Poly/ML

polyml@lists.polyml.org

704 discussions

[RFC]: Modules to support C FFIs -- a proposal and a minor patch
by domagoj.stolfa＠gmail.com 23 Oct '18

23 Oct '18

Hi all: I've only been using Poly for a few days and haven't done too much SML before a project I'm working on now -- but I've noticed a couple of things that seemed a little tedious while implementing some modules as a part of that project. I may have missed something, and if that's the case please do point it out to me. However, in case what I've observed is in fact the current state of the world, I've written a small proposal and a patch to eliminate these little annoyances (suggested to me by DGASAU on freenode). In essence, I've been building an FFI in the past few days and noticed an absence of a number of things. Namely, there seems to be a lack of support for standard C types such as uint16_t, int16_t, int8_t, int64_t and float (not sure about int64_t). This can make building FFIs a bit of a hassle sometimes, as it requires explicit checks of what value the integer/word in SML might contain before deciding to try and write it. An example of such a workaround would be: ??? ??? if (#NOFFS a) > 0w65535 orelse (#NENOFFS a) > 0w65535 ??? ??? then raise RunCall.Conversion "Invalid Word16 constant" ??? ??? else (...) where #NOFFS a and #NENOFFS a are otherwise manipulated as Word32.words. Moreover, this seems to sometimes cause people developing software or extracting executable specifications from proof assistants such as Isabelle that depend on things like Word16 to discard or impose a limitation on SML used with PolyML as it simply does not support these types out of the box [1, 2] Would it make sense to support these types in PolyML, namely through Word16.word, Int16.int, Int8.int and Int64.int, where Word16 would be of WORD signature, while Int16, Int8 and Int64 would be of INTEGER signature. I've attached a sample patch for supporting Word16, very heavily based on David Matthews' work on Word8 in order to demonstrate what a very crude implementation of the proposal would look like. Is this a real problem, and if so, is implementing these things a sensible way forward? Thanks! [1]: http://rohandrape.net/?t=smlsc3&e=README [2]: https://www.isa-afp.org/browser_info/current/AFP/Native_Word/outline.pdf -- Domagoj

2 2

LInkage and the Foreign structure
by mark.clements＠ki.se 17 Oct '18

17 Oct '18

I am enjoying using polyml's Foreign structure. However, in using the GNU Scientific Library (GSL), I have run into a problem with linkage. The GSL library (libgsl.so) calls standard cBLAS matrix functions with unknown symbols to allow the user to link with different cBLAS implementations (NB: GSL provides an unoptimised cBLAS implementation named libgslcblas.so). When I call a GSL function that depends on the cBLAS functions, I get a symbol not defined error. I have tried unsuccessfully (a) loading the libgslcblas library either before or after loading libgsl, (b) changing the loadLibrary implementation to use dlopen with RTLD_LAZY | RTLD_GLOBAL, and (c) compiling the SML code to an object file and then linking against the libraries. How could this be made to work? Sincerely, Mark. PS test code for (a) is: open Foreign; val gsl = loadLibrary "/usr/lib/x86_64-linux-gnu/libgsl.so"; val gslcblas = loadLibrary "/usr/lib/x86_64-linux-gnu/libgslcblas.so"; type gsl_matrix_t = Memory.voidStar; (* for demonstration only *) local val call1 = buildCall2((getSymbol gsl "gsl_matrix_calloc"), (cInt,cInt), cPointer) val call2 = buildCall4((getSymbol gsl "gsl_matrix_set"), (cPointer,cInt,cInt,cDouble), cVoid) val call3 = buildCall3((getSymbol gsl "gsl_linalg_exponential_ss"), (cPointer, cPointer, cDouble), cVoid) in fun gsl_matrix_alloc (size1,size2) = call1(size1,size2) fun gsl_matrix_set (ptr : gsl_matrix_t, i, j, x) = call2(ptr, i, j, x) fun gsl_matrix (arr : real Array2.array) = let val size1 = Array2.nRows arr val size2 = Array2.nCols arr val ptr = gsl_matrix_alloc(size1, size2) val _ = Array2.tabulate Array2.RowMajor (size1, size2, fn (i,j) => gsl_matrix_set(ptr, i, j, Array2.sub(arr, i, j))) in ptr end fun gsl_mexp (arr : real Array2.array) = let val size1 = Array2.nRows arr val size2 = Array2.nCols arr val ptr = gsl_matrix arr val ptr2 = gsl_matrix_alloc(size1, size2) val _ = call3(ptr, ptr2, 0.01) in ptr2 end end; fun main() = let val a = Array2.fromList [[1.0,2.0],[3.0,4.0]]; val ptr = gsl_matrix(a); val m = gsl_mexp(a) (* fails *) in () end; (* symbol lookup error: /usr/lib/x86_64-linux-gnu/libgsl.so: undefined symbol: cblas_dgemm *) For (c): polyc -c test.sml g++ test.o -L/usr/lib/x86_64-linux-gnu -lgsl -lgslcblas `pkg-config --libs polyml` ./a.out # same error

