ocaml-lib-devel — Developer's mailing list

[Ocaml-lib-devel] No takewhile and dropwhile functions yet

[Richard J. <ri...@an...>]

I don't know whether I'm using an old version of extlib, but my
version doesn't include these vital functions:

let rec takewhile f =3D function
    [] -> []
  | x :: xs when f x -> x :: takewhile f xs
  | _ -> []

let rec dropwhile f =3D function
    [] -> []
  | x :: xs when f x -> dropwhile f xs
  | xs -> xs

Rich.

--=20
Richard Jones. http://www.annexia.org/ http://www.j-london.com/
Merjis Ltd. http://www.merjis.com/ - improving website return on investment
If I have not seen as far as others, it is because I have been
standing in the footprints of giants.  -- from Usenet

Re: [Ocaml-lib-devel] new version?

[Nicolas C. <war...@fr...>]

> So... we've got optParse, Dllist, idxtree (if Brian
> will ever commit it :) ), Fifo (again, commit?) new
> since the 1.2 release. You think we should wrap this
> up, commit the outstanding modules, and roll another
> release?

This might be a good idea to Release now.
I'll try to make it when I have spare time.

Nicolas

[Ocaml-lib-devel] new version?

[Jesse G. <je...@wi...>]

So... we've got optParse, Dllist, idxtree (if Brian
will ever commit it :) ), Fifo (again, commit?) new
since the 1.2 release. You think we should wrap this
up, commit the outstanding modules, and roll another
release?

-- 
Jesse Guardiani, Systems Administrator
WingNET Internet Services,
P.O. Box 2605 // Cleveland, TN 37320-2605
423-559-LINK (v)  423-559-5145 (f)
http://www.wingnet.net

Re: [Ocaml-lib-devel] Re: Re: optParse doesn't compile under 3.07 (Buffer.nth)

[Bardur A. <oca...@sc...>]

On Mon, Sep 20, 2004 at 09:59:08AM -0400, Jesse Guardiani wrote:

> Bardur Arantsson wrote:
> 
> > On Mon, Sep 20, 2004 at 09:33:12AM -0400, Jesse Guardiani wrote:
> > 
> >> Bardur Arantsson wrote:
> >> 
> >> > On Mon, Sep 20, 2004 at 12:27:14AM -0400, Jesse D. Guardiani wrote:
> >> > 
> >> >> Buffer.nth isn't included in 3.07's stdlib. Not sure if there are more
> >> >> problems...
> >> > 
> >> > Could you please try compiling with the the attached
> >> > patch? It corrects the 1 use of Buffer.nth to just use
> >> > Buffer.sub...
> >> 
> >> No, sorry. That one is new too. You can view the changes made from 3.07
> >> to 3.08 here:
> > 
> > Dang. Alright, here's one which uses Buffer.contents
> > instead...
> 
> That works.
> 

Committed.

-- 
Bardur Arantsson
<ba...@im...>
<ba...@sc...>

- Do you wanna stay?
- Yeah, why not? All that awaits me at home is a masturbating
Welshman.
                                  Anna and William | Notting Hill

[Ocaml-lib-devel] Re: Re: optParse doesn't compile under 3.07 (Buffer.nth)

[Jesse G. <je...@wi...>]

Bardur Arantsson wrote:

> On Mon, Sep 20, 2004 at 09:33:12AM -0400, Jesse Guardiani wrote:
> 
>> Bardur Arantsson wrote:
>> 
>> > On Mon, Sep 20, 2004 at 12:27:14AM -0400, Jesse D. Guardiani wrote:
>> > 
>> >> Buffer.nth isn't included in 3.07's stdlib. Not sure if there are more
>> >> problems...
>> > 
>> > Could you please try compiling with the the attached
>> > patch? It corrects the 1 use of Buffer.nth to just use
>> > Buffer.sub...
>> 
>> No, sorry. That one is new too. You can view the changes made from 3.07
>> to 3.08 here:
> 
> Dang. Alright, here's one which uses Buffer.contents
> instead...

That works.


