- Haskell-Cafe - Haskell.org

Re: Debugging Haskell
by Jerry, JiJie 13 Mar '02

13 Mar '02

Thank you very much, it works :) As I think it might be of help to the others, I'm cc'ing this mail to the mailing list where my question was initially posted with the revised code attached. Sincerely, Jerry * Simon Marlow <simonmar(a)microsoft.com> [020313 20:16]: > [ redirected to glasgow-haskell-bugs(a)haskell.org ] > > > Good day everyone, I was fiddling around with this tiny echo > > client/server haskell program from 'The Great Language Shootout' > > site (http://www.bagley.org/~doug/shootout/) and got stuck. > > > > The code (attached) has been reformatted with minimal API tweak > > (mkPortNumber, writeSocket, readSocket) to please my ghc-5.02.2, and > > all what I get is something stuck forever after the first > > iteration: > > > > $ ./echo 3 > > Client wrote: Hello there sailor > > Server recv: Hello there sailor > > Server read: Hello there sailor > > Server wrote: Hello there sailor > > > > After adding all these print statement, I still don't have a clue > > what's jammed there. Hope someone here can shred some light. > > It turns out to be a bug in the network library; we weren't putting the > socket returned from accept into non-blocking mode. It works fine if > you use Handles rather than send/recv because the act of making a Handle > from a file descriptor sets non-blocking mode on the FD, so a workaround > for your program is just to insert a call to socketToHandle on the > socket returned from accept (you don't have to use the Handle, just > calling socketToHandle has the desired effect). > > Thanks for the report. > > > BTW, I'd also like to take this chance to ask how to debug a haskell > > program in general? > > With putStr or IOExts.trace, or using one of the more sophisticated > debugging tools such as the nhc98 tracing system or Andy Gill's Observe > library. You should be able to find links on the www.haskell.org pages > to these projects. > > Cheers, > Simon >

1 0

Question about something in Hudak's book
by Ludovic Kuty 13 Mar '02

13 Mar '02

Hello, I have just begun reading Hudak's The Haskell School of Expression and i am wondering something about the manner he calculates the area of a polygon (made of vertices) on page 27. He wrote: area(Polygon (v1:vs)) = polyArea vs where polyArea :: [Vertex] -> Float polyArea (v2:v3:vs') = triArea v1 v2 v3 + polyArea(v3:vs') polyArea _ = 0 i do not understand why variable v1 is pattern-matched in the definition of area and made "global" to polyArea instead of doing it with polyArea like this: area(Polygon vs) = polyArea vs where polyArea :: [Vertex] -> Float polyArea (v1:v2:v3:vs') = triArea v1 v2 v3 + polyArea(v3:vs') polyArea _ = 0 Is it an idiom or some sort of optimization ? Or just the way he likes it ? TIA Ludovic Kuty

2 2

Re: IO and syntax
by Jyrinx 13 Mar '02

13 Mar '02

> readIL' should be defined differently. Try: > > readIL' ans = > do > a <- ... > b <- ... > if a==0 && b==0 then return ans else readIL' ((a,b):ans) > > (note that I added "return".) > I don't know what the ...s should be, but they definitely should not be > what they are: readDec is some kind of number reader, but it doesn't > deal with actually getting input. Other people here will know more. > One way is to use getStr (or is it getStrLn??), split it at > the space, and then use read to get the numbers. However, I suspect > there is a better way. Good luck. Ah! You were right - I was confused as to the purpose of readDec and friends. I changed it to the following, which works: readIncompatList :: IO [(Player, Player)] readIncompatList = readIL' [] where readIL' ans = do ln <- getLine let [a, b] = map read (words ln) if a == 0 && b == 0 then return ans else readIL' ((a,b):ans) Thanks! Jyrinx jyrinx_list(a)mindspring.com (It occurs to me ... I love Haskell: that "let [a, b] = map read (words ln)" would be so much longer in C ... :-) )

1 0

Hiring and porting
by Mark Carroll 13 Mar '02

13 Mar '02

Thanks, everyone, for your responses! It's all been very helpful. Some things I should mention, then: We're based in central Ohio, but are not currently hiring FPers. Whether we will be in the future depends somewhat on this porting issue. However, if we do decide to hire any Haskell programmers, I shall mention that here. Ideally, yes, we'd look to hire practising programmers who already know Haskell and wish that they could be using it in their job, although I'd fear that there may not be enough such people available to us. The larger application is a simulation engine that involves parsing, symbolic processing, etc. in which correctness in important and that will require ongoing improvement and modification, so I was thinking that use of Haskell would be appropriate. (Use of C++ would make it faster, though, I expect. (-:) However, I do fear that Ashley's correct in suggesting that you'd probably need to rewrite everything to sensibly translate the Haskell to C or Java or whatever, and it is both reasonable and plausible that some larger clients will demand use of a more mainstream language in anything that we deliver to them so that they don't rely on us for maintenance. So, first we have to figure out if we should use Haskell at all, because a likely need for a non-trivial port in the future could easily negate, in time and cost, the initial productivity benefits we might gain from Haskell. -- Mark

2 1

Re: IO and syntax
by Ashley Yakeley 13 Mar '02

13 Mar '02

At 2002-03-12 20:18, Jyrinx wrote: >Of course, this "probable cause" is ridiculous at face value ... anyway, >my code makes as much sense to me as it can at the moment ... how should >I get it to make sense to GHC? (I couldn't find any examples on I can see the bug. It's a very common one when using monads and the 'do' notation -- confusing 'a' with 'IO a'. When in doubt, add type signatures to everything. You'll soon narrow it down. -- Ashley Yakeley, Seattle WA