4 12

"Exception- InternalError: asGenReg raised while compiling"
by mark.clements＠ki.se 13 Sep '18

13 Sep '18

Does anyone know why the following will not compile? I am using Poly/ML 5.7.1 Release (Git version v5.7.1-88-g04ba183) on Ubuntu 16.04LTS. fun invDet2(M) = let val a = Array2.sub(M,0,0) val b = Array2.sub(M,0,1) val c = Array2.sub(M,1,0) val d = Array2.sub(M,1,1) val det = (a*d - b*c) val x = 1.0/det val inv = Array2.fromList [[d*x, ~b*x], [~c*x, a*x]] in (inv,det) end; (* see http://www.cs.cmu.edu/afs/cs/project/pscico/pscico/src/matrix/RealMatrix.sml *) Note that the following is okay: fun invDet2(M) = let val a = Array2.sub(M,0,0) val b = Array2.sub(M,0,1) val c = Array2.sub(M,1,0) val d = Array2.sub(M,1,1) val det = (a*d - b*c) val inv = Array2.fromList [[d/det, ~b/det], [~c/det, a/det]] in (inv,det) end; --- Mark

2 3

socket from fd
by Ramana.Kumar＠cl.cam.ac.uk 03 Aug '18

03 Aug '18

Hi Poly/ML, Is it possible to retrieve a socket from a file descriptor in Standard ML or Poly/ML? E.g., if Posix.FileSys.ST.isSock (Posix.FileSys.fstat fd) is true, then what can I do to fd to use functions in the Socket structure on it? Cheers, Ramana

4 3

Error with poly_smlnj-lib.sml
by mark.clements＠ki.se 31 Jul '18

31 Jul '18

Dear all, This is my first post. I am trying to use poly_smlnj-lib.sml (with a plan to see whether mlnffigen can be ported from MLton to Poly/ML). Following instructions from http://www.tbrk.org/software/poly_smlnj-lib.html (see also https://github.com/standardml/smackage/blob/master/src/poly_smlnj-lib.sml), I called: $ poly > use "poly_mlyacc.sml"; (* ok *) > use "poly_smlnj-lib.sml"; with the latter giving an error: /usr/local/lib/mlton/sml/smlnj-lib/Util/bit-array.sml:781: error: Structure (Word8Array) has not been declared in structure Unsafe Found near Unsafe.Word8Array.create (Word8Array.length bits) I am using Poly/ML 5.7.1 Release and MLton 20180207. Any suggestions would be welcomed. Kindly, Mark.