-- 
Jesse Guardiani, Systems Administrator
WingNET Internet Services,
P.O. Box 2605 // Cleveland, TN 37320-2605
423-559-LINK (v)  423-559-5145 (f)
http://www.wingnet.net

Re: [Ocaml-lib-devel] Re: optParse doesn't compile under 3.07 (Buffer.nth)

[Bardur A. <oca...@sc...>]

On Mon, Sep 20, 2004 at 09:33:12AM -0400, Jesse Guardiani wrote:

> Bardur Arantsson wrote:
> 
> > On Mon, Sep 20, 2004 at 12:27:14AM -0400, Jesse D. Guardiani wrote:
> > 
> >> Buffer.nth isn't included in 3.07's stdlib. Not sure if there are more
> >> problems...
> > 
> > Could you please try compiling with the the attached
> > patch? It corrects the 1 use of Buffer.nth to just use
> > Buffer.sub...
> 
> No, sorry. That one is new too. You can view the changes made from 3.07 to
> 3.08 here:

Dang. Alright, here's one which uses Buffer.contents
instead...

-- 
Bardur Arantsson
<ba...@im...>
<ba...@sc...>

- Your ideas are intriguing to me and I wish to subscribe to your
newsletter.
                                     Homer Simpson | The Simpsons

[Ocaml-lib-devel] Re: optParse doesn't compile under 3.07 (Buffer.nth)

[Jesse G. <je...@wi...>]

Bardur Arantsson wrote:

> On Mon, Sep 20, 2004 at 12:27:14AM -0400, Jesse D. Guardiani wrote:
> 
>> Buffer.nth isn't included in 3.07's stdlib. Not sure if there are more
>> problems...
> 
> Could you please try compiling with the the attached
> patch? It corrects the 1 use of Buffer.nth to just use
> Buffer.sub...

No, sorry. That one is new too. You can view the changes made from 3.07 to
3.08 here:

http://wwwtcs.inf.tu-dresden.de/~tews/htmlman-3.08/libref/Buffer.html

-- 
Jesse Guardiani, Systems Administrator
WingNET Internet Services,
P.O. Box 2605 // Cleveland, TN 37320-2605
423-559-LINK (v)  423-559-5145 (f)
http://www.wingnet.net

Re: [Ocaml-lib-devel] optParse doesn't compile under 3.07 (Buffer.nth)

[Bardur A. <oca...@sc...>]

On Mon, Sep 20, 2004 at 12:27:14AM -0400, Jesse D. Guardiani wrote:

> Buffer.nth isn't included in 3.07's stdlib. Not sure if there are more 
> problems...

Could you please try compiling with the the attached
patch? It corrects the 1 use of Buffer.nth to just use
Buffer.sub...

-- 
Bardur Arantsson
<ba...@im...>
<ba...@sc...>

If a tree falls in the forest, and kills a mime, does anyone
care?
                                                      Gary Larson

[Ocaml-lib-devel] Re: [Caml-list] Re: Set and Map question

[skaller <sk...@us...>]

On Mon, 2004-09-20 at 14:18, Seth Fogarty wrote:
> Yes, sorry, sorted. I know it is possible, I was wondering if it was
> implemented.

I'm not aware of it, but it seems like a very useful
function to have. 

I would guess there would be some interest
for the extlib project (I added a CC to that list)

-- 
John Skaller, mailto:sk...@us...
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net

Re: [Ocaml-lib-devel] Dllist committed (I think)

[Bardur A. <oca...@sc...>]

On Mon, Sep 20, 2004 at 01:00:40AM -0400, Jesse D. Guardiani wrote:

> BTW, is there lag between a commit and when the changes show up in the
> web interface and anonymous checkout?

Yes, there seems to be a lag of about 4 hours (at least
that's what I remember it as -- could be longer).

> Or did I do it wrong?

Nope, don't worry. It'll show up... eventually. :)

-- 
Bardur Arantsson
<ba...@im...>
<ba...@sc...>

- If I seem drunk to you, it might be because I've been drinking.
                                                               Me

Re: [Ocaml-lib-devel] Dllist committed (I think)

[skaller <sk...@us...>]

On Mon, 2004-09-20 at 15:00, Jesse D. Guardiani wrote:

> BTW, is there lag between a commit and when the changes show up in the
> web interface and anonymous checkout? Or did I do it wrong?

Some time ago there was a HUGE lag -- 8-24 hours.
I believe the have implemented a better system
with no lag, and are progressively moving the CVS 
repositories across to it.

Your report tends to indicate this process is not
yet complete :(

-- 
John Skaller, mailto:sk...@us...
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net

[Ocaml-lib-devel] Dllist committed (I think)

[Jesse D. G. <je...@wi...>]

This is my first time using CVS as a developer, so bare with me.
I think I successfully committed dllist.ml and dllist.mli, along with
changes to Makefile and install.ml.

Before the commit I added the "add" function (append with unit return 
value),
changes my spaces to tabs, and removed all ";;"s.

