqui [~Thaqui@Nightstar-123.jetstream.xtra.co.nz] has left #code [Leaving]
11:02 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code
11:02 mode/#code [+o gnolam] by ChanServ
11:30 AnnoDomini [AnnoDomini@83.21.28.ns-26444] has quit [Ping Timeout]
11:32 AnnoDomini [AnnoDomini@83.21.41.ns-3845] has joined #Code
11:32 mode/#code [+o AnnoDomini] by ChanServ
12:10 GeekSoldier is now known as GeekSoldier|store
12:18 GeekSoldier|store is now known as GeekSoldier
13:22
<@ToxicFrog>
jerith: it does? Where
13:22
<@ToxicFrog>
?
15:07
<@McMartin>
I suspect he means "there exists a third-party port of it"
15:11 Serah [~Z@87.72.35.ns-26506] has quit [Quit: Don't try to read the quit message, that is impossible. Instead only realize the truth; "there is no quit message" and you will see it is not you who read the quit message but the quit message who reads you.]
15:12
<@ToxicFrog>
Oh, well, yes
15:12
<@ToxicFrog>
I'm assuming that since this is schoolwork you can't do that.
15:37
<@jerith>
Oh.
15:37
<@jerith>
That's silly.
15:37
<@jerith>
I wasn't aware that it was third-party -- I only hacked on the code someone else wrote.
15:37
<@McMartin>
RMS hates Sun with the fire of a thousand, well, suns.
15:38
<@jerith>
I can imagine.
15:38 * Vornicus-Latens still doesn't get why though
15:38
<@McMartin>
Vorn: Originally because they weren't releasing their (and thus third-party) code under GPL
15:39
<@McMartin>
Now ANAICT it's that they didn't do so fast enough.
15:40
<@McMartin>
They also occasionally show up as bit players in MS/SCO conspiracy theories about trying to slay all open source everywhere, but I can't really blame RMS for that since it's one of those lolinterwubs things.
15:43
<@ToxicFrog>
What does RMS's hatred of Sun have to do with getopts?
15:44
<@McMartin>
ISTR getopt being a GNU project, which tends to mirror his prejudices.
15:46
< Vornicus-Latens>
arg, cargo cult programming
15:46
<@ToxicFrog>
Well, "GNU Getopt" is.
15:47
<@ToxicFrog>
Getopt-the-specification is POSIX, though, and there's plenty of non-GNU implementations thereof in various languages.
15:47
<@ToxicFrog>
And considering that Java comes with libraries for every goddamn thing under the sun, I'm very surprised it doesn't have a java.util.Getopt
15:49
<@McMartin>
Aha.
15:49
<@McMartin>
Now that you mention it, it might be hiding somewhere in Runtime or System or something.
15:50 * McMartin is so used to writing basic option parsers in other langauges that his Java equivalents just do an array scan.
15:56
<@jerith>
import gnu.getopt.Getopt; <-- From my code.
15:56
<@jerith>
So there /is/ a GNU implementation for Java.
15:57
<@jerith>
Although I was under the somewhat vague impression that it was in the stdlib.
16:04
<@ToxicFrog>
None of the gnu.* stuff is part of standard Sun Java.
16:04
<@ToxicFrog>
It might come standard with the gcj toolchain.
16:07
<@jerith>
Yeah, My vague impression was mistaken in this case.
16:42 Brother_Willibald [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code
16:44 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has quit [Ping Timeout]
16:54 Brother_Willibald [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has quit [Ping Timeout]
17:01 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code
17:01 mode/#code [+o gnolam] by ChanServ
17:05 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has quit [Ping Timeout]
17:05 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code
17:05 mode/#code [+o gnolam] by ChanServ
17:11 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has quit [Ping Timeout]
17:14 You're now known as TheWatcher[afk]
17:17 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code
17:17 mode/#code [+o gnolam] by ChanServ
17:18
<@AnnoDomini>
Hm.
17:18
<@AnnoDomini>
Why is there a 1.3Gb hidden directory called System Volume Information on one of my drives?
17:19
<@AnnoDomini>
I'm mostly wondering why it's so large, mind.
17:19
< Vornicus-Latens>
SVI I believe holds bad blocks.
