Thanks, Bulat. I'm looking forward to trying it out this weekend. Is there any indication what fast IO approach might work its way into the standard libraries? It would be nice for idiomatic Haskell to be really fast by default, and I'd love to be able to show off the language shootout implications to coworkers. Cabal doesn't seem obvious to me. Currently my uninformed point of view is that the whole thing is too complicated to be worth the benefit it provides. Since it generally seems popular, I'm guessing that means it's either much simpler than it seems, or else the benefit provided is just enormous. Can anyone point me to a sales pitch, or provide some motivation for me to RTFM? Thanks, Chad On Sat, 2006-05-20 at 10:17 +0400, Bulat Ziganshin wrote:

Hello Chad,

Friday, May 19, 2006, 10:40:56 PM, you wrote:

...

It sounds like Bulat has gotten some impressive I/O speedups with his Streams library. I'd like to try this out, but I'm having some trouble installing it. I'm using GHC on Linux.

yes, and current (still unpublished) version is even better than previous one

...

My first attempt was looking around on this page: http://www.haskell.org/haskellwiki/Library/Streams

...

There's a really nice description, but no signs of where to actually get the library.

the old hawiki don't allowed to publish links to site narod.ru, because it's public site used also by spammers. i plan to request some place on the haskell.org site for my projects to make possible links publishing

...

From here I was able to download it, but there's no information regarding how this needs to be set up. There are directories "Data", "Examples", and "System", which I assume are supposed to be plugged into the hierarchical module structure, but how do I do that? I thought this might have something to do with Cabal (I've not yet used that), but the Cabal manual talks about a .cabal file, which doesn't exist here.

...

Does this follow some standard approach that I'm not familiar with? Where should I look to learn more?

you will be laughing but i still not learned how cabal should be used and just pack all the library modules together. you can either copy Data & System dirs to the place where your program lies or use "-i" ghc switch to point compiler where it can find additional modules:

ghc --make -i/Haskell/StreamsLib YourProgram.hs

in this case, ghc will search, for example, module "Data.Ref" in file /Haskell/StreamsLib/Data/Ref.hs (after looking to ./Data/Ref.hs)

directory Examples contains examples of using library, you should copy files from this dir to root library directory in order to compile them

On 2006-05-20 at 12:00+0200 "Sebastian Sylvan" wrote:

A quick sales pitch: usually you, the library user, can just type:

./runhaskell Setup.hs configure ./runhaskell Setup.hs build ./runhaskell Setup.hs install

And it will Do The Right Thing(TM), which is nice.

This is something I've never understood about the current fashion. Make allows one to set up rules about what depends on what, so why can't we just arrange it so that someone who wants to install the thing just hast to type ./runhaskell Setup.hs install ? I'm aware that part of this process might require root privileges, and that it would be bad to do the rest of it as root, but there are ways around that, surely? Even if not, the first line ought to be unnecessary, since configure produces something that build depends upon. Jón -- Jón Fairbairn Jon.Fairbairn at cl.cam.ac.uk

Hi

FWIW, it's almost identical to the incantation necessary for projects based on autotools, where it usually reads something like:

./configure make make install

But sadly compared to Windows tools, where its just a case of "double click" this sounds a bit too complex. I am hoping that I can register Cabal with Windows in some way to automate this down to "double click" once more. Thanks Neil

On 5/20/06, Brian Hulley <brianh@metamilk.com> wrote:

It's strange that with the most advanced programming language in existence we're still unable to automate the task of having to manually type a sequence of 3 commands at an ancient UNIX command prompt!!! :-)

No, it's not strange. What appears to have been overlooked is that different people will execute these commands differently: ./configure --prefix=/opt/freeware --sysconfdir=/etc --disable-nls make make install DESTDIR=/home/taral/staging Many people will run "make install" as root but the main compile as non-root. You can't blindly mix all that together. -- Taral <taralx@gmail.com> "You can't prove anything." -- Gödel's Incompetence Theorem

On 2006-05-20 at 11:58EDT Robert Dockins wrote:

On Saturday 20 May 2006 06:53 am, Jon Fairbairn wrote:

...