1 0

IO and syntax
by Jyrinx 12 Mar '02

12 Mar '02

I just got back from a high-school programming competition where the only allowed languages were C[++] and Java (*grmbl*). To see just how cool Haskell is, I'm redoing some of the problems in this more enlightened language :-) Anyway, that explains the rather strict requirements for my program's interface. At one point, the user must enter pairs of numbers like so: 1 4 2 4 3 2 4 1 4 3 0 0 (where two zeros terminate the input). Here is my latest attempt at a function to do this part of the problem (and it's the easy part ...) --- snip --- -- type Player = Int readIncompatList :: IO [(Player, Player)] readIncompatList = readIL' [] where readIL' ans = do a <- readDec b <- readDec if a == 0 && b == 0 then ans else readIL' ((a, b):ans) --- snip --- GHC complains with the following: --- snip --- prob13.hs:37: Couldn't match `String -> t' against `[t1]' Expected type: String -> t Inferred type: [t1] Probable cause: `(:)' is applied to too many arguments in the call ((:) (a, b) ans) In the first argument of `readIL'', namely `((a, b) : ans)' --- snip --- Of course, this "probable cause" is ridiculous at face value ... anyway, my code makes as much sense to me as it can at the moment ... how should I get it to make sense to GHC? (I couldn't find any examples on something as mundane as console I/O; funny how people are so fixated on the real stuff :-) ) Thanks! Jyrinx jyrinx_list(a)mindspring.com

1 0

TrafficMagnet - Special Offer!
by Christine Hall 12 Mar '02

12 Mar '02

ReplyTo:"Christine Hall"<christine(a)trafficmagnet.net> Hi! Did you know that 85% of your potential customers will be using search engines to find what they are looking for on the Internet? Have you ever thought about getting your website listed on search engines worldwide? TrafficMagnet offers a unique technology that will submit your website to over 300,000 search engines and directories every month. We can help your customers find you! Submit your website to more than 300,000 search engines and directories http://haskell.org Normal Price: $14.95 per month Special Price: $9.95 per month You Save: more than 30% off Sign up today at http://www.trafficmagnet.net Benefit now! It's easy, it's affordable and you can sign up online. I look forward to hearing from you. Best Regards, Christine Hall Sales and Marketing E-mail: christine(a)trafficmagnet.net http://www.TrafficMagnet.net

1 0

RE: Hiring Haskell programmers
by Konst Sushenko 12 Mar '02

12 Mar '02

I have always been wondering what exactly does "quickly learn Haskell" mean? Quickly learn Haskell syntax? Can one learn how to paint quickly? konst > -----Original Message----- > From: Eray Ozkural [mailto:erayo@cs.bilkent.edu.tr] > Sent: Monday, March 11, 2002 4:24 PM > To: Mark Carroll; haskell-cafe(a)haskell.org > Subject: Re: Hiring Haskell programmers > > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On Tuesday 12 March 2002 01:17, Mark Carroll wrote: > > How easy is it to hire reasonable Haskell programmers? Of > course, this may > > mean, hiring people with the aptitude and interest to quickly learn > > Haskell. Has anyone any experience of this that they can share? > > > > I know it's easy. Try me. ;) > > Sincerely, > > - -- > Eray Ozkural (exa) <erayo(a)cs.bilkent.edu.tr> > Comp. Sci. Dept., Bilkent University, Ankara > www: http://www.cs.bilkent.edu.tr/~erayo > GPG public key fingerprint: 360C 852F 88B0 A745 F31B EA0F > 7C07 AE16 874D 539C > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.0.6 (GNU/Linux) > Comment: For info see http://www.gnupg.org > > iD8DBQE8jUqcfAeuFodNU5wRAt1/AKCaN881FMXqzpx+xh1EpnFa/b6k9ACeNuPX > aYrkL3CfiF2C6uBzH+3chPs= > =U7Xm > -----END PGP SIGNATURE----- > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe >

4 3

RE: Hiring Haskell programmers
by Ashley Yakeley 12 Mar '02

12 Mar '02

At 2002-03-11 17:05, Mark Carroll wrote: >When I first learned Standard ML, after years of imperative >programming, my brain almost hurt for the first few weeks. For me the difference is that in imperative you tell the computer what to do, whereas in FP you tell the computer what things are. -- Ashley Yakeley, Seattle WA

2 1

Conversion/porting to "mainstream" languages
by Mark Carroll 12 Mar '02

12 Mar '02

One criticism I've received of the suggestion that we use Haskell in our business is that some particularly large clients will demand code in some 'standard' language that they know they can deal with, especially if they end up stuck with the software and we stop supporting it. How are others dealing with this? How big a project would it be to translate, say, from Haskell to C or Java? I'm guessing that you'd borrow a lot of the type-inferencing, etc. from the current compilers - how easy is their code to get to grips with? I suppose that at least something to translate Haskell to Java bytecode would be useful someday, and then you could 'decompile' back to Java, perhaps, but as we wouldn't be doing graphical stuff with Haskell anyway then C might be best. Anyhow, any thoughts or comments on this difficulty would be of interest. -- Mark

2 1