17:20
<@AnnoDomini>
Most of the other drives contain smaller versions of this directory. C: has just over 700Mb, but other than that they're only about 50-70Mb heavy.
17:43 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has quit [Ping Timeout]
17:47 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code
17:47 mode/#code [+o gnolam] by ChanServ
18:16 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has quit [Ping Timeout]
18:17 gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code
18:17 mode/#code [+o gnolam] by ChanServ
18:47 You're now known as TheWatcher
18:59
<@ToxicFrog>
It also contains information on system restore.
19:06 Vornicus-Latens is now known as Vornicus
19:07 Vornicus is now known as NSGuest-5494
19:07 NSGuest-5494 is now known as Vornicus
19:37 Pi [~sysop@ServicesAdmin.Nightstar.Net] has quit [Ping Timeout]
19:44
<@McMartin>
...ahahahahahaaaaaa
19:44
<@McMartin>
Knuth++
19:44
<@McMartin>
On his list of paper publications on his official CV
19:45
<@McMartin>
The first entry is a silly article he sent to MAD Magazine at age 19
19:45
<@McMartin>
The Potrzebie system of Weights and Measures
--- Log closed Mon Mar 10 19:52:47 2008
--- Log opened Mon Mar 10 20:04:18 2008
20:04 chris_ [~chris@Nightstar-29731.dsl.in-addr.zen.co.uk] has joined #code
20:04 Irssi: #code: Total of 16 nicks [8 ops, 0 halfops, 0 voices, 8 normal]
20:04 You're now known as TheWatcher
20:05 Irssi: Join to #code was synced in 53 secs
20:06 mode/#code [+o TheWatcher] by ChanServ
21:34 Pi-2 [~sysop@67.183.88.ns-4261] has joined #code
22:14 Vornicus [~vorn@ServicesOp.Nightstar.Net] has quit [Ping Timeout]
22:52 Molgorn [~moltare@Nightstar-29340.cable.ubr02.bath.blueyonder.co.uk] has joined #code
22:52
< Molgorn>
Good evening, gentlemen!
22:52
< EvilDarkLord>
And the rest of the meme. Hello, Molt.
22:54
< Molgorn>
My compiler is complaining about "implicit declaration of function `int malloc(...)'" in the tokens.h file, and I have no idea why. Any advice?
22:56
< EvilDarkLord>
#include <stdlib.h>
22:56
< Molgorn>
...I'm a muppet
22:57
<@ToxicFrog>
"implicit declaration of function foo" means you're using a function without first declaring it.
22:57
<@ToxicFrog>
Typically, this means you've forgotten an #include.
22:57 * Molgorn hangs his head in shame
23:00 You're now known as TheWatcher[T-2]
23:04 You're now known as TheWatcher[zZzZ]
23:04
< Molgorn>
yay it all compiles in the face
23:05
< Molgorn>
but doesn't link. never mind! hash() first.
23:05 Vornicus [~vorn@Admin.Nightstar.Net] has joined #code
23:05 mode/#code [+o Vornicus] by ChanServ
23:06
< Molgorn>
Does anyone have a link to a decent basic-making-a-C-hashing-function tutorial which'll let me hash a string into a suitable int?
23:06
< Molgorn>
I've looked but I keep finding .c files people tell one to include, which is Contraindicated
23:08
<@ToxicFrog>
The Dragon Book has a quick look, and Levitin - Algorithms has a more in depth overview
23:09
<@ToxicFrog>
In practice what you should probably look for is a tutorial on hash algorithms, which you can then implement in C yourself
23:09
< Molgorn>
Ooh, I've got the Levitin
23:10
<@ToxicFrog>
Page 265, if you have the same edition I do
23:10
< Molgorn>
Looking at it now
23:11
< Molgorn>
Not what I need, though
23:11
<@ToxicFrog>
How so?
23:11
<@ToxicFrog>
I don't think you're going to find "how to implement a hash specifically in C" that isn't just a simple hash library
23:11
< Molgorn>
I know I want to sum up the values of the characters of the strings and mod them against the hashtable size
23:12
<@ToxicFrog>
As a rule, this sort of thing is explained as an algorithm independent of any language
23:12
<@ToxicFrog>
And you are then responsible for writing the actual code that implements that algorithm.
