]> granicus.if.org Git - gc/log
gc
5 years ago.gitignore: Ignore CMake-generated libgc.dll.a
Ivan Maidanski [Thu, 13 Jun 2019 08:02:15 +0000 (11:02 +0300)]
.gitignore: Ignore CMake-generated libgc.dll.a

5 years agoRemove unnecessary checks of __ELF__ for NaCl
Ivan Maidanski [Tue, 11 Jun 2019 07:09:21 +0000 (10:09 +0300)]
Remove unnecessary checks of __ELF__ for NaCl
(code refactoring of commit 9738a14a0)

* dyn_load.c [NACL]: Assume __ELF__ is always defined.

5 years ago.gitignore: Ignore CMake-generated libgc.a
Ivan Maidanski [Tue, 11 Jun 2019 06:37:20 +0000 (09:37 +0300)]
.gitignore: Ignore CMake-generated libgc.a

5 years agoUpdate AUTHORS file (add Maya Rashish)
Ivan Maidanski [Tue, 11 Jun 2019 06:36:09 +0000 (09:36 +0300)]
Update AUTHORS file (add Maya Rashish)

5 years agoGet rid of some non-ELF ifdefs
Maya Rashish [Thu, 6 Jun 2019 02:43:51 +0000 (05:43 +0300)]
Get rid of some non-ELF ifdefs
(code refactoring)

Issue #283 (bdwgc).

Linux started switching to ELF in 1995 (Linux-2.0), due to the previous
a.out being really bad.
NetBSD has used ELF since 2002 (NetBSD 1.6), it's OK to let it go now.
FreeBSD switched to ELF in FreeBSD 3.0 (1998).

* include/private/gcconfig.h [(POWERPC || SPARC || MIPS || ALPHA
|| AARCH64 || ARM32 || X86_64) && FREEBSD] (DYNAMIC_LOADING): Define
regardless of __ELF__ (assuming __ELF__ is always defined).
* include/private/gcconfig.h [(ARM32 || X86_64 || HEXAGON) && LINUX]
(DYNAMIC_LOADING, SEARCH_FOR_DATA_START, DATASTART, DATAEND): Likewise.
* include/private/gcconfig.h [X86_64 && NETBSD] (DATASTART,
DYNAMIC_LOADING): Likewise.

5 years agoFix mmap(PROT_NONE) failure on AIX
Peter Wang [Thu, 6 Jun 2019 07:36:54 +0000 (17:36 +1000)]
Fix mmap(PROT_NONE) failure on AIX

Issue #285 (bdwgc).

Unmap memory using mprotect(PROT_NONE) instead of mmap(PROT_NONE) on AIX.

* os_dep [USE_MUNMAP && AIX] (GC_unmap, GC_unmap_gap): Call mprotect()
instead of mmap(); update comment.

5 years agoAdd CMake option to compile all library .c files into single object file
Ivan Maidanski [Fri, 7 Jun 2019 18:33:32 +0000 (21:33 +0300)]
Add CMake option to compile all library .c files into single object file

Issue #281 (bdwgc).

* CMakeLists.txt (enable_single_obj_compilation): New option (off by
default).
* CMakeLists.txt [enable_single_obj_compilation] (SRC): Override value
to extra/gc.c.
* CMakeLists.txt [enable_single_obj_compilation
&& CMAKE_USE_PTHREADS_INIT]: Define GC_PTHREAD_START_STANDALONE macro.
* CMakeLists.txt [enable_single_obj_compilation
&& CMAKE_USE_PTHREADS_INIT] (SRC): Append pthread_start.c (in addition
to extra/gc.c).

5 years agoFix build for OS X (CMake)
Victor Romero [Fri, 7 Jun 2019 07:35:31 +0000 (10:35 +0300)]
Fix build for OS X (CMake)

Issue #275 (bdwgc).

Having "ppc" in CMAKE_OSX_ARCHITECTURES fails the build on Mac
OS X 10.13, at least.

* CMakeLists.txt [APPLE && $CMAKE_OSX_ARCHITECTURES==""]
(CMAKE_OSX_ARCHITECTURES): Set to "x86_64;i386" (instead of
"ppc;i386;x86_64").

5 years agoDefine _CRT_SECURE_NO_DEPRECATE macro in CMake script only for MS VC
Victor Romero [Fri, 7 Jun 2019 06:40:20 +0000 (09:40 +0300)]
Define _CRT_SECURE_NO_DEPRECATE macro in CMake script only for MS VC
(fix of commit f5006fcc5)

Issue #281 (bdwgc).