However, I haven't touched the enum functions. I hope to look at these more
later.

BTW, is there lag between a commit and when the changes show up in the
web interface and anonymous checkout? Or did I do it wrong?


--
Jesse Guardiani, Systems Administrator
WingNET Internet Services,
P.O. Box 2605 // Cleveland, TN 37320-2605
423-559-LINK (v)  423-559-5145 (f)
http://www.wingnet.net

[Ocaml-lib-devel] optParse doesn't compile under 3.07 (Buffer.nth)

[Jesse D. G. <je...@wi...>]

Buffer.nth isn't included in 3.07's stdlib. Not sure if there are more 
problems...
We're still supporting 3.07 at this point, right?

--
Jesse Guardiani, Systems Administrator
WingNET Internet Services,
P.O. Box 2605 // Cleveland, TN 37320-2605
423-559-LINK (v)  423-559-5145 (f)
http://www.wingnet.net

[Ocaml-lib-devel] =?koi8-r?B?98naz9fB0SDwz8TExdLWy8Eh?=

[<ty...@ho...>]

ЖИTЬ B ЧEXИИ !!!



-Пpиглaшeния в Чexию нa oтдыx и paбoтy

-Пoлyчeниe BHЖ и ПMЖ

-Peгиcтpaция фиpm нa тeppитopии Чexии

-Пpиoбpeтeниe и apeндa нeдвижиmocти

a тak жe тpyдoycтpoйcтвo в Чexии!!! Mы тpyдoycтpoиm kaждoгo!!!



Bвecь юpидичeckий cnekтp ycлyг для вac!



Пишитe нam: pra...@ho...



Пpиглaшaem k coтpyдничecтвy вce юpидичeckиe и иmmигpaциoнныe aгeнтcтвa,

a тak жe изичeckиe лицa, koтopыe жeлaют зaнятьcя дaнныm бизнecom и paзвить

eгo y ceбя в peгиoнe.







Your mail is oca...@li...

Your  fax 



1w get ready

[Ocaml-lib-devel] regexps

[skaller <sk...@us...>]

I'm currently unable to get the regexp 'marks' to 
stick except in trailing position. This is enough
for tokenisation but not group extraction.

In principle, marks label epsilon arcs rather than nodes of the 
underlying NFA. The transformation to dfa must respect
these labels.

If anyone knows how to do this, please let me know.
The current algorithm bypasses NFA construction.
[See Dragon book]

-- 
John Skaller, mailto:sk...@us...
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net

Re: [Ocaml-lib-devel] regexps

[Martin J. <mar...@em...>]

On Fri, 10 Sep 2004, Richard Jones wrote:

> Not really sure I understand it from a user's point of view.
>
> For me what really matters is that I could write something like:
>
>   if string =~ /(a+)(b*)/ then ...
>
> The regular expression (Perl-compatible, please!) ought to be compiled
> and checked at compile time, and the compiler should handle caching
> compiled expressions containing variables.
>
> Actually, the example syntax above isn't really great.  I'd need some
> way to bind captures.

You can have a look at the solution I have implemented recently:

  http://martin.jambon.free.fr/micmatch.html

The syntax is based on the syntax which is already in use in ocamllex.
This is just a syntax: it currently produces bindings to the Str library,
but it could use OCaml-PCRE or any other library.
It is currently useful only for matching (from the first character), but
it is integrated in the regular match ... with construct.
I have projects (but not too much time these days) for the following:

- bindings to OCaml-PCRE (including lazy quantifiers and possessive
  groups, and possibly assertions - all of this is not available in Str)
- additional macros (REPLACE, SEARCH, and more)

Good points:
- early error detection (compilation-time checks)
- syntax highlighting (since it uses an OCaml friendly syntax)
- group binding and referencing is readable and safe
- macros such as RE digit = ['0'-'9'] and RE num = digit+   can be defined
  and used

Bad points:
- complex compilation commands
- it does not follow the Perl syntax (is it a real problem?)


Cheers,

Martin

[Ocaml-lib-devel] Re: [Caml-list] Gripes with array

[Brian H. <bh...@sp...>]

The long conservation about the limits of arrays in Ocaml cause a though 
in me: the current dynarray implementation in extlib has exactly the same 
limit- worse than that, it's more likely to hit it (do to array 
expansion).  The solution is to change the internal data structure from an 
array to an array of arrays.

I don't have time to do this at the moment- but if somebody else wants to 
improve extlib, this would be a great (not too hard) project.  If not, 
I'll try to get around to it eventually.

