Poly/ML

polyml@lists.polyml.org

2 participants
711 discussions

Release 5.2
by David.Matthews＠prolingua.co.uk 31 Jul '08

31 Jul '08

Timothy Bourke wrote: > I've just had some time to test the new version on FreeBSD (6.3-RELEASE). > > On Jun 3 at 15:14 +0100, David Matthews wrote: >> Changes to handling of signals by the Signal structure. Rather than >> forking a new thread to process each signal there is a single thread that >> executes all the signal handlers. This should improve the responsiveness >> to an asynchronous signals such as SIGINT. The original control-C handler >> has been removed and the Poly/ML read-eval-print loop now installs a simple >> handler for control-C that just generates an Interrupt exception. > > I found that the final polyimport -H 10 polytemp.txt kept crashing > with a segmentation fault. It seems that PTHREAD_STACK_MIN is not > enough space on FreeBSD 6.3. Adding 100 (found by trial and error) > fixes the problem as per the attached patch. Tim, Thanks for reporting this. I've installed FreeBSD 6.3 in an emulator and tested Poly/ML. It seems that PTHREAD_STACK_MIN is only 2k on FreeBSD whereas it is 16k on Linux. I've added a test for when it's less than 4k so that there is at least 4k for the stack. The idea was to try and use a smallish stack for this thread because it only calls a single function but there's no need to shave the stack space to an absolute minimum. It's likely to be rounded up to a page boundary anyway. I also fixed a few other things so that it compiles on FreeBSD without any other changes. Regards, David. Copied to the mailing list in case there are other FreeBSD users out there.

2 1

pretty printing oddities
by ldixon＠inf.ed.ac.uk 15 Jul '08

15 Jul '08

A while ago I grumbled about pretty printing of types in PolyML. I have just managed to condense my code into a sensibly viewable example which illustrates what I think the problem is: signature N = sig type T end; signature A = sig type T structure Ns : N sharing type Ns.T = T val mk : T val afoo : T -> T end; signature B = sig include A val bfoo : T -> T end; structure As : A = struct structure Ns = struct type T = string; end; type T = string val mk = "*" fun afoo s = s ^ "a"; end; structure Bs : B = struct open As; fun bfoo s = s ^ "b"; end; structure SafeBs :> B = Bs; Observe the output from PolyML for the signature SafeBs. It prints it out wrongly. And I think this is to do with the sharing constraint: without it, I get what I expect: structure SafeBs : sig structure Ns : N type T val afoo : SafeBs.T -> SafeBs.T val bfoo : SafeBs.T -> SafeBs.T val mk : SafeBs.T end However, with the sharing constraint, I get: structure SafeBs : sig structure Ns : N type T val afoo : A.T -> A.T val bfoo : A.T -> A.T val mk : A.T end I think, given that SafeBs is hiding the types, printing A.T is very confusing, one might even be tempted to say wrong. I would ideally like to have some programmatic control over the printing; perhaps choosing the left or right hand side of the last sharing constraint to be the name of the type that it printed. would be lovely if this was easy to fix - it would help my understanding of type errors enormously! :) cheers, lucas

1 0

polyml input, echo, pipes and PolyML.export oddity
by ldixon＠inf.ed.ac.uk 08 Jul '08

08 Jul '08

Hello, Perhaps this is something to do with my failure to understand unix pipes, but the following confuses me... I open poly from a shell and I type the following: fun loopy () = let val sopt = TextIO.inputLine(TextIO.stdIn); in case sopt of SOME s => (print s; loopy ()) | NONE => () end; PolyML.export ("foo", loopy); OS.Process.exit OS.Process.success; then I make this into a nice program call foo with: cc -o foo foo.o -L${POLYLIB} -lpolymain -lpolyml (${POLYLIB} is the local location of my poly libs) That does what I expect it to, it echos. However, the following misbehaves: echo 'fun loopy () = let val sopt = TextIO.inputLine(TextIO.stdIn); in case sopt of SOME s => (print s; loopy ()) | NONE => () end; PolyML.export ("foo", loopy); OS.Process.exit OS.Process.success;' | poly Somehow the "OS.Process.exit OS.Process.success;" becomes the first input whenever I foo... Is this my confusion, or is it really misbehaving? (PolyML5.2, Ubuntu 8.04) cheers, lucas

2 1

strange failures (problem with shell on Debian?)
by Michael.Norrish＠nicta.com.au 04 Jul '08

04 Jul '08

I have a relatively short piece of ML looking like val _ = PolyML.Compiler.prompt1 := "" val _ = PolyML.print_depth 0 val dir = OS.FileSys.getDir() val _ = OS.FileSys.chDir "somepath" val _ = use "somefile" val _ = OS.FileSys.chDir dir val _ = load "another file" val _ = print "Done\n"; (This text is also in the file foo.) The various files that are 'use'-d are quite big, and the first one includes further calls to use. I have found that on my Debian Linux system (2.6.18), I can't reliably do any of the following: poly <<EOF ..text above.. EOF or poly < foo or echo "use \"foo\";" | poly The error is not in the ML. Each time any of these is run, I see the final "Done". However, a high proportion of the time (33% perhaps), my shell (standard Debian bash) aborts the execution of the process. For example, with the echo "use" piped to poly, I get ...output... Done val it = () : unit Aborted If I use the here document or the redirect from a file, I also get "Aborted" appended to the expected output. So, this really doesn't look like poly's fault, but it beats me as to what the problem may be. Has anyone seen anything like this before? And, is there another route by which I might be able to get my program into poly that won't suffer this problem? On MacOSX, the system uses the here document approach and I've never seen it fail. Michael.

2 1

structure sharing
by ldixon＠inf.ed.ac.uk 02 Jul '08

02 Jul '08

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I've been trying to use the module system and come across a few issues, some of which I cannot easily pin down, and one which I can. The issue I can clearly identify concerns what seems to be failure of associativity in structure sharing, and is demonstrated in the following: signature ORD_DATA = sig type data end; signature VERTEX = sig type name type data end; signature BVERTEX = sig structure Boundary : ORD_DATA include VERTEX end; signature GRAPH = sig structure Vertex : VERTEX type T; end; signature BGRAPH = sig include GRAPH; structure BVertex : BVERTEX sharing Vertex = BVertex end; signature RG_VERTEX = sig include BVERTEX where type Boundary.data = unit end; signature RG_GRAPH = sig include BGRAPH structure RGVertex : RG_VERTEX sharing BVertex = RGVertex (* why is it OK if I change BVertex to Vertex? *) end Running this gives the error: Error: in '/tmp/sml25644bEn', line 26. Cannot share: (unit) is a type function Found near sig include BGRAPH structure RGVertex : RG_VERTEX sharing BVertex = RGVertex end Exception- Fail Static errors (pass2) raised Apart from just being confused by what the error means, I am confused that given, in BGraph, "sharing Vertex = BVertex", later I get an error if "BVertex = RGVertex" but not if "Vertex = RGVertex". Is there some subtle (or obvious?) thing I'm missing? Another thing I don't quite understand, but which I suspect is just part of the definition of ML, is that you cannot write "sharing type localtype = globaltype", is there some clever reason why this isn't allowed? Or even "sharing local_structure = global_structure", would sometimes be nice. :) The other issue I have is with the names given to types being printed in error messages; is there any way to control this? I sometimes seem to get mis-typed types; which makes errors hard to spot. A common thing I find is that a parameterised type loses it's parameter. I have example of this I can repeat, but they require a large amount of code... I'll try to find a smaller example. any suggestions/corrections/infomration, and possibly bug fixes, very welcome! :) cheers, lucas -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFINuQvogMqU4edHsMRAnq3AKDEOOZK6W4Yl9qLIybMvIYRmrGu8ACeKZIG 2RtB76HxIEiw+Kxd/xiYvGw= =mxpJ -----END PGP SIGNATURE-----

4 8

Real.fromDecimal and Real.~
by fluet＠tti-c.org 18 Jun '08

18 Jun '08

* One bug with the implementation of Real.fromDecimal (which also affects Real.scan and Real.fromString) in PolyML 5.2. The string_buffer[30] in Real_convc in <src>/libpolyml/reals.cpp limits the length of the string seen by strtod to 29 characters. However, the construction of the ML string at <src>/basis/Real.sml:359 can construct an arbitrarily large string. In particular, the "E" and exponent can be pushed beyond the end of the string seen by strtod: Poly/ML 5.2 Release > Real.fromString "1234567890.12345678901234"; val it = SOME 1234567890.0 : Real.real option > Real.fromString "1234567890.123456789012345"; val it = SOME 1.23456789 : Real.real option > Real.fromString "1234567890.1234567890123456"; val it = NONE : Real.real option > Real.fromString "1234567890.1234567890123456789"; val it = SOME 0.123456789 : Real.real option * Real.~ doesn't properly flip the sign bit on zero: Poly/ML 5.2 Release > Real.signBit 0.0; val it = false : bool > Real.signBit (Real.~ 0.0); val it = false : bool > Real.signBit 1.0; val it = false : bool > Real.signBit (Real.~ 1.0); val it = true : bool This, in turn, makes 'Real.fromString "~0.0"' get the sign bit wrong: > Real.signBit (valOf (Real.fromString "~0.0")); val it = false : bool

