Re: [Haskell-cafe] [iteratee] how to do nothing .. properly
Hi Sergey, I can't explain this; maybe it's a bug in enumWith? I'll look into it. Thanks, John Message: 20 Date: Thu, 2 Jun 2011 02:46:32 +0400 From: Sergey Mironov ier...@gmail.com Subject: [Haskell-cafe] [iteratee] how to do nothing .. properly To: haskell-cafe@haskell.org Message-ID: BANLkTimMFRWgH9Nopt-eua+L7jQcGq+u=g...@mail.gmail.com Content-Type: text/plain; charset=ISO-8859-1 Hi. Would anybody explain a situation with iter6 and iter7 below? Strange thing - first one consumes no intput, while second consumes it all, while all the difference is peek which should do no processing (just copy next item in stream and return to user). What I am trying to do - is to write an iteratee consuing no input, but returning a constant I give to it. I thought (return a) should do it, but it seems I was wrong as return actually consumes all unparsed stream. iter6 experience tells me that (peekreturn a) is what I need, but it's completely confusing and not what I expected. Thanks, Sergey import Data.Iteratee as I import Data.Iteratee.IO import Control.Monad import Control.Exception import Data.ByteString import Data.Char import Data.String -- countBytes :: (..., Num b) = Iteratee s m a - Iteratee s m (a, b) countBytes i = enumWith i I.length iter6 = do h - countBytes $ (peek return 0) s - I.stream2list return (h,s) iter7 = do h - countBytes $ (return 0) s - I.stream2list return (h,s) print6 = enumPure1Chunk [1..10] (iter6) = run = print print7 = enumPure1Chunk [1..10] (iter7) = run = print Here is example ghci session *Main print6 ((0,0),[1,2,3,4,5,6,7,8,9,10]) -- read 0 items, returns 0 *Main print7 ((0,10),[]) -- read 10 items (???) returns 0 *Main -- ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe End of Haskell-Cafe Digest, Vol 94, Issue 3 *** ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] [iteratee] how to do nothing .. properly
Ok. I've checked iteratee-0.8.3.0 and 0.8.4.0. Results are same. Sergey 2011/6/2 John Lato jwl...@gmail.com: Hi Sergey, I can't explain this; maybe it's a bug in enumWith? I'll look into it. Thanks, John Message: 20 Date: Thu, 2 Jun 2011 02:46:32 +0400 From: Sergey Mironov ier...@gmail.com Subject: [Haskell-cafe] [iteratee] how to do nothing .. properly To: haskell-cafe@haskell.org Message-ID: BANLkTimMFRWgH9Nopt-eua+L7jQcGq+u=g...@mail.gmail.com Content-Type: text/plain; charset=ISO-8859-1 Hi. Would anybody explain a situation with iter6 and iter7 below? Strange thing - first one consumes no intput, while second consumes it all, while all the difference is peek which should do no processing (just copy next item in stream and return to user). What I am trying to do - is to write an iteratee consuing no input, but returning a constant I give to it. I thought (return a) should do it, but it seems I was wrong as return actually consumes all unparsed stream. iter6 experience tells me that (peekreturn a) is what I need, but it's completely confusing and not what I expected. Thanks, Sergey import Data.Iteratee as I import Data.Iteratee.IO import Control.Monad import Control.Exception import Data.ByteString import Data.Char import Data.String -- countBytes :: (..., Num b) = Iteratee s m a - Iteratee s m (a, b) countBytes i = enumWith i I.length iter6 = do h - countBytes $ (peek return 0) s - I.stream2list return (h,s) iter7 = do h - countBytes $ (return 0) s - I.stream2list return (h,s) print6 = enumPure1Chunk [1..10] (iter6) = run = print print7 = enumPure1Chunk [1..10] (iter7) = run = print Here is example ghci session *Main print6 ((0,0),[1,2,3,4,5,6,7,8,9,10]) -- read 0 items, returns 0 *Main print7 ((0,10),[]) -- read 10 items (???) returns 0 *Main -- ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe End of Haskell-Cafe Digest, Vol 94, Issue 3 *** ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] How on Earth Do You Reason about Space?