2 2

PolyML and paths for load files.
by kostirya＠gmail.com 05 Jul '18

05 Jul '18

Hello. Let's take into account Env to load files in use function? > diff -u FinalPolyML.sml_orig FinalPolyML.sml --- FinalPolyML.sml_orig 2018-07-05 11:17:32.489398000 +0300 +++ FinalPolyML.sml 2018-07-05 11:20:20.012445000 +0300 @@ -674,6 +674,7 @@ end val suffixes = ref ["", ".ML", ".sml"] + val paths = ref ["POLYML_LIB", "SMACKAGE_HOME", "SMACKAGE"] (*****************************************************************************) @@ -689,15 +690,20 @@ (* We use the functional layer and a reference here rather than TextIO.input1 because that requires locking round every read to make it thread-safe. We know there's only one thread accessing the stream so we don't need it here. *) - fun trySuffixes [] = + fun trySuffixes name [] = (* Not found - attempt to open the original and pass back the exception. *) (TextIO.getInstream(TextIO.openIn originalName), originalName) - | trySuffixes (s::l) = + | trySuffixes name (s::l) = (TextIO.getInstream(TextIO.openIn (originalName ^ s)), originalName ^ s) - handle IO.Io _ => trySuffixes l + handle IO.Io _ => trySuffixes name l + (* First in list is the name with no suffix. *) - val (inStream, fileName) = trySuffixes("" :: ! suffixes) + fun tryPath name [] = trySuffixes name ("" :: ! suffixes) + | tryPath name (p::ps) = trySuffixes (p ^ name) ("" :: ! suffixes) handle IO.Io _ => tryPath name ps + + val (inStream, fileName) = tryPath originalName (rev (foldl (fn (p, r) => case OS.Process.getEnv p of NONE => r | SOME p => p ^ "/"::r) [""] (!paths))) + val stream = ref inStream (* Record the file name. This allows nested calls to "use" to set the correct path. *) >

1 1

Pattern matching compilation is very slow
by kostirya＠gmail.com 26 Jun '18

26 Jun '18

Hello. I found that Poly/ML very very slowly compiles (30 seconds) functions with very large count of pattern matching items (1400 items). I need to replace the pattern matching with many if-then-else. But the pattern matching works a faster for mlton and smlnj. Therefore, I want to use the pattern matching. :-) Also noticed that Poly/ML very slow for it. Even SML/NJ is faster for it. Below is the code for benchmark and result. fun fc 0wx41 f r = f (0wx61, r) | fc 0wx42 f r = f (0wx62, r) | fc 0wx43 f r = f (0wx63, r) ... | fc 0wx1f80 f r = f (0wx3b9, f (0wx1f00, r)) | fc 0wx1f81 f r = f (0wx3b9, f (0wx1f01, r)) | fc 0wx1f82 f r = f (0wx3b9, f (0wx1f02, r)) ... | fc 0wx1e920 f r = f (0wx1e942, r) | fc 0wx1e921 f r = f (0wx1e943, r) | fc w f r = f (w, r) 1400 items fun runBench N name f = let fun loop 0 f = () | loop i f = (f (); loop (i - 1) f) val t0 = Time.now () val _ = loop N f val t1 = Time.now () in print (name ^ " " ^ Real.toString(Time.toReal(Time.-(t1, t0))) ^ "\n") end val _ = runBench 1000000 "latin a" (fn () => (fc 0w97 (op ::) [])) val _ = runBench 1000000 "Sharp s" (fn () => (fc 0wxdf (op ::) [])) val _ = runBench 1000000 "Cyr Yi " (fn () => (fc 0wx407 (op ::) [])) val _ = runBench 1000000 "last " (fn () => (fc 0wx1e921 (op ::) [])) > time poly --script unicode-case.sml ; echo "===" ; time poly --script unicode-if.sml ; echo "===" ; time poly --script unicode-if-opt.sml latin a 4.557194 Sharp s 4.073306 Cyr Yi 2.902266 last 0.015322 47.325u 1.456s 0:48.23 101.1% 11246+419k 0+0io 0pf+0w === latin a 4.584374 Sharp s 0.065816 Cyr Yi 0.290468 last 4.698412 9.848u 0.046s 0:09.88 100.0% 11307+421k 0+0io 0pf+0w === latin a 0.015102 Sharp s 0.024983 Cyr Yi 0.030565 last 0.101628 0.463u 0.030s 0:00.43 113.9% 11499+428k 0+0io 0pf+0w MLton > ./unicode-case ; echo "===" ; ./unicode-if ; echo "===" ; ./unicode-if-opt latin a 0.00119 Sharp s 0.001167 Cyr Yi 0.001176 last 0.001172 === latin a 0.541467 Sharp s 0.024442 Cyr Yi 0.121234 last 0.537045 === latin a 0.013061 Sharp s 0.025635 Cyr Yi 0.154532 last 0.550286 > sml unicode-case.sml Standard ML of New Jersey v110.77 [built: Sun Jun 24 00:25:15 2018] [opening unicode-case.sml] [autoloading] [library $SMLNJ-BASIS/basis.cm is stable] [autoloading done] latin a 0.001608 Sharp s 0.00166 Cyr Yi 0.001546 last 0.001552 val fc = fn : word -> word list -> word list val runBench = fn : int -> string -> (unit -> 'a) -> unit - ^D > sml unicode-if.sml Standard ML of New Jersey v110.77 [built: Sun Jun 24 00:25:15 2018] [opening unicode-if.sml] [autoloading] [library $SMLNJ-BASIS/basis.cm is stable] [autoloading done] latin a 0.550017 Sharp s 0.02571 Cyr Yi 0.121336 last 0.548638 val fc = fn : word -> word list -> word list val runBench = fn : int -> string -> (unit -> 'a) -> unit - ^D > sml unicode-if-opt.sml Standard ML of New Jersey v110.77 [built: Sun Jun 24 00:25:15 2018] [opening unicode-if-opt.sml] [autoloading] [library $SMLNJ-BASIS/basis.cm is stable] [autoloading done] latin a 0.014573 Sharp s 0.027239 Cyr Yi 0.124949 last 0.557338 val fc = fn : word -> word list -> word list val runBench = fn : int -> string -> (unit -> 'a) -> unit - ^D