-- 
"Usenet is like a herd of performing elephants with diarrhea -- massive,
difficult to redirect, awe-inspiring, entertaining, and a source of
mind-boggling amounts of excrement when you least expect it."
                                - Gene Spafford 
Brian

Re: [Ocaml-lib-devel] Indexed tree module

[skaller <sk...@us...>]

On Sun, 2004-09-12 at 10:30, Brian Hurt wrote:
> On 12 Sep 2004, skaller wrote:
> > > 
> > > Actually, (first, last) would be easier to implement.  Is there a 
> > > preference in the group?
> > 
> > First,len has the advantage that the last node must

> Um, first and last are both indexes.  Ints.  I think you're still thinking
> of dllist.

Duh .. I was sorry :)

-- 
John Skaller, mailto:sk...@us...
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net

Re: [Ocaml-lib-devel] Indexed tree module

[Brian H. <bh...@sp...>]

On 12 Sep 2004, skaller wrote:

> On Sun, 2004-09-12 at 01:23, Brian Hurt wrote:
> > On Sat, 11 Sep 2004, Bardur Arantsson wrote:
> > 
> > > I would have suggested "slice" just to be consistent with
> > > ExtString.String.slice, but since it uses the (index,len)
> > > approach instead of (first,last) and is also much stricter
> > > it might be confusing to use the same name.
> > 
> > Actually, (first, last) would be easier to implement.  Is there a 
> > preference in the group?
> 
> First,len has the advantage that the last node must
> be in the same list. However, first,last is probably
> more useful in some circumstances -- and works even
> when the two nodes are in a distinct list. 

Um, first and last are both indexes.  Ints.  I think you're still thinking
of dllist.

-- 
"Usenet is like a herd of performing elephants with diarrhea -- massive,
difficult to redirect, awe-inspiring, entertaining, and a source of
mind-boggling amounts of excrement when you least expect it."
                                - Gene Spafford 
Brian

[Ocaml-lib-devel] Re: Re: Dllist ? IndexTree ?

[Jesse G. <je...@wi...>]

Nicolas Cannasse wrote:

>> On the subject of Dllist.enum... I noticed that the count ()
>> function always computes the length of the remaining nodes
>> by traversing the list. It would be more efficient to do
>> this, I think:
>> 
>> 1) The first time count () is called, compute the length
>> and store it "in the enum".
>> 
>> 2) Subsequent calls to count () just return the computed
>> length.
>> 
>> 2) Every time next () is called, decrese the number of
>> remaining elements by 1 (if the number hasn't been set,
>> you don't need to do anything).
>> 
>> This would only incur minimal overhead for next () and for
>> any code the uses count () it basically reduces count ()
>> to O(1) averaged over all the elements.
> 
> This is actually the implementation of List.enum.
> Jesse you can have a look at it to implement it.

OK. I'll see what I can do.

-- 
Jesse Guardiani, Systems Administrator
WingNET Internet Services,
P.O. Box 2605 // Cleveland, TN 37320-2605
423-559-LINK (v)  423-559-5145 (f)
http://www.wingnet.net

[Ocaml-lib-devel] Re: Re: Dllist ? IndexTree ?

[Jesse G. <je...@wi...>]

Brian Hurt wrote:

> On Fri, 10 Sep 2004, Jesse Guardiani wrote:
> 
>> Actually, I tried this before I submitted the current implementation and
>> I couldn't get it to work. The problem seemed to be that Dllist has no
>> concept of Empty, so the length function becomes really difficult to
>> implement.
> 
> I moved the concept of "empty" one level up in my implementation.  This is
> because I didn't want to deal with options on the next/prev links.
> 
>> 
>> Brian, do you think it's possible to implement enums without the "valid"
>> test? If everyone thinks it's possible then I'll give it another shot. I
>> just didn't have much luck the first time around.
> 
> Actually, changing things so that x.next == x is the sign of the end of
> the list should make enums easier.
> 
> let rec enum_done = { data = Obj.magic (); next = enum_done;
>                       prev = enum_done };;
> 
> let enum dlst =
>     match dlst.head with

The concept of "head" was removed along with the concept of "Empty"
in the last release.


-- 
Jesse Guardiani, Systems Administrator
WingNET Internet Services,
P.O. Box 2605 // Cleveland, TN 37320-2605
423-559-LINK (v)  423-559-5145 (f)
http://www.wingnet.net

[Ocaml-lib-devel] new regexp stuff

[skaller <sk...@us...>]

Attached mini-project for regexp stuff.
Use:

make
make test

There may be a bug in the reported number of states.
A state minimisation routine would be useful.