Since frequency counts are an important use of map-like data structures, I did a brief test of the available options. First using regular strings for input, and Data.Map.fromListWith - i.e. the operational bit being: freqs :: [String] - M.Map String Int freqs = M.fromListWith (+) . map (,1) This runs on a 14M corpus consisting of King James Bible, collected works of Shakespeare, and War and Peace. ./freqstr1 +RTS -s 5,093,386,368 bytes allocated in the heap 2,682,667,904 bytes copied during GC 261,110,368 bytes maximum residency (20 sample(s)) 9,018,000 bytes maximum slop 623 MB total memory in use (10 MB lost due to fragmentation) ./freqstr1 +RTS -s 21.43s user 0.78s system 99% cpu 22.285 total Kinda expensive, 250MB to store word frequencies of 14MB text. Now, changing to freqs :: [String] - M.Map String Int freqs = foldl' (\m w - M.insertWith' (+) w 1 m) M.empty i.e. using strict insertion, avoiding the buildup of lazy thunks for the counts. ./freqstr2 +RTS -s -- strings, using strict insertion 4,754,110,096 bytes allocated in the heap 2,089,527,240 bytes copied during GC 27,039,112 bytes maximum residency (66 sample(s)) 613,192 bytes maximum slop 80 MB total memory in use (2 MB lost due to fragmentation) ./freqstr2 +RTS -s 17.48s user 0.13s system 99% cpu 17.665 total This reduced maximam memory consumption to one tenth, still bigger than input corpus, but clearly not too bad. A bit faster, too, in spite of probably doing more work. Using ByteStrings instead, first fromListWith: ./freq +RTS -s (Just 77432,113931) 3,880,059,568 bytes allocated in the heap 1,467,507,808 bytes copied during GC 174,573,752 bytes maximum residency (14 sample(s)) 8,222,600 bytes maximum slop 385 MB total memory in use (6 MB lost due to fragmentation) ./freq +RTS -s 14.26s user 0.49s system 99% cpu 14.798 total About half the memroy of Strings, and 25% faster. With strict insertion: ./freq2 +RTS -s -- map using strict insertion 3,761,614,312 bytes allocated in the heap 849,806,000 bytes copied during GC 23,950,328 bytes maximum residency (35 sample(s)) 2,376,904 bytes maximum slop 58 MB total memory in use (1 MB lost due to fragmentation) ./freq2 +RTS -s 11.14s user 0.13s system 99% cpu 11.295 total Parallel to the String case, this is a lot more frugal with memory, and 30% faster. Now, I tried Data.HashMap from the hashmap library: ./freqH1 +RTS -s-- hashmap using fromListWith 4,552,922,784 bytes allocated in the heap 2,990,287,536 bytes copied during GC 401,247,912 bytes maximum residency (14 sample(s)) 42,098,016 bytes maximum slop 957 MB total memory in use (15 MB lost due to fragmentation) ./freqH1 +RTS -s 15.68s user 1.53s system 99% cpu 17.277 total ./freqH2 +RTS -s -- hashmap using foldl' insertWith 4,518,146,968 bytes allocated in the heap 2,986,973,352 bytes copied during GC 394,502,832 bytes maximum residency (14 sample(s)) 41,020,040 bytes maximum slop 957 MB total memory in use (15 MB lost due to fragmentation) ./freqH2 +RTS -s 15.86s user 1.62s system 99% cpu 17.537 total HashMap doesn't provide a strict insertWith, so this is similar to the lazy insertions above. A bit worse, actually, probably due to the overhead of hashing. Then, I discovered that Johan's hashmap is a different library, and thought I'd try that too for completeness. ./freqHS +RTS -s -- hashmap strict (unordered-containers) 2,628,628,752 bytes allocated in the heap 945,571,872 bytes copied during GC 26,635,744 bytes maximum residency (32 sample(s)) 2,433,504 bytes maximum slop 66 MB total memory in use (1 MB lost due to fragmentation) ./freqHS +RTS -s 6.90s user 0.16s system 99% cpu 7.082 total Memory residency like the other strict versions, but really fast, probably due to faster comparisons of hash values vs comparisons of strings. Conclusion: make sure you are using a strict map, and if your keys are strings or otherwise have expensive comparisons, unordered-containers is the library for you. -k PS: I also tried mapping 'copy' on the input words to avoid storing large slices of the input, but it only worsened things: ./freqHS3 +RTS -s (Just 77432,113931) 3,109,585,024 bytes allocated in the heap 936,724,184 bytes copied during GC 87,831,888 bytes maximum residency (19 sample(s)) 8,835,440 bytes maximum slop 164 MB total memory in use (3 MB lost due to fragmentation) ./freqHS3 +RTS -s 12.71s user 0.31s system 99% cpu 13.060 total Perhaps if you managed to only copy new words it would look better? PPS: I tried to be careful juggling the results around, but there's always the possiblity of a mistake. Caveat lector! (Or should that be 'cave scriptor'?) PPPS: There are some small interface annoyances around, it'd be nice
[Haskell-cafe] Iteratees again (Was: How on Earth Do You Reason about Space?)
By the way, what is the advantage of using iteratees here? For my testing, I just used: main = printit . freqs . B.words = B.readFile words (where 'printit' writes some data to stdout just to make sure stuff is evaluated, and you've already seen some 'freqs' examples) I have a bunch of old code, parsers etc, which are based on the 'readFile' paradigm: type Str = Data.ByteString.Lazy.Char8.ByteString -- usually decodeFoo :: Str - Foo encodeFoo :: Foo - Str readFoo = decodeFoo . readFile writeFoo f = writeFile f . encodeFoo hReadFoo = decodeFoo . hRead : (etc) This works pretty well, as long as Foo is strict enough that you don't retain all or huge parts of input, and as long as you can process input in a forward, linear fashion. And, like my frequency count above, I can't really see how this can be made much simpler. I haven't used iteratees or enumartors in anger, but it appears to me that they are most useful when the input is unpredictable or needs to be controlled in some way - for instance, when recv() can return a blocks of data that may be too little or too much. Would there be any great advantage to rewriting my stuff to use iterators? Or at least, use iterators for new stuff? As I see it, iterators are complex and the dust is only starting to settle on implementations and interfaces, and will introduce more dependencies. So my instinct is to stick with the worse-is-better approach, but I'm willing to be educated. -k -- If I haven't seen further, it is by standing in the footprints of giants ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] *GROUP HUG*
On Thursday 02 June 2011 01:12:37, Tom Murphy wrote: How about this: myFoldr :: (a - b - b) - b - [a] - b myFoldr f z xs = foldl' (\s x v - s (x `f` v)) id xs $ z Cheers, Ivan Great! Now I really can say Come on! It's fun! I can write foldr with foldl! Unfortunately, you can't, not quite. A left fold cannont return anything before it has reached the end of the list, so it doesn't work on infinite lists. With a suitable combining function (and suitable input), a right fold can start delivering the result early, hence right folds (can) produce results for infinite lists (and my example was meant to draw attention to that fact). Cheers, Daniel ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Iteratees again (Was: How on Earth Do You Reason about Space?)