23:12
< Molgorn>
What I don't know is how the blazes one takes a *char or whatever and turns it into a set of integer values for the chars in that *char
23:12
<@Vornicus>
That's not hard
23:12
<@Vornicus>
Seeing as a char is in reality a number
23:13
<@ToxicFrog>
And a char * is actually an array of characters
23:13
<@ToxicFrog>
(well, for this purpose)
23:13
<@ToxicFrog>
Recall that in C, strings are just arrays of characters, and arrays are just pointers to the first element.
23:13
< Molgorn>
So can you just iterate through from char*[0] to char*[length] and add up the chars you get?
23:13
<@Vornicus>
Sure, except
23:13
<@ToxicFrog>
replace char* with the name of the variable and yes.
23:13
<@ToxicFrog>
char * foo;
23:13
< Molgorn>
well, yes
23:13
<@Vornicus>
C strings are null terminated
23:14
<@ToxicFrog>
for (int i = 0; i < strlen(foo); ++i) { total += foo[i]; }
23:14
<@gnolam>
Assuming C99.
23:14
<@ToxicFrog>
This is not terribly efficient (it's O(n^2) due to repeated calls to strlen; a memoized version would be O(n)) but it gets the concept across.
23:15
< C_tiger>
Well, you can make a new variable len = strlen(foo)
23:15
< Molgorn>
Efficiency is the least of my concerns
23:15
<@ToxicFrog>
Indeed.
23:15
<@ToxicFrog>
So go with what makes the most sense, and worry about efficiency later, if at all.
23:16
<@ToxicFrog>
In this case, you can get the length of a string with strlen() (it's in <string.h>) and then you just iterate over it like an array and treat its elements like ints.
23:18
< C_tiger>
well, kinda... depends if you don't mind mapping 65 to a or if you want a = 1
23:18
<@ToxicFrog>
As far as the table itself goes, the easiest is probably closed hashing, with linear probing and a rehash once it gets too full
23:18
<@ToxicFrog>
But open hashing isn't too difficult either, once you're comfortable with pointers and structs.
23:19
<@ToxicFrog>
C_tiger: how do you handle, say, '!' then?
23:19
< C_tiger>
True, I'm only kinda paying attention to what mol's trying to do :P
23:20 * ToxicFrog throws C to the headcrabs~
23:21
<@ToxicFrog>
I also note that if implementing the hash becomes too much of a headache, you can always swap in a prefab hash and use that to test the rest of the problem
23:21
< C_tiger>
Sorry, just confused... it seemed like a hashtable was the goal but summing would mean "ha" hashes like "ah"
23:22
<@ToxicFrog>
And then revisit it later and implement your own.
23:23
<@ToxicFrog>
C_tiger: yes, which is why Levitin recommends using a more sophisticated hash.
23:23
<@ToxicFrog>
However, ?i str[i] is easy to implement.
23:23
<@ToxicFrog>
Which at the moment is more important than minimizing collisions or maximizing performance.
23:24
< C_tiger>
Ah.
23:25
< Molgorn>
C: This is merely an exercise. It doesn't have to be astoundingly efficient, just robust enough to not kill itself.
23:28
<@ToxicFrog>
(personally, the way I'd handle this is to use a prefab hash - drop that in, finish the rest of the program, test it for correctness with a hash that I know works - then write my own hash and replace the prefab one with that)
23:28
<@ToxicFrog>
Which reminds me - with the lexer done, how's the parser coming?
23:28 * Vornicus fiddles with a logic simulator.
23:31 * C_tiger simulates illogic
23:31
< Molgorn>
Fairly well. Doesn't yet do anything, because the various components have no actions associated, and it doesn't link because I can't seem to work out how to #include my yy.lex.c without the linker screaming about multiple definitions
23:31
< Molgorn>
But it doesn't complain about my syntax, at least ¬¬
23:32
< Molgorn>
From it not existing yesterday evening, this is Progress
23:32
<@Vornicus>
gah
23:32 * Vornicus misclicks, joins the "include" channel
23:32 * C_tiger points, laughs.
23:32
<@McMartin>
If you really don't care about performance, just use an association list and claim it's a one-bucket hash~
23:33
< Molgorn>
(and if I try to just #include the yy.lex.h file that defines yylex(), the compiler explodes into 25-odd errors)
23:33
<@McMartin>
(Duplicate the yylex prototype in your tokens.h)
23:34
< Molgorn>
(instead or as well?)