Make allows one to set up rules about what depends on what, so why can't we just arrange it so that someone who wants to install the thing just hast to type

./runhaskell Setup.hs install

FWIW, it's almost identical to the incantation necessary for projects based on autotools, where it usually reads something like:

./configure make make install

I know. That's the fashion I'm complaining about! I think we shouldn't follow it if we can possibly avoid it. -- Jón Fairbairn Jon.Fairbairn at cl.cam.ac.uk

As a haskell beginner, I saw the lack of a simple and fast standard library as THE major dificulty. Yes, more than monads and lazy programming and all other stuff. This happened because I simply could not create simple programs to simple task as I needed then in production work. I work at forensic area, and have to deal with large ammounts of data (HD images for example). The processing time is important to me and so is the ability to accept files of any size. If an easy fast IO package become part of standard libraries, I am sure will make my life and of more beginners a lot easier. Atila Duncan Coutts wrote:

On Sat, 2006-05-20 at 00:00 -0700, Chad Scherrer wrote:

...

Thanks, Bulat. I'm looking forward to trying it out this weekend.

Is there any indication what fast IO approach might work its way into the standard libraries? It would be nice for idiomatic Haskell to be really fast by default, and I'd love to be able to show off the language shootout implications to coworkers.

Personally I rather like the idea of a new IO library based on lazy byte strings or in this context it'd be more helpful to call them lazy byte streams. I believe this could give us a system which is much more idiomatic Haskell than an explicit buffer/IO based system and with almost as good performance (by almost I mean at most 10%, not 50%).

The idea is that instead of pushing things like string encoding or compression into the IO part of the system that we can leave it in the pure part and get nicer code and still get good performance. That is, instead of layering encoding etc into a stream that we can write things like:

print . count . lines . decode utf8 . gunzip =<< readFile "foo"

and have that run within 10% of the speed of C code.

Donald Stewart has been posting benchmarks to the libraries mailing list showing that the Data.ByteString and Data.ByteString.Lazy modules can give us near-C speed and constant memory use on huge files while still using idiomatic Haskell code.

So instead of layering streams I want to compose functions.

...

Cabal doesn't seem obvious to me. Currently my uninformed point of view is that the whole thing is too complicated to be worth the benefit it provides. Since it generally seems popular, I'm guessing that means it's either much simpler than it seems, or else the benefit provided is just enormous. Can anyone point me to a sales pitch, or provide some motivation for me to RTFM?

The general idea is that it provides a default build system for Haskell libraries and programs. Otherwise people use various build systems which are usually broken in various ways for corner cases that their authors don't encounter.

As someone who packages Haskell progs/libs for Gentoo I can assure you that the shift to Cabal has been a great improvement. Previously, packages needed a great deal of manual work and often patching to package them adequately. That's not to say that Cabal is perfect, but as we improve it everyone benefits which was not previously the case when everyone rolled their own build system. For example we still don't support win32 very well but we should be able to improve that by improving Cabal (or layering tools on top of Cabal) where as previously you'd have to go replacing every make-based build system to get it to work on win32.

Duncan

_______________________________________________ Haskell mailing list Haskell@haskell.org http://www.haskell.org/mailman/listinfo/haskell

____________________________________________________

On Wed, May 24, 2006 at 10:51:31AM +0100, Simon Marlow wrote:

Now you could make purely functional code raise I/O exceptions, but that gives rise to a few problems: imprecise exceptions are difficult to program with (need deepSeq etc.), and they aren't widely implemented (Hugs doesn't have them, JHC doesn't have them and may not get them IIUC).

Hugs has had imprecise exceptions (but not asynchronous ones) since Nov 2003.

Hello Ross, Wednesday, May 24, 2006, 4:50:39 PM, you wrote:

...

Now you could make purely functional code raise I/O exceptions, but that gives rise to a few problems: imprecise exceptions are difficult to program with (need deepSeq etc.), and they aren't widely implemented (Hugs doesn't have them, JHC doesn't have them and may not get them IIUC).

Hugs has had imprecise exceptions (but not asynchronous ones) since Nov 2003.

can you please explain or give a pointer - what is an imprecise exceptions? i thought that it the synonym for async ones -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com

Hello Chad, Saturday, May 20, 2006, 11:00:48 AM, you wrote:

Is there any indication what fast IO approach might work its way into the standard libraries? It would be nice for idiomatic Haskell to be really fast by default, and I'd love to be able to show off the language shootout implications to coworkers.