Hi Ketil, By the way, what is the advantage of using iteratees here? For my testing, I just used: My initial move to iteratees was more a clutch call I made when I was still using bytestring-trie, and was having immense memory consumption problems. bytestring-trie uses strict byte strings as an index, and since I was getting only lazy byte strings, the only way to make them strict would be to use (S.concat . L.toChunks) (L and S being the lazy/strict byte string imports,) which felt *wrong*. In short, I thought iteratee would give me enough magic fairy dust to actually have a decent control over how much data I'm holding in RAM at any given point — that was not the case, since I didn't know about the pointer mechanic of strict ByteStrings and hence was oblivious to the bad impact that would have on garbage collection performance. Even so, I think I can still justify using iteratees in the current design: a) I don't like lazy IO (conceptually,) b) I'm gonna write a left-fold somewhere anyway, might as well use a decent infrastructure for it c) I can strictly control the chunk size, and I'm not going to have any bad effects with accidental eager evaluation somewhere down the pipe. c) being the only legitimate reason (though the reason for a) is c) ) — adjusting the chunk size might actually yield noticeable performance differences when reading through files that are well into the realm of gigabytes. And the chunk size limit will protect me from an accidental strict fold or so that would leave me with a 4GB file in memory. About a): Lazy IO just doesn't feel right for me. I want my pure computations to actually be pure. If I put a ' on one of my functions *within* my pure code, this might have *side effects* — now, instead of reading in only part of the file, this will demand the *whole* file, and that is *quite* a side effect! So, suddenly I have to worry about side effects in my pure code. Ugh. That's why I'm going to continue using iteratees. I don't know if that's the right justification, but it's a hey, it works for me! justification I can comfortably live with. Besides, I don't think the iteratee interface is all that opaque. I found arrows in HXT, for example, much more difficult to deal with conceptually. (That said, I'm still using HDBC over Takusen, because the latter's API just didn't make sense to me.) Regards, Aleks signature.asc Description: Digital signature ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] Attoparsec concatenating combinator
I often find while using attoparsec and attoparsec-text that I need to match a number of text parsers consecutively and concatenate the result. By text parser I mean Parser ByteString for attoparsec and Parser Text for attoparsec-text. It seems the best I can do is to collect them all in a list and then apply concat. But that still copies the text several times. Is there a combinator that does this without all that copying? If not, does the internal representation easily admit such a combinator? Thanks, Yitz ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] [iteratee] how to do nothing .. properly
Hi Sergey, I've got an explanation; quite surprisingly it's a bug in enumPure1Chunk. Even though it is an odd case, I'm surprised that it hasn't come up before now since enumPure1Chunk appears frequently. I've just uploaded 0.8.5.0 which has the fix. There's now an additional Monoid constraint on enumPure1Chunk, unfortunately. Thanks very much for reporting this. John L On Thu, Jun 2, 2011 at 10:02 AM, Sergey Mironov ier...@gmail.com wrote: Ok. I've checked iteratee-0.8.3.0 and 0.8.4.0. Results are same. Sergey 2011/6/2 John Lato jwl...@gmail.com: Hi Sergey, I can't explain this; maybe it's a bug in enumWith? I'll look into it. Thanks, John Message: 20 Date: Thu, 2 Jun 2011 02:46:32 +0400 From: Sergey Mironov ier...@gmail.com Subject: [Haskell-cafe] [iteratee] how to do nothing .. properly To: haskell-cafe@haskell.org Message-ID: BANLkTimMFRWgH9Nopt-eua+L7jQcGq+u=g...@mail.gmail.com Content-Type: text/plain; charset=ISO-8859-1 Hi. Would anybody explain a situation with iter6 and iter7 below? Strange thing - first one consumes no intput, while second consumes it all, while all the difference is peek which should do no processing (just copy next item in stream and return to user). What I am trying to do - is to write an iteratee consuing no input, but returning a constant I give to it. I thought (return a) should do it, but it seems I was wrong as return actually consumes all unparsed stream. iter6 experience tells me that (peekreturn a) is what I need, but it's completely confusing and not what I expected. Thanks, Sergey import Data.Iteratee as I import Data.Iteratee.IO import Control.Monad import Control.Exception import Data.ByteString import Data.Char import Data.String -- countBytes :: (..., Num b) = Iteratee s m a - Iteratee s m (a, b) countBytes i = enumWith i I.length iter6 = do h - countBytes $ (peek return 0) s - I.stream2list return (h,s) iter7 = do h - countBytes $ (return 0) s - I.stream2list return (h,s) print6 = enumPure1Chunk [1..10] (iter6) = run = print print7 = enumPure1Chunk [1..10] (iter7) = run = print Here is example ghci session *Main print6 ((0,0),[1,2,3,4,5,6,7,8,9,10]) -- read 0 items, returns 0 *Main print7 ((0,10),[]) -- read 10 items (???) returns 0 *Main -- ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe End of Haskell-Cafe Digest, Vol 94, Issue 3 *** ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] [iteratee] how to do nothing .. properly