This formulation provides better factoring and a simpified
interface for clients, in particular the routine
'recognize' can be used to match a labelled set of
regexps and return the set of labels (try doing THAT in Perl :)

Next version I'll try to get groups working (see reggrp.ml)
and perhaps provide a lexer and searcher routine.

NOTE: I usually provide *.mli files for everything
except mainlines. However, I can't figure out how
to type everything properly -- if you run ocamlc -i src/regsug.ml 
you'll see what I mean.

-- 
John Skaller, mailto:sk...@us...
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net

Re: [Ocaml-lib-devel] Indexed tree module

[Nicolas C. <war...@fr...>]

> > > I would have suggested "slice" just to be consistent with
> > > ExtString.String.slice, but since it uses the (index,len)
> > > approach instead of (first,last) and is also much stricter
> > > it might be confusing to use the same name.
> >
> > Actually, (first, last) would be easier to implement.  Is there a
> > preference in the group?
>
> First,len has the advantage that the last node must
> be in the same list. However, first,last is probably
> more useful in some circumstances -- and works even
> when the two nodes are in a distinct list.
>
> So I can't decide. I think the semantics are
> sufficiently different to provide both. First,len
> can be implement by starting at first, advancing len,
> and then calling first,last. Or even use options
>
>  ?first ?last ?len

Brian, please don't listen to him :)
Since the two are really same, I would prefer the current index, len. You
can simply add index to len if you want to get last.  Slice or splice are
both ok, sub also.

Nicolas

Re: [Ocaml-lib-devel] Indexed tree module

[skaller <sk...@us...>]

On Sun, 2004-09-12 at 01:23, Brian Hurt wrote:
> On Sat, 11 Sep 2004, Bardur Arantsson wrote:
> 
> > I would have suggested "slice" just to be consistent with
> > ExtString.String.slice, but since it uses the (index,len)
> > approach instead of (first,last) and is also much stricter
> > it might be confusing to use the same name.
> 
> Actually, (first, last) would be easier to implement.  Is there a 
> preference in the group?

First,len has the advantage that the last node must
be in the same list. However, first,last is probably
more useful in some circumstances -- and works even
when the two nodes are in a distinct list. 

So I can't decide. I think the semantics are
sufficiently different to provide both. First,len
can be implement by starting at first, advancing len,
and then calling first,last. Or even use options

 ?first ?last ?len

and treat all 3 being provided as requiring an assertion

  first + len = last

to check the implied invariant. Hmmm...

-- 
John Skaller, mailto:sk...@us...
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net

Re: [Ocaml-lib-devel] Regular expressions and groups

[skaller <sk...@us...>]

On Sun, 2004-09-12 at 00:17, Nicolas Cannasse wrote:

> Well I'm sorry I had first some talks privatly with Alain.
> Speaking french is more easy for exposing my idea and getting feedback :)

I understand .. although I'm not sure my 'ramblings' in
English are any more comprehenisble to an English speaker
than your French :00


> I'm thinking of keeping "core-ML" and adding several functionnalities in
> order to get introspection and dynamic loading. From a language point of
> view, it will looks more like CamlLight than OCaml, but a lot of work is
> needed for runtime and librairies. In particular, it will be possible to
> generate, compile, and execute code at runtime , then examine results using
> reflexion, so the language will be more dynamic. The goal is to have a
> language that helps writing other languages but also enable them to use the
> same runtime and interact together, a little like .net but more
> functionnal-oriented.

Ouch! Felix has a related goal, although our technical
approach is different. It's intended that the client
be able to extend it to implement Domain Specific
Sub Languages. Regular expressions are an example
of that, except this DSSL is currently built in
since it is needed for bootstrapping (I need
a parser and term rewriting system too)

My approach here is to use existing C libraries to
allow the client to use the domain specific functionality
easily, and Felix generated wrappers which translate
some DSSL layer to function calls on that library.

Both the C libraries and DSSL library use traditional 
dlopen() style dynamic linkage.

For me, actually obtaining this layer isn't so hard,
at least once it is bootstrapped (which is still some
way off).

The big problem is probably the same as yours --
making it typesafe.

BTW: why don't you use the same sublanguage of Ocaml
that Ocaml is itself written in, and use the same
bootstrapping mechanism? Or is that what you're doing?

-- 
John Skaller, mailto:sk...@us...
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net