23:35
<@McMartin>
(as well. Multiple prototypes that are 'the same' should be permitted)
23:35
< Molgorn>
hrm
23:36
<@McMartin>
Or, for that matter, have a completely separate lexer.h file that eveyrone that isn't yy.lex.c can include.
23:36
< Molgorn>
"tokens.h:3: parse error before `token'"
23:36
<@McMartin>
Oh. Right. tokens.h it is, thn.
23:37
<@McMartin>
You have to define the token type before you can put it in a signature.
23:37
<@McMartin>
Put it after the defenition of the token struct.
23:37
< Molgorn>
I did
23:37
<@McMartin>
Also, except under very special and unspeakably heinous conditions, one must never ever ever #include a .c file
23:37
<@McMartin>
Not if your error is on line 3 you didn't.
23:37
< Molgorn>
It's on line 15 as of two seconds after I pasted that line
23:38
<@McMartin>
Same error?
23:38
< Molgorn>
(And I note that, in defence of my #include yy.lex.c, that The Lecturer Made Me Do It)
23:38
<@McMartin>
Something's screwy with the typedef if so.
23:38
< Molgorn>
Same error
23:38 * McMartin steeples his fingers
23:39
<@ToxicFrog>
To the pastebin?
23:39
<@McMartin>
Right, it's because in including yy.lex.c you're including tokens.h twice
23:39 AnnoDomini [AnnoDomini@83.21.41.ns-3845] has quit [Quit: Why is there a jalapeno in the Aegean Sea?]
23:39
<@McMartin>
And typedeffing the same type twice will confuse it.
23:39
<@McMartin>
So, put at the top of tokens.h
23:39
<@McMartin>
#indef TOKENS_H_
23:39
<@McMartin>
#define TOKENS_H_
23:39
<@McMartin>
And at the very bottom,
23:39
<@McMartin>
#endif
23:40
<@McMartin>
That will insure that its text is only ever #included once per source file.
23:40
<@McMartin>
Er
23:40
<@McMartin>
#ifndef, not #indef.
23:41
<@ToxicFrog>
Molgorn: did the lecturer give any reason for including lex.yy.c?
23:41 mode/#code [+v Molgorn] by ChanServ
23:41
<@ToxicFrog>
Because doing so is Stupid And Wrong.
23:42
<+Molgorn>
More or less "Because you can get away with it"
23:42
<+Molgorn>
tokens.h:16: parse error before `token' still, idly
23:42
<@McMartin>
OK, pastebin tokens.h
23:42
<@ToxicFrog>
Hurt the lecturer for me, plz
23:42
< C_tiger>
Man, I wish I knew about #code when I took my first CS class.
23:43
<+Molgorn>
http://rafb.net/p/g2dXiT11.html
23:43
<@McMartin>
token *yylex(void)
23:43
< C_tiger>
put #endif at the bottom, also.
23:44
<@McMartin>
The struct name is "struct token_struct"
23:44
<@McMartin>
token is a shorthand for that.
23:44
<@ToxicFrog>
That should be either "token *" or "struct token_struct *"
23:44
<@McMartin>
and the * goes after the typename, yeah.
23:44
<@ToxicFrog>
Also, move the #endif at the top to the bottom
23:44
<@ToxicFrog>
The purpose of those directives is: if TOKENS_H_ is not already defined, define it and <here's the rest of the file>
23:44
<@ToxicFrog>
Otherwise, skip the entire header
23:44
<@McMartin>
Also, new_token really shouldn't be defined in tokens.h
23:44
<@McMartin>
Line 22 should be part of some .c file
23:45
<@ToxicFrog>
So, the #ifndef at the start needs to match up with an #endif at the bottom.
23:45
<@McMartin>
Possibly yy.lex.c, if you define it right above your definition of main.
23:45
<@McMartin>
(in lexer.l, that is)
23:45
<@ToxicFrog>
McMartin: if it's declared in the header, he can define it wherever he likes
23:45
<+Molgorn>
main's defined in parser.c now
23:45
<+Molgorn>
and no longer in lexerl.
23:45
<@McMartin>
OK
23:45
<+Molgorn>
.l*
23:45
<@McMartin>
Well, it should go in some .c file.