1 0

RPC in SML
by Ramana.Kumar＠cl.cam.ac.uk 08 Jun '18

08 Jun '18

Dear Poly/ML list, I'm exploring how to make remote procedure calls (RPC) from Standard ML. My understanding is that this is not part of the language or basis library in any form, but that there is FFI support in Poly/ML for it. Is that correct? Is there any documentation for the Poly/ML FFI or examples of its use that someone could point me to? Thanks much, Ramana

2 1

Building on Windows in batch mode -- Visual Studio
by makarius＠sketis.net 08 Mar '18

08 Mar '18

Dear Windows experts, the task is to build the Poly/ML such that: * the build works on the command-line / in batch mode * the resulting poly.exe is a command-line tool, not a Windows desktop application. Based on recommendations by David Matthews from some years ago, I have done that so far with MinGW -- using a rather old version of gcc. See also: http://isabelle.in.tum.de/repos/isabelle/file/839de121665c/Admin/polyml/INS… http://isabelle.in.tum.de/repos/isabelle/file/839de121665c/Admin/polyml/REA… http://isabelle.in.tum.de/repos/isabelle/file/839de121665c/src/Pure/Admin/b… (with specific options for x86-windows and x86_64-windows) I wonder if it would be better to use the free community version of Visual Studio instead: https://www.visualstudio.com/downloads Has anybody tried this and can report some experience with it? Makarius

3 2

fancy tty tricks (with curses?)
by Michael.Norrish＠data61.csiro.au 27 Feb '18

27 Feb '18

Has anyone used the FFI interface or similar to link libcurses or similar? I could try to talk ANSI terminal sequences directly, but querying terminal state with these seems pretty painful. (At the very least one has to put standard input into a non-blocking state?, and I found the Basis documentation on how I might do this very confusing.) And then one is breaking the very abstraction that curses is trying to provide... As a bare minimum for example, I'd like to get the width of the current terminal. Michael

2 1

← Newer
1
...
8
9
10
11
12
13
14
...
71
Older →

Jump to page:

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