I am glad to help! Looks like upgrading to 0.8.5.0 also fixes initial problem that involved me into testing! I'll take the opportunity and ask another thing about iteratee: Is it expected behavior that throwErr consumes all data in current chunk? I wish it to stop in place and let after-checkErr code to continue the parsing. Well, I already found solution (or workaround?) - I wrap Iteratee with ErrorT monad and use ErrorT's raiseError instead of throwErr. Is it correct? Here is example code instance Exception Int iter4 = do I.dropWhile (/= 3) h-I.head throwErr $ toException $ (-4::Int) -- doesn't meter what exactly to throw return h -- catch the error with checkErr iter5 = do (_,b)-countBytes $ I.checkErr $ iter4 s - I.stream2list return (b,s) print5 = enumPure1Chunk [1..10] (iter5) = run = print Thanks a lot! Sergey 2011/6/2 John Lato jwl...@gmail.com: Hi Sergey, I've got an explanation; quite surprisingly it's a bug in enumPure1Chunk. Even though it is an odd case, I'm surprised that it hasn't come up before now since enumPure1Chunk appears frequently. I've just uploaded 0.8.5.0 which has the fix. There's now an additional Monoid constraint on enumPure1Chunk, unfortunately. Thanks very much for reporting this. John L On Thu, Jun 2, 2011 at 10:02 AM, Sergey Mironov ier...@gmail.com wrote: Ok. I've checked iteratee-0.8.3.0 and 0.8.4.0. Results are same. Sergey 2011/6/2 John Lato jwl...@gmail.com: Hi Sergey, I can't explain this; maybe it's a bug in enumWith? I'll look into it. Thanks, John Message: 20 Date: Thu, 2 Jun 2011 02:46:32 +0400 From: Sergey Mironov ier...@gmail.com Subject: [Haskell-cafe] [iteratee] how to do nothing .. properly To: haskell-cafe@haskell.org Message-ID: BANLkTimMFRWgH9Nopt-eua+L7jQcGq+u=g...@mail.gmail.com Content-Type: text/plain; charset=ISO-8859-1 Hi. Would anybody explain a situation with iter6 and iter7 below? Strange thing - first one consumes no intput, while second consumes it all, while all the difference is peek which should do no processing (just copy next item in stream and return to user). What I am trying to do - is to write an iteratee consuing no input, but returning a constant I give to it. I thought (return a) should do it, but it seems I was wrong as return actually consumes all unparsed stream. iter6 experience tells me that (peekreturn a) is what I need, but it's completely confusing and not what I expected. Thanks, Sergey import Data.Iteratee as I import Data.Iteratee.IO import Control.Monad import Control.Exception import Data.ByteString import Data.Char import Data.String -- countBytes :: (..., Num b) = Iteratee s m a - Iteratee s m (a, b) countBytes i = enumWith i I.length iter6 = do h - countBytes $ (peek return 0) s - I.stream2list return (h,s) iter7 = do h - countBytes $ (return 0) s - I.stream2list return (h,s) print6 = enumPure1Chunk [1..10] (iter6) = run = print print7 = enumPure1Chunk [1..10] (iter7) = run = print Here is example ghci session *Main print6 ((0,0),[1,2,3,4,5,6,7,8,9,10]) -- read 0 items, returns 0 *Main print7 ((0,10),[]) -- read 10 items (???) returns 0 *Main -- ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe End of Haskell-Cafe Digest, Vol 94, Issue 3 *** ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Attoparsec concatenating combinator
On Thu, Jun 2, 2011 at 7:02 AM, Yitzchak Gale g...@sefer.org wrote: It seems the best I can do is to collect them all in a list and then apply concat. But that still copies the text several times. Right. I'd like a no-copy combinator for the same reasons, but I think it's impossible to do without some low-level support. If not, does the internal representation easily admit such a combinator? Not very easily. Internally, attoparsec maintains just three pieces of data for its state: - The current input - Any input we received via continuations, in case we need to backtrack - A flag that denotes whether we were fed EOF by a continuation There are no line numbers, no bytes consumed counters, nothing else. If there was a bytes consumed counter, it would be possible to write a try-like combinator that would hold onto the current input, run a parser, tack on any input received via continuations to the original input, and then use the counter to slice off a portion of that bytestring without copying. I can't think of another way to do it. Adding that counter would be a moderate amount of work, and would presumably have a negative effect on performance. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Oracle Sessions in Takusen
Dmitry,
I'm not directly familiar with Takusen or its use with OracleDB, but I
would hazard a guess that the withSession is doing FFI resource management
and that resources obtained inside the withSession environment are no
longer valid outside of the withSession.
