SCM Repository

[smlnj] Diff of /sml/trunk/HISTORY

[smlnj] / sml / trunk / HISTORY

Diff of /sml/trunk/HISTORY

-revision 1447, Mon Jan 26 17:24:47 2004 UTC
+revision 1580, Tue Aug  3 19:26:26 2004 UTC
 Line 14
  ----------------------------------------------------------------------
  Name: Matthias Blume (blume (at) tti - c (dot) org)
- Date: 2004/01/26
+ Date: 2004/08/03 14:25:00 CDT
+ Tag: blume-20040803-callingconv
+ Description:
+ Added low-level support for choosing C calling conventions by
+ twiddling the type of rawccall.  (See
+ src/compiler/Semant/types/cproto.sml for details.)
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/08/02 15:55:00 CDT
+ Tag: blume-20040802-backout
+ Description:
+ Backed out of change to win32-filesys.c.  The earlier patch to
+ get_file_time caused CM to produce files with the wrong time stamp.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/08/02 14:45:00 CDT
+ Tag: blume-20040802-nlffi-win32
+ Description:
+ Added NLFFI support for Win32, adapted from a patch provided by David
+ Hansel.  This is currently completely untested.  Also, the issue
+ concerning stdcall vs. ccall is still unresolved.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/07/30 17:55:00 CDT
+ Tag: blume-20040730-various
+ Description:
+ Gearing up towards 110.47...
+ - various minor bugfixes to ml-nlffigen
+ - a beginning of a manual for nlffi
+ - eliminated 'export name=value' in config/install.sh as this does
+   not work with certain versions of /bin/sh
+   (Thanks to David King at Motorola for catching this.)
+ - several bugfixes provided or suggested by David Hansel at Reactive Systems:
+   - added a test for tm==NULL to gmtime.c and localtime.c
+   - applied patch for incorrect GetFileTime under win32
+   - toSeconds -> toMilliseconds in Win32/win32-process.sml
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/07/21 18:20:00 CDT
+ Tag: blume-20040721-nlffigen
+ Description:
+ - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
+   for a typedef to an incomplete type, but leave out the "typ" member.
+   (This is just for consistency.)
+ - Started to produce what is supposed to become better (i.e., comprehensive)
+   documentation of what ml-nlffigen does and produces.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/07/14 16:25:00 CDT
+ Tag: blume-20040714-union
+ Description:
+ Added C_UNION to c-calls/c-types.sml and updated the machinery
+ (ml-nlffigen, cproto.sml) that conveys C function interface
+ information to the code generator.
+ However, the actual architecture-specific implementation of function
+ arguments and results that are C unions is still not implemented.
+ ----------------------------------------------------------------------
+ Name: Allen Leung (leunga (at) reservoir (dot) com)
+ Date: 2004/07/14 14:38:00 EST
+ Tag: leunga-110_46_1-ppc-lwzu
+ Description:
+     Added these instructions to the PowerPC architecture:
+     LBZU(X), LHZU(X), LWZU(X),
+     STWU(X), STFDU, STFSU
+     etc...
+     Note: I haven't added their instruction encoding into the description.
+ ----------------------------------------------------------------------
+ Name: Allen Leung (leunga (at) reservoir (dot) com)
+ Date: 2004/07/13 15:04:00 EST
+ Tag: leunga-110_46_1-ppc-lwarx
+ Description:
+    Added the two instructions LWARX and STWCX to the PowerPC
+ instruction set.
+    A (untested) rewrite of loop-structure.sml.  The old version
+ is completely broken.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/07/13 13:50:00 CDT
+ Tag: blume-20040713-nlffi
+ Description:
+ - use paramAlloc to report c-calls with too many arguments
+   (for PPC version where parameter area is pre-allocated)
+ - added ccall_maxargspace to machspec (to implement the above)
+ - made "make" commend in CM's "make" tool configurable
+ - added option (default: on) for passing the name of the SML/NJ's "bin"
+   directory to "make"; the call looks like this:
+      make <options> SMLNJ_BINDIR=<dir> <target>
+   This can be used by the Makefile to, e.g., pick the "right" version
+   of ml-nlffigen.
+ - minor code tweaks
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/07/12 22:50:00 CDT
+ Tag: blume-110_46_1-macosx-nlffi
+ Description:
+ NLFFI under Mac OS X now working (sort of).  This is largely untested,
+ though.
+ Note:  1. You have to make a new, clean build of the runtime system.
+. There are new BOOTFILES, you have to use them!
+           (Doing the bootstrap process yourself would be *very* painful!
+            If you absolutely have to do it, build the system under
+            a different architecture and then cross-compile.)
+ Version bumped to 110.46.1 to account for runtime data format changes.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/06/18 14:30:00 CDT
+ Tag: blume-20040618-unix
+ Description:
+ Changed the implementation of structure Unix so that the same stream
+ is returned every time one of the {text,bin}{In,Out}streamOf functions
+ is invoked on the same proc.  This is not what the spec currently
+ says -- although IMO it arguably should.  (See discussion below.)