first, about the speed. i've just implemented ByteString I/O and it shows speed of 20-100 mb/s on my box. so, as i've previously said, the speed of text I/O using ByteStrings+Streams will be comparable to disk speed itself. that means that my "quest for text I/O speed" now finished :) second - i developed Streams to be usable both as add-on lib for existing hugs/ghc versions and as future part of base libraries what will replace current Handle-based I/O. in particular, i will include in new version module what emulates System.IO types and functions. afaik, Simon Marlow supports my plans. but the current version still lacks some functionality that System.IO provides: network I/O via sockets overlapping I/O in multi-threaded programs non-blockable I/O correctness with asynchronous interrupts and 'trace (putStr ..)' LineBuffering/NoBuffering large number of functions, such as hUngetChar the last three things just needs my attention, i know how to do it. the first three things, though, are from areas where i don't have any experience. so i think that i now should publish library in it's current state and interest other developers to work on it. as you understand, we can't say about replacing System.IO without full implementation of it's features and it's no good to have two I/O libraries in the base package :) and third that i want to say is what Streams isn't only "fast I/O library", it's general purpose I/O library that includes support for binary serialization, Char encoding, I/O to memory buffers and strings, memory-mapped files and so on, so on. i think that its main selling points is modularity and extensibility, that greatly simplifies addition of new features, maintenance and improves code readability. i believe that everyone is able to understand library structure and add new features what one needs or replace old modules with better new ones in February i wrote to Andrew Pimlott about this library goals: Wednesday, February 08, 2006, 8:24:59 PM, you wrote:

...

AP> Bulat, it wouldn't hurt to include a motivation section at the top. As AP> I understand, it's ultimately all about speed, right? Otherwise, we AP> would all be happy with lists (and unsafeInterleave*). So maybe a AP> comparison between Stream and [] should be given.

you guessed wrong :) this library is ultimately about replacing System.IO library (i.e. Handles)

AP> Let me rephrase my question: Why not just reimplement the Handles API AP> (with some extensions like binary IO)? Is there really a need to use a AP> handle-like API for more than real IO? If so, what is the need, AP> expressivity or performance (or both)? Maybe a use case showing what AP> you can do with your library, and how you would have to do it otherwise? now i understood you. actually, my presentation was meant only for a few people who are already know about limitations of current library, who are already requested additional features but don't got it. here i need to give some history: when a System.IO interface was developed, it implements much less features than now, and its implementation was enough simple and straightforward. as time goes, the more and more features was added to this library: complex buffering scheme, several async i/o implementations, locking, networking. And at current moment, GHC's System.IO implementation consists of about 3000 lines with a rather monolithic structure. you can't easily add new feature or make using of some "heavyweight" feature optional because it needs to make changes through the entire library. As the result, GHC users can't get improvements in this library, required for his work. Some of them even develop his own libraries what implements just that they need. for example, Einar Karttunen developed networking library with advanced async i/o and support for i/o of fast packed strings. But such solutions is not universal - his library can't be used for file i/o, fo example, although the code is essentially the same. what i done? the main merit of Streams library is not implementation of any particular feature, but birth of framework for the I/O sub-libraries. and my library essentially is just a collection of such sublibs. first, for example, implements file i/o, second implements buffering, third - utf-8 encoding, and so on. the most important property of all these sublibs is that no one of them is greater than 300 lines. that means that it is far easier to understand, modify or even replace any of them. and that will have no impact to other part of library because all these sublibs binded together not via data fields, but with well defined interfaces now, implementing any new I/O feature or new I/O source means only implementing Stream class-comforming interface - all other features, including locking, buffering, encoding, compression, serialization, binary and text i/o, async i/o, will become available automatically. the same for transformers - once implemented gzip compression or UTF-16 encoding support will become automatically available for all the I/O sources, present and future. is not that great? :) moreover, user apllication or third party lib can easily implement new stream types or transformers without bothering the original library. so, this lib in some aspect is meta-meta-instrument, whose capital will be automatically increasing as new sublibs will appear. just at this moment its advantages over System.IO is in the following areas: faster i/o support for optional utf-8 encoding binary i/o and serialization user-controlled locking if you will look inside the archive, you will find directory Examples, which demonstrates usage of all these features. as i said in docs, i also plan to implement other user requests to the System.IO library another consequence of emerging this library is that all these features will become available on Hugs and other Haskell compilers, that never had enough man resources to develop such great library as GHC's System.IO. and about using streams in monads other than IO. i really don't know, whether it will be used or not. at least, for seriazliation it looks promising. for example, there is functions "encode" and "decode" that is like show/read pair, but implements binary encoding according to the instances of Binary class. and of course, it is implemented through the StringBuffer instance of Stream class, working in the ST monad. comparing to the Handles, library provides essentially the same interface. again, you can find information about swithcing from Handles to Streams in doc. i plan to provide in future "legacy layer" which will emulate 100% of System.IO interface, but use the streams internally. It will be essential for old apps, especially if Streams will become official and sole GHC i/o library about internal organization - Streams is somewhat like that if Simon himself splitted up Handles library to the small independent parts and then replaced part of them with simpler/faster implementations. nothing more, except for common Stream class interface, developed by John Goerzen. my work was mainly to bring the best ideas together :) -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com