2 1

Release 5.2 issues
by David.Matthews＠prolingua.co.uk 17 Jun '08

17 Jun '08

I've been away for the last week or so since releasing version 5.2 and there have been a few emails about various issues. Rather than deal with each individually I'm responding to all of them together. > I have installed polyml 5.2 (sources) under > MacOSX 10.4.* and 10.5.*. Both times I had > to manually change in the "make install"-step > the attribute of the file install-sh to 755. > Otherwise the make process would refuse to > finish with permission denied message. Maybe > I have been stupid, but it could also be a bug. The version of install-sh in the CVS repository on sourceforge does not have execute permission and I don't believe it's possible to add it to an existing file. I exported from the repository and forgot to set it explicitly before creating the tar file for the version 5.2 release. It only seems to be a problem on certain platforms. I don't generally like changing a distribution once it has been uploaded but I think the easiest way would be for me to upload a new tar file with execute permissions set on install-sh. > The PolyML.get_print_depth function appears to be missing from 5.2. > Should I rework my code to do without it or was this an unintentional > omission? The PolyML structure has been reworked and in particular PolyML.Compiler has changed. In the old persistent storage version (PolyML 4 and earlier) the print depth and similar values were special kinds of references. From version 5 these are just simple references. The reference set by PolyML.print_depth is now in PolyML.Compiler.printDepth so you can define get_print_depth as ! PolyML.Compiler.printDepth. This is used if the compiler is invoked by means of the usual top-level or with "use" but if PolyML.compiler is called with a value passed in with PolyML.Compiler.CPPrintDepth then this is used in place of the reference. > Has there been a change to the semantics of the function Unix.execute? > > Previously, it executed the programme with respect to the current > path. Now, it seems to require an absolute pathname. There were some changes to Unix.execute and while I was doing it I took a closer look at the Basis Library document. This says that the "cmd" argument should name an executable file. Previously the code used "/bin/sh -c" to run the command. It's easy to add this in the call if you want it to search the path. David.

1 0

Release 5.2
by David.Matthews＠prolingua.co.uk 03 Jun '08

03 Jun '08

I've now released version 5.2 on SourceForge. There are no major changes in this release but it contains some updates and bug fixes. I'll update the release notes shortly but these are the major changes. Change to PolyML.compiler to allow more control over parameters and addition of "name space". The current PolyML.compiler has a different type from previously so this will affect code that uses it. Improvement to real numbers on X86 (32 and 64 bit). Fix to a hot-spot in the compiler. This and other changes seem to have speeded up the compiler by about 10%. Improvement to the source-level debugger (i.e. when compiling with PolyML.debug set to true), especially tracing. Addition of weak references. The Weak structure contains functions to create weak references and arrays and a condition variable and associated mutex that is signalled when a weak reference has been deleted by the garbage collector. Changes to handling of signals by the Signal structure. Rather than forking a new thread to process each signal there is a single thread that executes all the signal handlers. This should improve the responsiveness to an asynchronous signals such as SIGINT. The original control-C handler has been removed and the Poly/ML read-eval-print loop now installs a simple handler for control-C that just generates an Interrupt exception. There have been various bug fixes and probably a few other things that I've forgotten. As always, let me know how it goes. David

1 0

Fwd: [polyml] Re: Interactive behaviour and ctrl-c
by David.Matthews＠prolingua.co.uk 31 May '08

31 May '08

Dear Tom, Your example is essentially doing nothing but growing the stack. The stack is held in the heap and whenever the current segment runs out a new, larger, segment is allocated and the current contents of the stack are copied into that. That can result in garbage collection. Both garbage collection and stack copying are non-interruptible so in either case an interrupt will be deferred until the action is complete. The handling of ctrl/C has been improved in the CVS version compared with the 5.1 release but there are still delays if ctrl/C is pressed during a garbage collection. While investigating this I did discover a bug that resulted in the garbage collector saying that there was insufficient space to extend the stack when in fact there was enough. It's likely now that your example would result in a Warning - Stack limit reached - interrupting process message since the stack will probably run out before the ctrl/C is processed. There may be more of a problem with the 64-bit version than the 32-bit version because the maximum size of a stack segment is larger in the 64-bit version so the limitation is likely to be the overall memory rather than the stack size. David Tom Ridge wrote: > Dear David, > > I just tried the following experiment in poly (see below). > > Under mosml, ctrl-c causes an almost instantaneous return to > top-level. Under poly, the time to return seems dependent on the depth > of recursion(!). Does this sound possible? Is there anyway to get poly > to raise the exception to toplevel in a more timely fashion? > > Thanks > > Tom > > ---------- Forwarded message ---------- > From: Tom Ridge <thomas.ridge(a)cl.cam.ac.uk> > Date: 2008/5/23 > Subject: Re: [polyml] Re: Interactive behaviour and ctrl-c > To: Michael Norrish <Michael.Norrish(a)nicta.com.au> > Cc: hol-developers(a)lists.sourceforge.net > > > Dear All, > > I've just tried this simple experiment in poly's top-level: > > --- > > fun f x = 1 + f x; > > f 1; > > (* leave this to run for 10 or 20 seconds *) > > (* type ctrl-c *) > > (* type f to raise an exception *) > > (* now wait while the exception is raised... this takes a while *) > > --- > > So it seems that the slow-ctrl-c behaviour is something to do with > (64bit) poly, and not HOL. Can someone confirm that this *doesn't* > happen under 32bit poly (linux and Mac). >

2 2

Poly/ML syntax for constructors too liberal?
by rda＠lemma-one.com 13 May '08

13 May '08

Poly/ML correctly allows the following binding: val op :: (h, t) = [1, 2, 3, 4]; However it also allows brackets around the constructor: val (op ::) (h, t) = [1, 2, 3, 4]; My reading of the syntax in the definition of Standard ML is that the latter form should not be allowed. Am I right? Regards, Rob.

4 5

← Newer
1
...
54
55
56
57
58
59
60
...
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