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/06/17 18:15:00 CDT
+ Tag: Release_110_46
+ Description:
+ New working version (110.46).  NEW BOOTFILES!
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/06/17 17:20:00 CDT
+ Tag: blume-20040617-timer-unix
+ Description:
+ Changed the interface of structures Timer and Unix to match the most
+ recent Basis spec.
+ In the case of Unix there still seems to be an open/weird issue:
+    The {text,bin}{In,Out}streamOf functions are supposed to create
+    fresh streams whenever they are called -- as opposed to have them
+    return the same stream every time.  This design is supposed to
+    prevent space leaks caused by proc values hanging on to streams.
+    The reap function, on the other hand, is supposed to close the
+    streams.  This cannot be done without having a handle on the
+    stream in proc after all...
+ I took the liberty to implement the following stopgap solution:
+   The proc value hangs on to the most recently created stream(s).
+   Reap closes those.  If either or both of the two streams hadn't
+   been created at all yet, then reap will close the corresponding
+   file descriptors directly.
+ PS: I don't understand the original space leak argument anymore.  If
+ a proc hangs on to the imperative stream, then I/O operations on those
+ will advance the state of the cached stream and avoid the space leak.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/05/28 16:45:00 CDT
+ Tag: blume-20040528-basis
+ Description:
+ Added signature PACK_REAL and exported functor PrimIO.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/05/25 16:00:00 CDT
+ Tag: blume-20040525-group-owner
+ Description:
+ CM now ignores (but still accepts) the "owner" information in group
+ descriptions.  The owner of a group is its next enclosing
+ library. Each group must have a unique owner.  (There is a virtual
+ "toplevel" library that owns groups which are not nested within a real
+ library.)  Previously, each group had to explicitly declare its owner,
+ and CM would check that such a declaration is correct.  The new scheme
+ is to have CM check that for each group there is precisely one owning
+ library.
+ The advantage of the new scheme is that the programmer no longer needs
+ to maintain the somewhat annoying owner information.  The downside is
+ that CM cannot enforce the ownership rule across multiple runs of
+ CM.make.  Fortunately, enclosing the same group in two different
+ libraries A and B which are not part of the same program does not
+ cause real problems.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/05/20 16:00:00 CDT
+ Tag: blume-20040520-win32
+ Description:
+ Made the win32 version work again.  (Strangely, a misplaced comma had
+ slipped into win32-process.c which prevented the runtime from being
+ compiled correctly.)
+ Also, included a minor addition to ml-build.bat analogous to what was
+ done in blume-20040519-ml-build.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/05/19 22:10:00 CDT
+ Tag: blume-20040519-ml-build
+ Description:
+ Arranged for ml-build to clean up after itself a little bit better.
+ The script generates a temporary SML source file and compiles it using
+ CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
+ gets rid of those at the end, so they don't accumulate under .cm.
+ This required a minor change to install.sh because the name of the
+ metadata directory (default: .cm) is actually configurable at
+ installation time.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/05/18 15:50:00 CDT
+ Tag: blume-20040518-mkreader
+ Description:
+ Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
+ implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/05/11 14:35:00 CDT
+ Tag: blume-20040511-win32sock
+ Description:
+ Added previously missing support for many socket-related functions
+ under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
+ for the voluminous patch!
+ (I have not tested this patch under win32 yet.)
+ Here is David's e-mail:
+ Hi,
+ Attached to this email you find a diff against sml/nj 110.45
+ that will enable socket support under Windows.
+ To apply the patch (using unix or cygwin)
+) gunzip runtime.diff.gz
+) "cd" into "src/runtime" in the source tree of a fresh
+.45 installation.
+) patch -p 1 < [your/path/to]runtime.diff
+ The code compiles fine but has NOT yet been extensively tested.
+ I only ran a few tests for basic socket client functionality
+ (which worked fine).  Especially the functions that use ioctl
+ are not tested at all and might not work (see below).
+ I implemented this since we want to move to a newer version of sml/nj
+ but need socket support in order to use it.  This is the first time I
+ even had a look at the sml/nj source,  so please review my changes
+ before making this part of the distribution!  Here are a few issues
+ that I think might be better for someone to solve who is more
+ familiar with the sml/nj source (and socket programming):
+ - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