23:45
<@McMartin>
It should not be in the header, or once you have two files that #include tokens.h, they won't link because the function will have been apparently defined twice.
23:46
<@McMartin>
As opposed to declared, which is what line 20 does.
23:46
<@ToxicFrog>
And even if the two definitions are the same, that is Not Allowed; you can declare something as many times as you like, as long as all the declarations agree, but you can only define it once.
23:46
<+Molgorn>
I note that changing struct * token yylex(void) to token yylex(void) makes the compiler cry
23:47
<@McMartin>
You forgot the * before yylex.
23:49
<+Molgorn>
So noted :) right, now it's back to the same "multiple definition of Thing in yy.lex.c and parser.c"
23:50
<@McMartin>
Stop #including yy.lex.c, and let the linker do its job
23:50
<@McMartin>
If the Lecturer *absolutely insists*, then remove yy.lex.c from the project entirely.
23:51
<+Molgorn>
If I don't #include yy.lex.c, the compiler explodes into 28 errors
23:51
<@McMartin>
Of the form?
23:51
<@McMartin>
And what is "Thing"?
23:52
<+Molgorn>
"thing" is a big list of the methods flex causes yy.lex.c to contain
23:52
<+Molgorn>
yyput, yyunput, yylex, yyhands, yyknees, yybumpsadaisy
23:52
<@McMartin>
Right, hence why #including it is a bad idea
23:52
<@McMartin>
Anyway, what are the 28 errors?
23:52
<@ToxicFrog>
The way this should work is that the compiler turns them into yy.lex.o and parser.o
23:52
<@ToxicFrog>
And then the linker combines them.
23:53
<@McMartin>
And while we're at it, it's entirely possible that the 28 errors are in fact linker errors, not compiler errors
23:53
<@McMartin>
Since you haven't told us in any detail what's gone wrong.
23:53
<+Molgorn>
Lots of "This isn't declared!" with a sprinkling of "implicit declaration"s and "syntax error"s
23:53
<@McMartin>
OK, so, syntax errors are problems
23:54
<@McMartin>
Does parser.c include tokens.h?
23:54
<@Vornicus>
undeclared/implicit declarations are linker errors.
23:54
<+Molgorn>
Ah. No, since it was #including yy.lex.c which does.
23:54
<@McMartin>
"implicit declaration" means you're calling a function it doesn't know about, either because you didn't declare it in that file (or one of its #includes) or because you misspelled the function name.
23:54
<+Molgorn>
9 linker errors now.
23:55
<@McMartin>
What are the missing methods?
23:56
<+Molgorn>
Likewise I forgot to replace stdio
23:56
<+Molgorn>
4 errors.
23:56
<@McMartin>
If one of the missing functions is yylex, then your dev-cpp project is misconfigured.
23:56
<+Molgorn>
No, there's only one left, which is yyin
23:56
<+Molgorn>
And that's because I was using it before for the command-line test rig
23:57
<@ToxicFrog>
...including stdio wouldn't fix linker errors
23:57
<@McMartin>
Another way around this is to do a redeclaration.
23:57
<@ToxicFrog>
Unless they were preceded by a bunch of "implicit declaration of foo" errors
23:57
<@McMartin>
TF: It will fix implicid declaration of printf and friends.
23:57
<@ToxicFrog>
Well, s/errors/warnings that really should be errors/
23:57
<@McMartin>
(To declare a variable, you put "extern" in front of it)
23:57
<@McMartin>
(So, "extern FILE *yyin;")
23:58
<+Molgorn>
nifty!
23:58
<+Molgorn>
It compiles! and links!
23:58
<+Molgorn>
zooomj
23:59
<+Molgorn>
Thanks, lads!
23:59
<@McMartin>
Not really. C is just unutterably stupid and needs to be told everything a million times.
23:59
<+Molgorn>
(of course, since it doesn't generate any output from any input, it just exits as soon as you give it anything, but that's the next step)
23:59
<@McMartin>
It's just good enough at what it does that nobody bothers writing a replacement.
23:59
<+Molgorn>
wheeeeee
--- Log closed Tue Mar 11 00:00:24 2008
code logs -> 2008 -> Mon, 10 Mar 2008< code.20080309.log - code.20080311.log >
 search of the:   for: