Poly/ML

polyml@lists.polyml.org

2 participants
711 discussions

A type checking bug in Poly/ML
by vesa.a.j.k＠gmail.com 08 Nov '07

08 Nov '07

While attempting to compile my random value generator library with Poly/ML, I encountered a type checking bug in Poly/ML (compiled from CVS). Steps to reproduce: 1. Get MLton lib revision 6146. 2. Run ./Make.sh in the <mltonlib>/org/mlton/vesak/use-lib/unstable directory. 3. Run poly in the <mltonlib> directory and enter the following two expressions: use "org/mlton/vesak/use-lib/unstable/polyml.use" ; use "com/ssh/random/unstable/lib.use" ; The output [...] Error: in '[...]/mltonlib-6146/com/ssh/random/unstable/detail/mk-random-gen.fun', line 7. Can't match 'a MkRandomGen().cod to 'a [...] indicates a type checking bug in Poly/ML. For convenience, here is a shell script for triggering the bug: svn -r 6146 co svn://mlton.org/mltonlib/trunk mltonlib-6146 cd mltonlib-6146/org/mlton/vesak/use-lib/unstable/ ./Make.sh cd ../../../../.. echo 'use "org/mlton/vesak/use-lib/unstable/polyml.use" ; use "com/ssh/random/unstable/lib.use" ;' | poly Part of the code that triggers the bug are the following type definitions in the file <mltonlib>/com/ssh/random/unstable/detail/mk-random-gen.fun: type 'a dom = Int.t * RNG.t and 'a cod = 'a type 'a t = 'a dom -> 'a cod If the definitions of 'a dom and 'a cod are substituted into 'a t type 'a t = Int.t * RNG.t -> 'a the code compiles with Poly/ML. This shouldn't be necessary however. For the curious, the dom and cod tycons are a leftover from an earlier signature that revealed that the t tycon is an arrow. The same code compiles fine without modification with SML/NJ, MLton, and MLKit. -Vesa Karvonen

1 0

Forthcoming 5.1 release
by David.Matthews＠prolingua.co.uk 08 Nov '07

08 Nov '07

There has been quite a lot of work on Poly/ML over the last year and I'm intending to release version 5.1 shortly. This is an invitation to try the CVS version and let me know if there are any bugs that need to be fixed before the final release. The two major changes are the addition of true multi-threading and the ability to save and load changes without having to build an executable. There have also been changes to the Windows interface structures and other minor changes. The multi-threading changes (thanks to the Verisoft project and TUM) involved a major re-working of the run-time system to allow the code to use kernel threads, in turn allowing multi-threaded programs to make use of multiple cores. A new Thread structure has been written to make effective use of this and the details are at http://www.polyml.org/docs/Threads.html . The old Process structure has been rewritten in terms of this but it will generally be better to use the Thread structure directly. A new substructure, PolyML.SaveState, has been added (thanks to Qinetiq), to allow the state of a session to be saved and then reloaded in another session. This is a lighter-weight version of the PolyML.export and does not require a linking step. Details of the structure can be found at http://www.polyml.org/docs/SaveState.html . As well as simple saved states it is also possible to build chains of saved states in which each saved state contains only the changes from the one above. The current plan is to release version 5.1 towards the end of this month so any feedback would be appreciated as soon as possible. David.

2 2

Int32 hack and patch to Word32.sml
by vesa.a.j.k＠gmail.com 06 Nov '07

06 Nov '07

UG9seS9NTCBkb2VzIG5vdCBjdXJyZW50bHkgc2VlbSB0byBwcm92aWRlIGFueSBmaXhlZCBwcmVj aXNpb24gaW50ZWdlcgptb2R1bGVzLiAgUG9ydGluZyBzb21lIG9mIG15IGNvZGUgdG8gUG9seS9N TCB3b3VsZCBiZSBlYXNpZXIgaWYKUG9seS9NTCB3b3VsZCBwcm92aWRlIGEgZml4ZWQgcHJlY2lz aW9uIGludGVnZXIgbW9kdWxlLCBsaWtlIEludDMyLAphbmQgdGhlIGFzc29jaWF0ZWQgb3Zlcmxv YWRzLiAgSGF2aW5nIGFyYml0cmFyeSBwcmVjaXNpb24gaW50ZWdlcnMgYXMKdGhlIGRlZmF1bHQg SW50IG1vZHVsZSBpcyBub3QgYSBwcm9ibGVtLiAgQXR0YWNoZWQgaXMgYSBxdWljayBoYWNrCnRo YXQgaW1wbGVtZW50cyBJbnQzMiBmb3IgUG9seS9NTC4gIEkgdGVzdGVkIGl0IGJyaWVmbHkgaW50 ZXJhY3RpdmVseQphbmQgaXQgc2VlbXMgdG8gd29yay4gIEkgYWxzbyBub3RpY2VkIHRoYXQgdGhl IG5lZ2F0aW9uIG9wZXJhdG9yLCB+LApoYXMgbm90IGJlZW4gb3ZlcmxvYWRlZCBmb3IgV29yZDMy IGFuZCBhZGRlZCB0aGUgYXBwcm9wcmlhdGUgb3ZlcmxvYWQuCgotVmVzYSBLYXJ2b25lbgotLS0t LS0tLS0tLS0tLSBuZXh0IHBhcnQgLS0tLS0tLS0tLS0tLS0KQSBub24tdGV4dCBhdHRhY2htZW50 IHdhcyBzY3J1YmJlZC4uLgpOYW1lOiB3b3JkMzItaW50MzIucGF0Y2gKVHlwZTogdGV4dC94LWRp ZmYKU2l6ZTogMTg5MiBieXRlcwpEZXNjOiBub3QgYXZhaWxhYmxlClVybCA6IGh0dHA6Ly9saXN0 cy5pbmYuZWQuYWMudWsvbWFpbG1hbi9wcml2YXRlL3BvbHltbC9hdHRhY2htZW50cy8yMDA3MTEw Ni9hODY1MmQ1ZC93b3JkMzItaW50MzIuYmluCi0tLS0tLS0tLS0tLS0tIG5leHQgcGFydCAtLS0t LS0tLS0tLS0tLQpBIG5vbi10ZXh0IGF0dGFjaG1lbnQgd2FzIHNjcnViYmVkLi4uCk5hbWU6IElu dDMyLnNtbApUeXBlOiBhcHBsaWNhdGlvbi9zbWlsClNpemU6IDQwMjAgYnl0ZXMKRGVzYzogbm90 IGF2YWlsYWJsZQpVcmwgOiBodHRwOi8vbGlzdHMuaW5mLmVkLmFjLnVrL21haWxtYW4vcHJpdmF0 ZS9wb2x5bWwvYXR0YWNobWVudHMvMjAwNzExMDYvYTg2NTJkNWQvSW50MzIuc21pCg==

2 2

Patch(es) to Poly/ML's Basis library (REAL, TEXT_IO, CHAR, STRING)
by vesa.a.j.k＠gmail.com 06 Nov '07

06 Nov '07

V2hpbGUgdXBkYXRpbmcgbXkgRXh0ZW5kZWQgQmFzaXMgbGlicmFyeQooaHR0cDovL21sdG9uLm9y Zy9jZ2ktYmluL3ZpZXdzdm4uY2dpL21sdG9ubGliL3RydW5rL2NvbS9zc2gvZXh0ZW5kZWQtYmFz aXMvdW5zdGFibGUvKQp0byBjb21waWxlIHdpdGggUG9seS9NTCwgSSBmaXhlZCBzb21lIGlzc3Vl cyBpbiBQb2x5L01MJ3MgQmFzaXMKbGlicmFyeSBpbXBsZW1lbnRhdGlvbjoKCi0gSSBmaXhlZCB0 aGUgUkVBTCBzaWduYXR1cmUsIHdoaWNoIGRpZCBub3QgY29ycmVzcG9uZCB0byB0aGUgUmVhbApz dHJ1Y3R1cmUuICBJIGFsc28gc2VhbGVkIHRoZSBSZWFsIHN0cnVjdHVyZSB3aXRoIHRoZSBSRUFM IHNpZ25hdHVyZQp0byBhdm9pZCBwcm9ibGVtcyBsaWtlIHRoYXQgaW4gdGhlIGZ1dHVyZS4KCi0g SSBhZGRlZCB0aGUgbWlzc2luZyBURVhUX0lPIHNpZ25hdHVyZS4gIEkgaGF2ZW4ndCBjaGVja2Vk IHdoZXRoZXIgaXQKY29uZm9ybXMgdG8gdGhlIEJhc2lzIGxpYnJhcnkgc3BlYywgYnV0IGF0IGxl YXN0IHRoZXJlIG5vdyBpcyBhClRFWFRfSU8gc2lnbmF0dXJlICh3aGljaCBpcyBuZWVkZWQgYnkg bXkgbGlicmFyeSkuCgotIEkgZml4ZWQgc29tZSBzcGVjaWZpY2F0aW9ucyBpbiB0aGUgQ0hBUiBh bmQgU1RSSU5HIHNpZ25hdHVyZXMgdG8KY29uZm9ybSB0byB0aGUgQmFzaXMgc3BlYywgd2l0aCB0 aGUgZXhjZXB0aW9uIG9mIHRoZSBTVFJJTkcuc2Nhbgp2YWx1ZSwgd2hpY2ggc2VlbXMgdG8gYmUg c3BlY2lmaWVkIGluY29ycmVjdGx5IGluIHRoZSBCYXNpcyBzcGVjIChzZWUKaHR0cDovL21sdG9u Lm9yZy9waXBlcm1haWwvbWx0b24vMjAwNy1NYXJjaC8wMjk3MDAuaHRtbCkuCgpBdHRhY2hlZCBp cyBhIHBhdGNoIG9mIHRoZXNlIGNoYW5nZXMuCgotVmVzYSBLYXJ2b25lbgotLS0tLS0tLS0tLS0t LSBuZXh0IHBhcnQgLS0tLS0tLS0tLS0tLS0KQSBub24tdGV4dCBhdHRhY2htZW50IHdhcyBzY3J1 YmJlZC4uLgpOYW1lOiBiYXNpcy1maXhlcy5wYXRjaApUeXBlOiB0ZXh0L3gtZGlmZgpTaXplOiAx MDgyNCBieXRlcwpEZXNjOiBub3QgYXZhaWxhYmxlClVybCA6IGh0dHA6Ly9saXN0cy5pbmYuZWQu YWMudWsvbWFpbG1hbi9wcml2YXRlL3BvbHltbC9hdHRhY2htZW50cy8yMDA3MTEwNi9kMTY2MjRk Yy9iYXNpcy1maXhlcy5iaW4K

2 1

Wrong type for Real.fromDecimal
by vesa.a.j.k＠gmail.com 06 Nov '07

06 Nov '07

Hi, The type of Real.fromDecimal Poly/ML 5.1 Testing > Real.fromDecimal ; val it = fn : IEEEReal.decimal_approx -> Real.real SML# 0.30 (2007-07-03 16:16:12 JST) # Real.fromDecimal ; val it = fn : IEEEReal.decimal_approx -> Real.Math.real Standard ML of New Jersey v110.66 [built: Thu Oct 25 11:43:30 2007] - Real.fromDecimal ; [autoloading] [library $SMLNJ-BASIS/basis.cm is stable] [autoloading done] val it = fn : IEEEReal.decimal_approx -> real disagrees with the Basis Library specification (http://www.standardml.org/Basis/real.html#SIG:REAL.toDecimal:VAL) val fromDecimal : IEEEReal.decimal_approx -> real option in all of the three compiler's implementation of the Basis Library. Also note that the description in the Basis Library spec includes the equation fromDecimal (toDecimal r) = r which is incorrect, of course. -Vesa Karvonen

2 1

polyml on macbook intel
by cusquinho＠gmail.com 31 Oct '07

31 Oct '07

Hi All, I'm sorry if this is not the proper place to ask, but I'll try to be "short". I've been trying to get Isabelle in my macbook and I've joined their discussion list for support. They told me intel macbooks wont work with the modified polyml_ppc_darwin, and sent me a patch to get it working with Polyml 5.0. I tried both Polyml 5.0 and the one from the top SVN tree, and when I try to compile I get the same error: /bin/sh ../libtool --tag=3DCXX --mode=3Dcompile g++ -DHAVE_CONFIG_H -I. -= I.. -DMACOSX -Wall -I/usr/X11/include -O3 -I/usr/X11/include -MT machoexport.lo -MD -MP -MF .deps/machoexport.Tpo -c -o machoexport.lo machoexport.cpp g++ -DHAVE_CONFIG_H -I. -I.. -DMACOSX -Wall -I/usr/X11/include -O3 -I/usr/X11/include -MT machoexport.lo -MD -MP -MF .deps/machoexport.Tpo -c machoexport.cpp -fno-common -DPIC -o .libs/machoexport.o machoexport.h:49: error: 'uint' has not been declared machoexport.h:49: error: 'uint' has not been declared machoexport.h:49: error: 'uint' has not been declared machoexport.cpp:241: error: 'uint' has not been declared machoexport.cpp:241: error: 'uint' has not been declared machoexport.cpp:241: error: 'uint' has not been declared make[2]: *** [machoexport.lo] Error 1 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2 sh-3.2# Any idea? thanks, Fernando

4 10

Question about possible compiler optimization
by David.Matthews＠prolingua.co.uk 19 Oct '07

19 Oct '07

Lukas Bulwahn wrote: > I was mistaken that this function is tail recursive. > But I was expecting that the function can be translated to the following > C code (using no stack allocations): > > while (p is not Nil) do > if (value of p == l) then p := p->next; > else break; > od > if (p is Nil) then raise Error... > > So, this program should not allocate any stack. That is a different function. It doesn't return the modified list but merely checks that the item is or is not present. The equivalent, tail-recursive function in ML is: fun test l [] = raise Fail "Not present" | test l (x::xs) = if l = x then () else test l xs; Poly/ML detects this as tail-recursive and turns it into a loop. David

4 11

Question about possible compiler optimization
by bulwahn＠in.tum.de 16 Oct '07

16 Oct '07

Hi all, I use the code export of Isabelle and this will generate (poly)ML source code of the following form: ... fun mem l (x :: xs) = (if ((x : IntInf.int) = l) then xs else let val v = mem l xs; in (x :: v) end) | mem l [] = raise (Fail "Cannot find literal"); ... The function (as many others of that form) can easily be transformed into a tail recursive function. Does PolyML figure out that it is possible to make this function tail-recursive (automatically) by "unfolding the let"? Thanks in advance. Lukas Bulwahn

2 1

RE: mlEdit crashes on Windows
by pahia＠t-online.hu 16 Sep '07

16 Sep '07

>It certainly doesn't crash when I do that although it doesn't run >properly. The problem seems to be that the Poly icon is not available >in a stand-alone executable and trying to load it results in a exception >that stops the program. I've fixed this in the CVS version. >Which compiler did you use to build PolyML 5.0 on Windows? >Regards, >David. Thanks for your quick response. I used VC6 and VC7.1 to build PolyML on Windows. I can build it with VC8 also, but it seems to me, that I do something wrong: I wrote a simple rc file which included poly.ico, so certainly the LoadIcon is not a problem for me. I have used the following commands to produce mlEdit.exe. In PolyML: > PolyML.make "mlsource/extra/Win"; > PolyML.use "mlsource/extra/Win/Examples/mlEdit"; > PolyML.export ("mlEdit", mlEdit); At the Windows prompt: > rc mlEdit.rc > cl /FemlEdit.exe mlEdit.obj mlEdit.res PolyMainLib.lib PolyLib.lib libcmt.lib /link /subsystem:windows What I am doing wrong ? Thanks, Andr?s P?hi

2 1

mlEdit crashes on Windows
by pahia＠t-online.hu 12 Sep '07

12 Sep '07

Dear all, I have compiled the mlEdit.sml example, exported to the native object file linked with PolyMainLib and PolyLib and with poly.ico and the resulting exe crashes inside PolyLib.dll somewhere. I am using PolyML 5.0 on Windows 2000. Could you give me any advice, how to get it to work ? Thanks, Andras Pahi

2 1

← Newer
1
...
58
59
60
61
62
63
64
...
72
Older →

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

Poly/ML