* CMakeLists.txt (add_definitions): Define _CRT_SECURE_NO_DEPRECATE
only if MSVC (instead of WIN32).

5 years agoSupport header files installation in CMake script
Victor Romero [Fri, 7 Jun 2019 06:34:27 +0000 (09:34 +0300)]
Support header files installation in CMake script

Issue #281 (bdwgc).

Now, "cmake --target install" copies the public header files (to the
installation destination) unless -Dinstall_headers=OFF is passed to
cmake.

* CMakeLists.txt (install_headers): New option (on by default).
* CMakeLists.txt [install_headers] (include/gc.h, include/gc_backptr.h,
include/gc_config_macros.h, include/gc_gcj.h, include/gc_inline.h,
include/gc_mark.h, include/gc_pthread_redirects.h,
include/gc_tiny_fl.h, include/gc_typed.h, include/gc_version.h,
include/javaxfc.h, include/leak_detector.h): Specify install to
"include/gc" folder.
* CMakeLists.txt [install_headers && enable_cplusplus]
(include/gc_allocator.h, include/gc_cpp.h): Likewise.
* CMakeLists.txt [install_headers && enable_disclaim]
(include/gc_disclaim.h): Likewise.
* CMakeLists.txt [install_headers] (include/extra/gc.h): Specify
install to "include" folder.
* CMakeLists.txt [install_headers && enable_cplusplus]
(include/extra/gc_cpp.h): Likewise.

5 years agoAppveyor CI: Turn on build_test and enable_cplusplus for CMake
Ivan Maidanski [Thu, 6 Jun 2019 07:29:56 +0000 (10:29 +0300)]
Appveyor CI: Turn on build_test and enable_cplusplus for CMake

5 years agoDo not build the tests by default (CMake)
Victor Romero [Thu, 6 Jun 2019 07:24:28 +0000 (10:24 +0300)]
Do not build the tests by default (CMake)

Issue #281 (bdwgc).

build_cord and build_tests options are introduced in the CMake script.

* CMakeLists.txt (build_cord): Add option (on by default); add comment.
* CMakeLists.txt (build_tests): Add option (off by default).
* CMakeLists.txt (cord): Specify add_subdirectory only if build_cord.
* CMakeLists.txt (tests): Specify add_subdirectory only if build_tests.
* cord/CMakeLists.txt: Skip cordtest and de tests unless build_tests.

5 years agoFix 'ulong undefined' compilation error on AIX
Peter Wang [Thu, 6 Jun 2019 02:48:17 +0000 (12:48 +1000)]
Fix 'ulong undefined' compilation error on AIX

* include/gc.h [_AIX] (GC_DATASTART, GC_DATAEND): Remove cast to ulong.

5 years agoAppveyor CI: Pass --build-config to ctest again
Ivan Maidanski [Thu, 6 Jun 2019 06:48:19 +0000 (09:48 +0300)]
Appveyor CI: Pass --build-config to ctest again
(fix of commit 906280234)

5 years agoAdd cordtest to CMake script and rename cord/cord executable to cord/de
Ivan Maidanski [Thu, 6 Jun 2019 06:40:06 +0000 (09:40 +0300)]
Add cordtest to CMake script and rename cord/cord executable to cord/de

Issue #281 (bdwgc).

* cord/CMakeLists.txt (cordtest): Specify add_executable,
set_target_properties, target_link_libraries and add_test.
* cord/CMakeLists.txt [WIN32] (cord): Rename executable to de.

5 years ago.gitignore: Ignore CMake-generated files in cord folder
Ivan Maidanski [Wed, 5 Jun 2019 08:50:29 +0000 (11:50 +0300)]
.gitignore: Ignore CMake-generated files in cord folder

5 years agoMove CMake cord specification and cord executable to cord folder
Ivan Maidanski [Wed, 5 Jun 2019 08:25:32 +0000 (11:25 +0300)]
Move CMake cord specification and cord executable to cord folder

Issue #281 (bdwgc).

* CMakeLists.txt [WIN32]: Move add_executable, set_target_properties,
target_link_libraries for cord to cord/CMakeLists.txt.
* CMakeLists.txt (cord): Specify add_subdirectory.
* Makefile.am (EXTRA_DIST): Add cord/CMakeLists.txt.
* cord/CMakeLists.txt: New file.

