Skip to content

Commit

Permalink
Merge remote-tracking branch 'refs/remotes/origin/master'
Browse files Browse the repository at this point in the history
# Conflicts:
#	m3-libs/m3core/src/thread/WIN32/ThreadWin32.m3
  • Loading branch information
jaykrell committed Dec 4, 2016
2 parents 68a5fbf + fac9d74 commit f6d7b2a
Show file tree
Hide file tree
Showing 214 changed files with 10,642 additions and 822 deletions.
78 changes: 77 additions & 1 deletion .gitignore
@@ -1,4 +1,80 @@
scripts/PKGS
ALPHA_LINUX
ALPHA_NETBSD
ALPHA_OPENBSD
ALPHA_OSF
ALPHA32_VMS
ALPHA64_VMS
AMD64_DARWIN
I386_DARWIN
AMD64_FREEBSD
AMD64_LINUX
AMD64_NETBSD
AMD64_NT
AMD64_OPENBSD
AMD64_SOLARIS
ARM_DARWIN
ARM_LINUX
ARM6_LINUX
ARM7_LINUX
ARMEL_LINUX
ARMV6_LINUX
ARMV7_LINUX
FBSD_ALPHA
FreeBSD2
FreeBSD3
FreeBSD4
FreeBSD5
HPPA
I386_CYGWIN
I386_DARWIN
I386_FREEBSD
I386_INTERIX
I386_LINUX
I386_MINGW
I386_MSDOS
I386_NETBSD
I386_NT
I386_OPENBSD
I386_SOLARIS
IA64_FREEBSD
IA64_HPUX
IA64_LINUX
IA64_NT
IA64_VMS
IBMR2
IRIX5
LINUXELF
LINUXLIBC6
MIPS32_IRIX
MIPS32_LINUX
MIPS64_IRIX
MIPS64_LINUX
MIPS64_OPENBSD
MIPS64EL_OPENBSD
NetBSD2_i386
NT386
NT386GNU
NT386MINGNU
PA_HPUX
PA32_HPUX
PA32_LINUX
PA64_HPUX
PA64_LINUX
PPC_DARWIN
PPC_LINUX
PPC_OPENBSD
PPC32_AIX
PPC32_OPENBSD
PPC64_AIX
PPC64_DARWIN
PPC64_LINUX
SOLgnu
SOLsun
SPARC
SPARC_LINUX
SPARC_SOLARIS
SPARC32_LINUX
SPARC32_SOLARIS
SPARC64_LINUX
SPARC64_OPENBSD
SPARC64_SOLARIS
25 changes: 12 additions & 13 deletions README-unicode
Expand Up @@ -5,14 +5,13 @@ The language has these changes:
ORD(LAST(WIDECHAR)) = 16_10FFFF, the entire code point range specified
by the Unicode standard.

Character and Text literals can have Unicode escapes of the form
\Uhhhhhh, where each h is a hexadecimal digit, in either upper or
lower case. These are lexically valid in any literal, wide or not.
It is a static error if the code point value exceeds the range of CHAR
or Unicode-sized WIDECHAR. In a wide literal, and when the compiler
is configured for 16-bit WIDECHAR, values outside the 16-bit range are
converted to the Unicode "replacement" character,
VAL(16_FFFD,WIDECHAR), with a warning.
Wide character and text literals can have Unicode escapes of the form
\Uhhhhhh or \uhhhhhh, where each h is a hexadecimal digit, in either
upper or lower case. It is a static error if the code point value
exceeds the range of Unicode-sized WIDECHAR. When the compiler is
configured for 16-bit WIDECHAR, values outside the 16-bit range are
converted to the Unicode "replacement" character, VAL(16_FFFD,WIDECHAR),
with a warning.

The subtype and assignability rules are relaxed as if CHAR and
WIDECHAR were the same base type. This allows assignments between
Expand Down Expand Up @@ -48,7 +47,7 @@ same-named procedures in the new interfaces.

The aforementioned interfaces do their own synchronization, and thus
provide atomic operations. The Unsafe* interfaces provide equivalent
funcions, but do not synchronize, expecting their callers to do it, or
functions, but do not synchronize, expecting their callers to do it, or
ensure it is unnecessary.