If this is the case then I would expect the following to work:
replicateM 2 (do
withSession (connect x x x) (do
res - doQuery ...
liftIO $ print res
)
)
If this really is the case then it seems that withSession shouldn't be
exporting FFI-based resources.
-KQ
On Wed, 01 Jun 2011 07:44:10 -0700, Dmitry Olshansky
olshansk...@gmail.com wrote:
Hello,
Could anyone explain strange behavior of Takusen with OracleDB (OraClient
11.x)? Several sequential sessions give Seqmentation Fault error. In
case
of nested sessions it works well.
{-# LANGUAGE ScopedTypeVariables #-}
module Main where
import Database.Oracle.Enumerator
import Control.Monad(replicateM)
import Control.Monad.Trans(liftIO)
main = do
{-
-- This gives an Segmentation Fault for the second session
replicateM 2 (do
res - withSession (connect x x x) (do
doQuery (sql SELECT dummy FROM dual) (\(d::String)
(_::Maybe String) - result' $ Just d) Nothing
)
print res
)
-}
-- This is works well
withSession (connect x x x) (do
r1 - doQuery (sql SELECT dummy FROM dual) (\(d::String)
(_::Maybe String) - result' $ Just d) Nothing
liftIO $ print r1
liftIO $ withSession (connect x x x) (do
r2 - doQuery (sql SELECT dummy FROM dual)
(\(d::String) (_::Maybe String) - result' $ Just d) Nothing
liftIO $ print r2
)
)
Best regards,
Dmitry
--
-KQ
___
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] *GROUP HUG*
Alberto G. Corona agocor...@gmail.com wrote: Haskell is an academic asset as well as a fun asset. I cannot agree with this for practical reasons. I'm using Haskell for real world commercial applications, and I'm very productive with it. There is however a variation of this statement, with which I could agree, namely: Learning Haskell will pay off much less than learning PHP, if your goal is to find a job. It takes a lot longer and there are a lot less companies in need of Haskell programmers. Greets, Ertugrul -- nightmare = unsafePerformIO (getWrongWife = sex) http://ertes.de/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] *GROUP HUG*
Ivan Tarasov ivan.tara...@gmail.com wrote: myFoldr :: (a - b - b) - b - [a] - b myFoldr f z xs = foldl' (\s x v - s (x `f` v)) id xs $ z That's not foldr. It's a function similar to foldr in Haskell and equal to foldr in a different language, which lacks bottom. Greets, Ertugrul -- nightmare = unsafePerformIO (getWrongWife = sex) http://ertes.de/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] *GROUP HUG*
Learning Haskell will pay off much less than learning PHP, if your goal is to find a job. Amen. I cannot agree with this for practical reasons. I'm using Haskell for real world commercial applications, and I'm very productive with it. I wish so much I could say that... Out of curiosity, what are you using Haskell for? 2011/6/2 Ertugrul Soeylemez e...@ertes.de Alberto G. Corona agocor...@gmail.com wrote: Haskell is an academic asset as well as a fun asset. I cannot agree with this for practical reasons. I'm using Haskell for real world commercial applications, and I'm very productive with it. There is however a variation of this statement, with which I could agree, namely: Learning Haskell will pay off much less than learning PHP, if your goal is to find a job. It takes a lot longer and there are a lot less companies in need of Haskell programmers. Greets, Ertugrul -- nightmare = unsafePerformIO (getWrongWife = sex) http://ertes.de/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Oracle Sessions in Takusen
On 3 June 2011 05:35, Kevin Quick qu...@sparq.org wrote: Dmitry, I'm not directly familiar with Takusen or its use with OracleDB, but I would hazard a guess that the withSession is doing FFI resource management and that resources obtained inside the withSession environment are no longer valid outside of the withSession. If this is the case then I would expect the following to work: replicateM 2 (do withSession (connect x x x) (do res - doQuery ... liftIO $ print res ) ) If this really is the case then it seems that withSession shouldn't be exporting FFI-based resources. -KQ You're right, withSession shouldn't be exporting FFI obtained resources, and I don't think it does. There are some known issues with the Oracle code, where it allows buffers to fall out of reference (and thus be gc'd) before the C libs have finished with them, thereby causing segfaults. The known problems are around bind variable buffers, so this looks like a new issue. At least we're collecting a nice corpus of programs that cause the Oracle backend to fail :-) I'd love to have more time to work on it... If you don't need the Oracle-specific functionality, for now I suggest using the ODBC driver as a substitute, as this seems to be (more) reliable. Alistair ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Iteratees again (Was: How on Earth Do You Reason about Space?)