5 years agoDefine _CRT_SECURE_NO_DEPRECATE macro in CMake script only for Win32
Victor Romero [Wed, 5 Jun 2019 07:49:30 +0000 (10:49 +0300)]
Define _CRT_SECURE_NO_DEPRECATE macro in CMake script only for Win32

Issue #281 (bdwgc).

* CMakeLists.txt (add_definitions): Define _CRT_SECURE_NO_DEPRECATE
only if WIN32.

5 years agoAdd tracetest and disclaim tests to CMake script
Ivan Maidanski [Wed, 5 Jun 2019 07:02:28 +0000 (10:02 +0300)]
Add tracetest and disclaim tests to CMake script

Issue #281 (bdwgc).

* tests/CMakeLists.txt [enable_gc_debug] (tracetest): Specify
add_executable, target_link_libraries, add_test.
* tests/CMakeLists.txt [enable_disclaim] (disclaim_bench,
disclaim_test, disclaim_weakmap_test): Likewise.

5 years agoTurn off C++ API in CMake script by default
Ivan Maidanski [Wed, 5 Jun 2019 06:53:15 +0000 (09:53 +0300)]
Turn off C++ API in CMake script by default

Issue #281 (bdwgc).

This is to match the behavior of configure script.
Now, -Denable_cplusplus=ON should be passed to cmake to enable C++ API.

* CMakeLists.txt (enable_cplusplus): Add option (off by default);
remove TODO.
* CMakeLists.txt (SRC): Add gc_cpp.cc only if enable_cplusplus.
* tests/CMakeLists.txt (leak_test.c, test.c): Set CXX langunage
property only if enable_cplusplus.
* tests/CMakeLists.txt [enable_cplusplus]: Add TODO to add test_cpp as
a test.

5 years agoSupport enable_sigrt_signals option by CMake script
Ivan Maidanski [Tue, 4 Jun 2019 20:20:43 +0000 (23:20 +0300)]
Support enable_sigrt_signals option by CMake script

Issue #281 (bdwgc).

* CMakeLists.txt (enable_sigrt_signals): Add option (default to OFF).
* CMakeLists.txt [CMAKE_USE_PTHREADS_INIT && enable_sigrt_signals]:
Define GC_USESIGRT_SIGNALS macro.

5 years agoAppveyor CI: Pass CMAKE_OPTIONS to cmake only when configuring
Ivan Maidanski [Tue, 4 Jun 2019 08:02:18 +0000 (11:02 +0300)]
Appveyor CI: Pass CMAKE_OPTIONS to cmake only when configuring

5 years ago.gitignore: Ignore libgcmt-dll.dll.a (produced by CMake)
Ivan Maidanski [Tue, 4 Jun 2019 07:08:57 +0000 (10:08 +0300)]
.gitignore: Ignore libgcmt-dll.dll.a (produced by CMake)

5 years agoTravis CI: Output content of threadleaktest.log
Ivan Maidanski [Fri, 31 May 2019 06:26:35 +0000 (09:26 +0300)]
Travis CI: Output content of threadleaktest.log

5 years agoUpdate AUTHORS file (add Victor Romero)
Ivan Maidanski [Thu, 30 May 2019 22:07:39 +0000 (01:07 +0300)]
Update AUTHORS file (add Victor Romero)

5 years agoGroup all options near the top of CMake script
Victor Romero [Thu, 30 May 2019 22:05:25 +0000 (01:05 +0300)]
Group all options near the top of CMake script
(code refactoring)

Issue #281 (bdwgc).

* CMakeLists.txt (option): Group all items together close to the
beginning of the file; reorder items to match that of configure.

5 years agoUpgrade cmake_minimum_required(version) to 3.1
Victor Romero [Thu, 30 May 2019 19:26:26 +0000 (12:26 -0700)]
Upgrade cmake_minimum_required(version) to 3.1

Issue #281 (bdwgc).

* CMakeLists.txt (CMAKE_LEGACY_CYGWIN_WIN32): Remove (do not set).
* CMakeLists.txt (VERSION): Change cmake_minimum_required from 2.6
to 3.1.