+   exception since I could not figure out what the windows equivalent
+   of these functions is
+ - In sockets-osdep.h there are a some #include statements that are
+   only used in a few files that include sockets-osdep.h
+ - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
+   does not process its return value since I don't know how to
+   report an error upwards.
+ - It would probably be good to have a call to WSACleanup() when
+   the library is unloaded (if there is such a possibility).
+   Otherwise I think Windows will take care of this automatically
+   when the process finishes.
+ - I used ioctlsocket() as a replacement for ioctl() but I have
+   no idea if that is actually the proper replacement on Windows.
+ - All these issues are marked in the code by "FIXME" comments.
+ We use sml/nj extensively in our products and are quite happy
+ with it.  I hope this contribution will help you.
+ Keep up the good work!
+ David
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/05/11 14:20:00 CDT
+ Tag: blume-20040511-installml
+ Description:
+ Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
+ the patch.)
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/05/11 14:05:00 CDT
+ Tag: blume-20040511-nlffi-netbsd
+ Description:
+ Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
+ the patch.)
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/05/11 12:05:00 CDT
+ Tag: blume-20040511-exports
+ Description:
+ As per request by Adam Chlipala <adam@hcoop.net>, extended various
+ export lists in compiler-related .cm-files.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/05/11 11:35:00 CDT
+ Tag: blume-20040511-allsource
+ Description:
+ The installer now honors the "src-smlnj" target again, although its meaning
+ has changed from "all sources required for the compiler" to "all sources
+ the installer knows about".  In other words, if you enable "src-smlnj"
+ in the "targets" file, then the installer will pull in sources for
+ everything.  (Notice that this refers to source code only.  Compiled
+ code is still only installed for modules that were requested explicitly
+ or which are required for other modules that were requested explicitly.)
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/04/23 17:40:00 CDT
+ Tag: blume-20040423-ieee-scan
+ Description:
+ Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
+ in the exponent calculation we get INF or ZERO (depending on the mantissa
+ and the sign of the exponent).
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/04/23 10:40:00 CDT
+ Tag: blume-20040423-ml-build
+ Description:
+ The ml-build script now terminates with a non-0 status when something
+ goes wrong.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/04/22 16:35:00 CDT
+ Tag: blume-20040422-Option
+ Description:
+ Made exception Option to be the same as exception Option.Option
+ (as it should be).
+ ----------------------------------------------------------------------
+ Name: Allen Leung (leunga (at) reservoir (dot) com)
+ Date: 2004/03/19 14:40:00 EST
+ Tag: leunga-20040319-cygwin-nlffi
+ Description:
+ Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
+ of SML/NJ.  The problem is that
+    lib = dlopen(NULL, ...)
+    f   = dlsym(lib, "malloc");
+ does not work on Windows unless we explicitly export symbols
+ such as 'malloc' during linking.  We fixed this by explicitly
+ exporting the required symbols with the magic gcc incantation:
+     -Wl,--export-all cygwin.def
+ where cygwin.def is a file containing all the symbols that
+ we wish to export.
+ I suspect this is a Windows problem and we'll have to
+ do the same (somehow with windows compilers) when
+ we build the native win32 version with the system
+ calls LoadLibrary/GetProcAddress.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/03/04 16:35:00 CST
+ Tag: blume-20040304-intinf-fmt
+ Description:
+ Fixed problem with IntInf.fmt (sign would show up on the right instead
+ of on the left for BIN, OCT, and HEX).
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/03/04 11:25:00 CST
+ Tag: blume-20040304-symlinks
+ Description:
+ Fixed problem with installer script (unix only) where bin/ml-yacc and
+ friends pointed (via symlinks) to absolute locations instead of just
+ .run-sml.  This was reported by Vesa A Norrman.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/02/13 14:50:00 CST
+ Tag: Release_110_45
+ Description:
+ New working version (110.45). New bootfiles.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/01/26 15:15:15 CST
+ Tag: blume-20040126-toplevel
+ Description:
+ Improved handling of exceptions at the interactive toplevel.
+ ----------------------------------------------------------------------
+ Name: Matthias Blume (blume (at) tti - c (dot) org)
+ Date: 2004/01/26 11:25:00
  Tag: blume-20040126-app
  Description:

 Legend:



Removed from v.1447
 


changed lines


 
Added in v.1580
 Legend:



Removed from v.1447
 


changed lines


 
Added in v.1580
-Removed from v.1447
+Added in v.1580

root@smlnj-gforge.cs.uchicago.edu	ViewVC Help
Powered by ViewVC 1.0.0