At Thu, 02 Jun 2011 13:52:52 +0200, Ketil Malde wrote: I have a bunch of old code, parsers etc, which are based on the 'readFile' paradigm: type Str = Data.ByteString.Lazy.Char8.ByteString -- usually decodeFoo :: Str - Foo encodeFoo :: Foo - Str readFoo = decodeFoo . readFile writeFoo f = writeFile f . encodeFoo hReadFoo = decodeFoo . hRead : (etc) This works pretty well, as long as Foo is strict enough that you don't retain all or huge parts of input, and as long as you can process input in a forward, linear fashion. And, like my frequency count above, I can't really see how this can be made much simpler. This is fine if you never have parse errors and always read to the end of the file. Otherwise, the code above is incorrect and ends up leaking file descriptors. In general, it is very hard to write parsers that parse every possible input and never fail. Thus, for anything other than a toy program, your code actually has to be: readFoo path = bracket (hOpen path) hclose $ hGetContents = (\s - return $! decodeFoo s) Which is still not guaranteed to work if Foo contains thunks, so then you end up having to write: readFoo path = bracket (hOpen path) hclose $ \h - do s - hGetContents h let foo = decodeFoo s deepseq foo $ return foo Or, finally, what a lot of code falls back to, inserting gratuitous calls to length: readFoo path = bracket (hOpen path) hclose $ \h - do s - hGetContents h length s `seq` return decodeFoo s The equivalent code with the iterIO package would be: readFoo path = enumFile path |$ fooI which seems a lot simpler to me... Would there be any great advantage to rewriting my stuff to use iterators? Or at least, use iterators for new stuff? In addition to avoiding edge cases like leaked file descriptors and memory, one of the things I discovered in implementing iterIO is that it's really handy to have your I/O functions be the same as your parsing combinators. So iteratees might actually admit a far simpler implementation of decodeFoo/fooI. More specifically, imagine that you have decodeFoo, and now want to implement decodeBar where a Bar includes some Foos. Unfortunately, having an implementation of decodeFoo in-hand doesn't help you implement decodeBar. You'd have to re-write your function to return residual input, maybe something like: decodeFooReal :: String - (Foo, String) decodeFoo :: String - Foo decodeFoo = fst . decodeFooReal and now you implement decodeBar in terms of decodeFooReal, but you have to pass around residual input explicitly, handle parsing failures explicitly, etc. As I see it, iterators are complex and the dust is only starting to settle on implementations and interfaces, and will introduce more dependencies. So my instinct is to stick with the worse-is-better approach, but I'm willing to be educated. I fully agree with the point about dependencies and waiting for the dust to settle, though I hope a lot of that changes in a year or so. However, iterIO should already significantly reduce the complexity. David ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] *GROUP HUG*
Yves Parès limestr...@gmail.com wrote: I cannot agree with this for practical reasons. I'm using Haskell for real world commercial applications, and I'm very productive with it. I wish so much I could say that... Out of curiosity, what are you using Haskell for? I use the Yesod web framework for web development and a combination of many libraries for network servers. Greets, Ertugrul -- nightmare = unsafePerformIO (getWrongWife = sex) http://ertes.de/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] ANN: fix-imports-0.1.0
I just uploaded a new package called fix-imports. Maintaining the import block is one of the less fun things I do while writing haskell. So I wrote a little program to do that for me. Basically, I edit source, then hit ,a in vim, and it figures out which modules need to be imported, which imports are not needed, and sorts and formats the import list. I've been using it for the last week or so and it seems to be working ok, so I decided to upload it to hackage in case anyone else doesn't like maintaining import lists manually. The catch: it only works for qualified imports. Here's copy paste from the package description: A small standalone program to manage the import block of a haskell program. It will try to add import lines for qualified names with no corresponding import, remove unused import lines, and sort the import block according to some heuristic you can define. This only works for qualified imports! Unqualified imports are left untouched. It's most convenient if bound to an editor key. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] License of hslogger, HDBC, etc.
Hi Jon all, I've decided that I'm OK with re-licensing hslogger, HDBC, and well all of my Haskell libraries (not end programs) under 3-clause BSD. My schedule is extremely tight right now but if someone wants to send me patches for these things I will try to apply them within the week. -- John ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] *GROUP HUG*
I disagree. I'm by no means proficient in Haskell. And, I never bothered learning PHP. I will when I need to. PHP programmers are a dime a dozen. It's been my experience that Haskell is a tool one may use to distinguish oneself from the hoi-poloi. This is important when you live in an area where the baker down the street has a CS degree. On Thu, Jun 2, 2011 at 11:52 AM, Yves Parès limestr...@gmail.com wrote: Learning Haskell will pay off much less than learning PHP, if your goal is to find a job. Amen. I cannot agree with this for practical reasons. I'm using Haskell for real world commercial applications, and I'm very productive with it. I wish so much I could say that... Out of curiosity, what are you using Haskell for? 2011/6/2 Ertugrul Soeylemez e...@ertes.de Alberto G. Corona agocor...@gmail.com wrote: Haskell is an academic asset as well as a fun asset. I cannot agree with this for practical reasons. I'm using Haskell for real world commercial applications, and I'm very productive with it. There is however a variation of this statement, with which I could agree, namely: Learning Haskell will pay off much less than learning PHP, if your goal is to find a job. It takes a lot longer and there are a lot less companies in need of Haskell programmers. Greets, Ertugrul -- nightmare = unsafePerformIO (getWrongWife = sex) http://ertes.de/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Names for pretty-printing combinators
Casey McCann schrieb: One drastic approach I've used in personal libraries--operator-heavy EDSLs specifically--is to define everything first with alphanumeric names, then put operators in their own modules. In some cases I'd have three such modules: One providing a minimal set of operators that don't clash with anything significant, one providing a larger set of operators that clash with one or more common modules (often deliberately, e.g. duplicating Arrow combinators for some type with similar semantics but no valid Arrow instance), and one providing a bunch of gratuitous Unicode operators that look pretty in my code editor but I don't know how to type in GHCi. I use explicit imports in order to avoid infix operator clashes. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] License of hslogger, HDBC, etc.