5 years agoFix 'GC_old_bus_handler defined but not used' compiler warning
Ivan Maidanski [Thu, 30 May 2019 20:57:46 +0000 (23:57 +0300)]
Fix 'GC_old_bus_handler defined but not used' compiler warning

* os_dep.c [!DARWIN && !MSWIN32 && !MSWINCE] (GC_old_bus_handler):
Define variable only if FREEBSD or HPUX, or HURD, or LINUX.
* os_dep.c [!DARWIN && !MSWIN32 && !MSWINCE && (HPUX || HURD
|| FREEBSD)] (GC_dirty_init): Do not set GC_old_bus_handler_used_si
to FALSE (because it is already initialized to FALSE).

5 years agoNew API standalone functions to acquire and release the allocator lock
Ivan Maidanski [Thu, 30 May 2019 19:53:12 +0000 (22:53 +0300)]
New API standalone functions to acquire and release the allocator lock

Issue #282 (bdwgc).

Public GC_alloc_lock() and GC_alloc_unlock() are introduced to simplify
dealing with the GC lock compared to GC_call_with_alloc_lock().

* include/gc.h [GC_THREADS] (GC_alloc_lock, GC_alloc_unlock): Declare
new API function.
* include/gc.h [!GC_THREADS] (GC_alloc_lock, GC_alloc_unlock): Define
as a no-op macro.
* misc.c [THREADS] (GC_alloc_lock, GC_alloc_unlock): Define API
function.
* tests/test.c (reachable_objs_count_enumerator): Remove function.
* tests/test.c (check_heap_stats): Call GC_alloc_lock(); call
GC_enumerate_reachable_objects_inner() directly; call GC_alloc_unlock().

5 years agoRename NO values to OFF in CMake script
Ivan Maidanski [Thu, 30 May 2019 06:55:26 +0000 (09:55 +0300)]
Rename NO values to OFF in CMake script
(code refactoring)

This is to better distinguish them from ON values.

* CMakeLists.txt (enable_threads, enable_gc_debug,
enable_redirect_malloc, enable_mmap, enable_large_config,
enable_gc_assertions, enable_checksums): Change NO to OFF.

5 years agoWorkaround 'checking if unsigned i < 0' cppcheck FP in is_heap_base
Ivan Maidanski [Thu, 30 May 2019 06:17:54 +0000 (09:17 +0300)]
Workaround 'checking if unsigned i < 0' cppcheck FP in is_heap_base

* os_dep.c [MSWIN32 || MSWINCE || CYGWIN32] (GC_is_heap_base): Change
type for i local variable from unsigned to int; cast GC_n_heap_bases to
int.

5 years agoEliminate 'struct member os_callback is never used' cppcheck warning
Ivan Maidanski [Wed, 29 May 2019 20:52:27 +0000 (23:52 +0300)]
Eliminate 'struct member os_callback is never used' cppcheck warning

* os_dep.c [MPROTECT_VDB && DARWIN && CPPCHECK] (GC_dirty_init): Call
GC_noop1(GC_ports.os_callback[0]).

5 years agoFix 'unused function GC_add_map_entry' compiler warning
Ivan Maidanski [Wed, 29 May 2019 20:46:41 +0000 (23:46 +0300)]
Fix 'unused function GC_add_map_entry' compiler warning

* include/private/gc_priv.h (GC_add_map_entry): Declare only if
MARK_BIT_PER_GRANULE.

5 years agoFix 'unused function GC_lock' compiler warning
Ivan Maidanski [Wed, 29 May 2019 08:16:04 +0000 (11:16 +0300)]
Fix 'unused function GC_lock' compiler warning

* include/private/gc_locks.h [GC_PTHREADS && !USE_SPIN_LOCK] (GC_lock):
Declare only if USE_PTHREAD_LOCKS and (GC_ASSERTIONS or
not NO_PTHREAD_TRYLOCK).
* pthread_support.c [!USE_SPIN_LOCK] (GC_lock): Define only if
USE_PTHREAD_LOCKS and (not NO_PTHREAD_TRYLOCK or GC_ASSERTIONS).

5 years agoDefine STATIC macro to static by default
Ivan Maidanski [Mon, 27 May 2019 19:58:41 +0000 (22:58 +0300)]
Define STATIC macro to static by default

This is reduce number of internal symbols exposed outside GC.

* include/private/gcconfig.h [!STATIC] (STATIC): Define to empty only
if GC_ASSERTIONS (regardless of NO_DEBUGGING); update comment.