Consistency of WIDECHAR size.
Expand Down Expand Up @@ -82,7 +81,7 @@ overridden by command line options -widechar-uni or widechar-16.
However, this is likely to be tedious because of the consistency
requirement.

Compatability of WIDECHARs in pickles.
Compatibility of WIDECHARs in pickles.

Programs with different WIDECHAR size can interchange pickles
containing WIDECHARs in either direction, if both are linked to a
Expand All @@ -97,11 +96,11 @@ reverse is not true. A pre-Unicode libm3core reading a pickle that
was written by a program compiled with Unicode-sized WIDECHAR will
raise an exception, even if the pickle contains no actual WIDECHARs.

Network object compatability.
Network object compatibility.

The compatability rules for WIDECHARs transferred by network object
The compatibility rules for WIDECHARs transferred by network object
calls are the same as for pickles. But remember that network object
calls involve two-way transfer of data, and thus will, in general,
require the two-way compabability of post-Unicode libm3core and
require the two-way compabibility of post-Unicode libm3core and
post-Unicode m3netobj to work.

2 changes: 1 addition & 1 deletion doc/help/gen_html/m3core/src/win32/WinBase.i3.html
Expand Up @@ -41,7 +41,7 @@ <H2>m3core/src/win32/WinBase.i3</H2></A><HR>
PSID, PACL, PVOID, PLUID, LARGE_INTEGER, HANDLE, PHANDLE;
FROM <A HREF="../C/Common/Ctypes.i3.html">Ctypes</A> IMPORT char, int, void_star;
</PRE><P>
* Compatability macros
* Compatibility macros


<P><PRE>CONST
Expand Down
2 changes: 1 addition & 1 deletion doc/help/gen_html/visualobliq/src/Dialog.m3.html
Expand Up @@ -437,7 +437,7 @@ <H2>visualobliq/src/Dialog.m3</H2></A><HR>

NodeVBT.LoadFromFile(screen[1], s);

IF loadingVerNum &gt;= 30 THEN (* only for backward compatability *)
IF loadingVerNum &gt;= 30 THEN (* only for backward compatibility *)
(* load the session constructor code - version 3.0 and above *)
RW.rtext(s, GenerateObliq.sessionConstructor);
(* load the global code - version 3.0 and above *)
Expand Down
2 changes: 1 addition & 1 deletion doc/notes/porting.txt
Expand Up @@ -24,7 +24,7 @@ Our portability rests on some large pillars:
See ThreadPosixC.c, it contains both get/setcontext and sigaltstack variants.
sigstack might be the way to go on e.g. VMS, where pthreads don't get context.
Lastly the old way of digging around in jmpbuf remains viable, but not portable,
if desparate.
if desperate.


In the past, a large tedious error-prone part of porting was "rewriting /usr/include in Modula-3".
Expand Down
4 changes: 2 additions & 2 deletions doc/notes/todo.txt
Expand Up @@ -74,8 +74,8 @@ and/or produce enough type information in backend
to make gdb much better and m3gdb moot
-----------------------------------------------------------------------------
write a C or C++ generating backend
C++ is useful for possbly more portable optimized exception handling.
That is, given a C++ compiler that implementations exception handling,
C++ is useful for possibly more portable optimized exception handling.
That is, given a C++ compiler that implements exception handling,
which is all these days, and most through time, but not all through time,
there is some good chance the exception handling is more efficient
than setjmp/longjmp
Expand Down
12 changes: 12 additions & 0 deletions doc/reference/complete/.gitignore
@@ -0,0 +1,12 @@
m3defn-complete.aux
m3defn-complete.css
m3defn-complete.dvi
m3defn-completeimages.aux
m3defn-completeimages.log
m3defn-completeimages.pl
m3defn-completeimages.tex
m3defn-completeimg1.old
m3defn-completeimg2.old
m3defn-completelabels.pl
m3defn-complete.log

2 changes: 2 additions & 0 deletions doc/reference/complete/WARNINGS
@@ -0,0 +1,2 @@
No implementation found for style `graphicx'
No implementation found for style `url'
7 changes: 7 additions & 0 deletions doc/reference/complete/html/.gitignore
@@ -0,0 +1,7 @@
images.aux
images.log
images.pl
images.tex
img1.old
img2.old
labels.pl

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

0 comments on commit f6d7b2a

Please sign in to comment.