On 2 June 2011 22:20, John Goerzen jgoer...@complete.org wrote: Hi Jon all, I've decided that I'm OK with re-licensing hslogger, HDBC, and well all of my Haskell libraries (not end programs) under 3-clause BSD. Awesome, thanks very much! ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] License of hslogger, HDBC, etc.
2011/6/2 John Goerzen jgoer...@complete.org: Hi Jon all, I've decided that I'm OK with re-licensing hslogger, HDBC, and well all of my Haskell libraries (not end programs) under 3-clause BSD. My schedule is extremely tight right now but if someone wants to send me patches for these things I will try to apply them within the week. Thanks! What was your line of reasoning to make the switch? Cheers, Thu ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Status of Haskell + Mac + GUIs graphics
Thanks for these instructions, John. I had to adapt to my 32-bit ghc by adding +universal to the 'port install' lines. I did not find a gtkglext.dpatch. Has it been roled into the gtkglext darcs repo? I guess not, since the last update was Nov 7 (Tag 0.12.0), according to 'darcs changes'. Eventually everything compiled, but I get a linking error. When building an executable, I get Linking dist/build/TestImage/TestImage ... Undefined symbols: _gdk_gl_config_get_screen, referenced from: _s8kj_info in libHSgtkglext-0.12.0.a(Config.o) _gdk_gl_context_is_direct, referenced from: _s73Y_info in libHSgtkglext-0.12.0.a(Context.o) _gdk_gl_context_get_gl_config, referenced from: _s777_info in libHSgtkglext-0.12.0.a(Context.o) _gdk_gl_context_get_render_type, referenced from: _s7lg_info in libHSgtkglext-0.12.0.a(Context.o) _gdk_gl_context_get_share_list, referenced from: _s75v_info in libHSgtkglext-0.12.0.a(Context.o) _gdk_gl_config_get_depth, referenced from: _s8r6_info in libHSgtkglext-0.12.0.a(Config.o) In ghci: *Main main Loading package array-0.3.0.2 ... linking ... done. Loading package bytestring-0.9.1.10 ... linking ... done. ... Loading package cairo-0.12.0 ... linking ... done. Loading package glib-0.12.0 ... linking ... done. Loading package gio-0.12.0 ... linking ... done. Loading package pango-0.12.0 ... linking ... done. Loading package gtk-0.12.0 ... linking ... done. Loading package gtkglext-0.12.0 ... linking ... interactive: unknown symbol `_gdk_gl_context_get_render_type' ghc: unable to load package `gtkglext-0.12.0' *Main Are these failures about the missing patch? - Conal On Thu, May 26, 2011 at 3:03 AM, John Lato jwl...@gmail.com wrote: fltk definitely has some good points, but I've always found it hideously ugly. Of course the default gtk on osx is ugly too, but some of the available themes are nice. However, getting gtk with OpenGL on osx was fairly easy. Everything worked out of the box except gtkglext (Haskell package). I've submitted some patches for that which hopefully will be applied to the repo soon. This is what I used: git darcs ghc-7.0.3 (self-compiled) cabal-install macports gtk2 +no_x11 +quartz macports gtkglext +no_x11 +quartz gtk2hs from source gtkglext (Haskell) from source I use a self-compiled ghc, but this should work with any ghc. Mixing ghc and macports can cause problems with libiconv, but the solutions are pretty well-known. Since I compile ghc, I configure it to use macports's libiconv so there aren't any conflicts. sudo port install gtkglext +no_x11 +quartz (wait a while...) darcs get --lazy http://code.haskell.org/gtk2hs cd gtk cabal install gtk2hs-buildtools chmod +x bootstrap.sh ./bootstrap.sh -fhave-quartz-gtk (wait a while...) cd .. darcs get --lazy http://code.haskell.org/gtkglext cd gtkglext darcs apply gtkglext.dpatch cabal install Until the source tree gets patched, you'll have to manually apply the patch bundle. If you want to avoid macports, it should be possible to use gtk-osx and gtkglext from source instead. However, I've found gtk-osx to be unstable and quite difficult to build in the past, so I'd strongly recommend macports, at least for now. The source install of gtkglext works just fine and can be used with macports gtk2 (I tested this). Since this path already uses macports there's not much benefit though. Cheers, John L On Wed, May 25, 2011 at 9:07 PM, Evan Laforge qdun...@gmail.com wrote: fltk supports OpenGL on windows, X11, and OS X, though I've never used it. The thing it doesn't have is a haskell binding, but as I mentioned I just bind whatever I need when I need it and since I don't need much it's easy. Dunno if this applies in your case though. Maybe it's my NIH, but I like to start with something too simple and add what I need rather than start with something that has more than I need and try to get it working. On Wed, May 25, 2011 at 11:58 AM, Conal Elliott co...@conal.net wrote: Thanks, John. Encouraging bit of news. Please do let us know what you learn when you try. - Conal On Tue, May 24, 2011 at 1:28 AM, John Lato jwl...@gmail.com wrote: You can use gtkglext to get OpenGL support. With the official release of gtkglext-1.2.0 there's a bit of hacking involved (that was probably me you're referring to), but it looks like the repo head has native Quartz. With any luck, you just need to build gtkglext from the repo, then the gtkglext package. I might have some time to try this later today; I'll report back if I get results. John Lato On Tue, May 24, 2011 at 6:01 AM, Conal Elliott co...@conal.net wrote: Last I tried, there wasn't native support for OpenGL with gtk, and I need OpenGL. Then more recently, I heard of some progress in that area, but