5 years agoUse lowercase naming of commands in CMake scripts
Ivan Maidanski [Mon, 27 May 2019 19:00:59 +0000 (22:00 +0300)]
Use lowercase naming of commands in CMake scripts
(code refactoring)

* CMakeLists.txt: Convert all commands from uppercase to lowercase.
* tests/CMakeLists.txt: Likewise.
* CMakeLists.txt: Reformat the whole file (indent is 2 spaces).

5 years agoWorkaround spurious gctest exit in incremental mode on Cygwin
Ivan Maidanski [Thu, 23 May 2019 07:47:00 +0000 (10:47 +0300)]
Workaround spurious gctest exit in incremental mode on Cygwin
(fix of commit fa7f7a72b)

* include/private/gcconfig.h [I386 && CYGWIN32 && !USE_WINALLOC]
(MPROTECT_VDB): Do not define; add comment.
* include/private/gcconfig.h [X86_64 && CYGWIN32 && !USE_WINALLOC
&& THREAD_LOCAL_ALLOC] (MPROTECT_VDB): Likewise.

5 years agoUpdate ChangeLog file
Ivan Maidanski [Fri, 17 May 2019 18:38:01 +0000 (21:38 +0300)]
Update ChangeLog file

5 years agoUpdate ChangeLog file (v8.0 changes)
Ivan Maidanski [Fri, 17 May 2019 18:22:38 +0000 (21:22 +0300)]
Update ChangeLog file (v8.0 changes)

5 years agoUpdate ChangeLog file (v7.6 changes)
Ivan Maidanski [Fri, 17 May 2019 17:55:47 +0000 (20:55 +0300)]
Update ChangeLog file (v7.6 changes)

5 years agoUpdate ChangeLog file (v7.2 - v7.4 changes only)
Ivan Maidanski [Fri, 17 May 2019 10:52:55 +0000 (13:52 +0300)]
Update ChangeLog file (v7.2 - v7.4 changes only)

5 years agoUse __data_start instead of _fdata on OpenBSD/mips64
Jasper Lievisse Adriaanse [Thu, 16 May 2019 20:42:49 +0000 (23:42 +0300)]
Use __data_start instead of _fdata on OpenBSD/mips64
(fix of commit d79f09665)

* include/private/gcconfig.h [MIPS && OPENBSD] (__data_start): Declare
extern variable (instead of _fdata).
* include/private/gcconfig.h [MIPS && OPENBSD] (DATASTART): Use
__data_start.

5 years agoUpdate AUTHORS file (add Jasper Lievisse Adriaanse)
Ivan Maidanski [Wed, 15 May 2019 08:03:27 +0000 (11:03 +0300)]
Update AUTHORS file (add Jasper Lievisse Adriaanse)

5 years agoFix word size and alignment for OpenBSD/mips64(el)
Jasper Lievisse Adriaanse [Sat, 27 Jun 2015 13:21:29 +0000 (13:21 +0000)]
Fix word size and alignment for OpenBSD/mips64(el)

* include/private/gcconfig.h [MIPS && OPENBSD] (CPP_WORDSZ): Define
macro; add comment.
* include/private/gcconfig.h [MIPS && OPENBSD] (ALIGNMENT): Define to 8
(instead of 4).

5 years agoUpdate AUTHORS file (add Christian Weisgerber)
Ivan Maidanski [Wed, 15 May 2019 07:30:38 +0000 (10:30 +0300)]
Update AUTHORS file (add Christian Weisgerber)

5 years agoAdd support of OpenBSD/aarch64
Christian Weisgerber [Sat, 15 Apr 2017 14:02:46 +0000 (14:02 +0000)]
Add support of OpenBSD/aarch64

* include/private/gcconfig.h [__aarch64__ && !LINUX && !DARWIN
&& !FREEBSD && !NETBSD] (NOSYS): Do not define if OPENBSD.
* include/private/gcconfig.h [OPENBSD && __aarch64__] (AARCH64,
mach_type_known): Define.
* include/private/gcconfig.h [AARCH64 && OPENBSD] (OS_TYPE, ELF_CLASS,
DATASTART, DATAEND, DYNAMIC_LOADING): Define macro.
* include/private/gcconfig.h [AARCH64 && OPENBSD
&& !GC_OPENBSD_THREADS]: Include sys/param.h, uvm/uvm_extern.h.
* include/private/gcconfig.h [AARCH64 && OPENBSD
&& !GC_OPENBSD_THREADS && USRSTACK] (STACKBOTTOM): Define macro.
* include/private/gcconfig.h [AARCH64 && OPENBSD
&& !GC_OPENBSD_THREADS && !USRSTACK] (HEURISTIC2): Define macro.
* include/private/gcconfig.h [AARCH64 && OPENBSD] (__data_start, _end):
Declare extern variable.