Re: [Haskell-cafe] *GROUP HUG*
I got hired at a company because one of the interviewers was impressed that I taught myself Haskell. I basically never use it at work, but I did in my old job. Dave On Thu, Jun 2, 2011 at 2:23 PM, Michael Litchard mich...@schmong.orgwrote: I disagree. I'm by no means proficient in Haskell. And, I never bothered learning PHP. I will when I need to. PHP programmers are a dime a dozen. It's been my experience that Haskell is a tool one may use to distinguish oneself from the hoi-poloi. This is important when you live in an area where the baker down the street has a CS degree. On Thu, Jun 2, 2011 at 11:52 AM, Yves Parès limestr...@gmail.com wrote: Learning Haskell will pay off much less than learning PHP, if your goal is to find a job. Amen. I cannot agree with this for practical reasons. I'm using Haskell for real world commercial applications, and I'm very productive with it. I wish so much I could say that... Out of curiosity, what are you using Haskell for? 2011/6/2 Ertugrul Soeylemez e...@ertes.de Alberto G. Corona agocor...@gmail.com wrote: Haskell is an academic asset as well as a fun asset. I cannot agree with this for practical reasons. I'm using Haskell for real world commercial applications, and I'm very productive with it. There is however a variation of this statement, with which I could agree, namely: Learning Haskell will pay off much less than learning PHP, if your goal is to find a job. It takes a lot longer and there are a lot less companies in need of Haskell programmers. Greets, Ertugrul -- nightmare = unsafePerformIO (getWrongWife = sex) http://ertes.de/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] *GROUP HUG*
Being able to use Haskell at such an early stage of my programming career has given me high expectations of what comes next. On Thu, Jun 2, 2011 at 3:22 PM, David Leimbach leim...@gmail.com wrote: I got hired at a company because one of the interviewers was impressed that I taught myself Haskell. I basically never use it at work, but I did in my old job. Dave On Thu, Jun 2, 2011 at 2:23 PM, Michael Litchard mich...@schmong.org wrote: I disagree. I'm by no means proficient in Haskell. And, I never bothered learning PHP. I will when I need to. PHP programmers are a dime a dozen. It's been my experience that Haskell is a tool one may use to distinguish oneself from the hoi-poloi. This is important when you live in an area where the baker down the street has a CS degree. On Thu, Jun 2, 2011 at 11:52 AM, Yves Parès limestr...@gmail.com wrote: Learning Haskell will pay off much less than learning PHP, if your goal is to find a job. Amen. I cannot agree with this for practical reasons. I'm using Haskell for real world commercial applications, and I'm very productive with it. I wish so much I could say that... Out of curiosity, what are you using Haskell for? 2011/6/2 Ertugrul Soeylemez e...@ertes.de Alberto G. Corona agocor...@gmail.com wrote: Haskell is an academic asset as well as a fun asset. I cannot agree with this for practical reasons. I'm using Haskell for real world commercial applications, and I'm very productive with it. There is however a variation of this statement, with which I could agree, namely: Learning Haskell will pay off much less than learning PHP, if your goal is to find a job. It takes a lot longer and there are a lot less companies in need of Haskell programmers. Greets, Ertugrul -- nightmare = unsafePerformIO (getWrongWife = sex) http://ertes.de/ ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Attoparsec concatenating combinator
On Thu, Jun 2, 2011 at 10:02 AM, Yitzchak Gale g...@sefer.org wrote: I often find while using attoparsec and attoparsec-text that I need to match a number of text parsers consecutively and concatenate the result. By text parser I mean Parser ByteString for attoparsec and Parser Text for attoparsec-text. I don't know if this helps, but the incremental-parser library has exactly the combinator you're looking for. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] Iteratees again (Was: How on Earth Do You Reason about Space?)
On 6/2/11 8:59 AM, Aleksandar Dimitrov wrote: Hi Ketil, By the way, what is the advantage of using iteratees here? For my testing, I just used: My initial move to iteratees was more a clutch call I made when I was still using bytestring-trie, and was having immense memory consumption problems. bytestring-trie also (intentionally) uses ByteString slicing in order to minimize copying. It does so semi-intelligently--- ensuring that of the two sharing options it chooses the one with a shorter spur. About half the time that spur will be used by continuing down the trie, but you can still end up with unwanted overhead especially if you're intermittently removing keys from the trie. I've been meaning to add functions to remove spurs (because sometimes the memory is more important than the running time) and meaning to add various other upgrades. I've just been too busy with other code for the last year or so. -- Live well, ~wren ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