5 years agoUpdate AUTHORS file (add Brian J. Cardiff)
Ivan Maidanski [Mon, 6 May 2019 09:07:49 +0000 (12:07 +0300)]
Update AUTHORS file (add Brian J. Cardiff)

5 years agoAdd API functions to get and set the stack bottom of each thread
Brian J. Cardiff [Mon, 6 May 2019 09:06:12 +0000 (12:06 +0300)]
Add API functions to get and set the stack bottom of each thread

Issue #277 (bdwgc).

This API is useful to support coroutines.

* include/gc.h (GC_get_my_stackbottom, GC_set_stackbottom): New API
function declaration.
* misc.c [!THREADS] (GC_set_stackbottom, GC_get_my_stackbottom): New
function definition.
* pthread_support.c [GC_PTHREADS && !GC_WIN32_THREADS]
(GC_set_stackbottom, GC_get_my_stackbottom): Likewise.
* win32_threads.c [GC_WIN32_THREADS] (GC_set_stackbottom,
GC_get_my_stackbottom): Likewise.
* tests/test.c (struct thr_hndl_sb_s): Define.
* tests/test.c (set_stackbottom): New function (which calls
GC_set_stackbottom).
* tests/test.c (run_one_test): Define thr_hndl_sb local variable;
call GC_get_my_stackbottom() and set_stackbottom().
* win32_threads.c [GC_WIN32_THREADS && I386] (struct GC_Thread_Rep):
Add initial_stack_base field.
* win32_threads.c [GC_WIN32_THREADS && I386] (GC_record_stack_base,
GC_call_with_gc_active): Set initial_stack_base field.
* win32_threads.c [GC_WIN32_THREADS && I386] (GC_push_stack_for): Handle
the case when GetThreadContext() might return stale register values,
thread stack_base != initial_stack_base but the stack is not inside
the TIB stack (use context.Esp but call WARN); add TODO.

5 years agoFix first_thread stack_base initialzation if custom GC_stackbottom (Win32)
Ivan Maidanski [Mon, 29 Apr 2019 06:44:42 +0000 (09:44 +0300)]
Fix first_thread stack_base initialzation if custom GC_stackbottom (Win32)

Stack bottom value of the primordial thread should be obtained from
GC_stackbottom.

* win32_threads.c [GC_ASSERTIONS] (GC_thr_init): Remove sb_result local
variable.
* win32_threads.c (GC_thr_init): Initalize sb mem_base and reg_base
fields with GC_stackbottom and GC_register_stackbottom, respectively;
do not call GC_get_stack_base().

5 years agoReplace 'stack base' with 'stack bottom' in the documentation
Ivan Maidanski [Fri, 26 Apr 2019 16:03:42 +0000 (19:03 +0300)]
Replace 'stack base' with 'stack bottom' in the documentation

* README.md (The C Interface to the Allocator): Outline that stack base
means its bottom; remove a note that client should set GC_stackbottom
on HP PA machines manually.
* doc/README.macros (USE_GET_STACKBASE_FOR_MAIN): Rename "stack base"
to "stack bottom".
* doc/porting.md (Adding Platforms to gcconfig.h): Likewise.
* doc/porting.md (Adding Platforms to gcconfig.h): Rename
"cool end of the stack" to "cold end of the stack"; refine
documentation about STACKBOTTOM.
* include/private/gcconfig.h (STACKBOTTOM): Rename "cool end" to
"cold end" in comment.
* include/gc.h (GC_stackbottom): Likewise.
* include/private/gcconfig.h (LINUX_STACKBOTTOM): Rename "stack base"
to "stack bottom" in comment.
* include/gc.h (GC_stack_base, GC_call_with_stack_base): Likewise.
* include/gc.h [GC_THREADS] (GC_register_my_thread): Likewise.
* misc.c [!THREADS] (GC_call_with_gc_active): Likewise.
* os_dep.c [LINUX_STACKBOTTOM] (GC_linux_main_stack_base): Likewise.
* os_dep.c (GC_get_stack_base): Likewise.
* pthread_support.c (GC_call_with_gc_active): Likewise.
* win32_threads.c (GC_call_with_gc_active): Likewise.

5 years agoFix the stack bottom variable name in README
Ivan Maidanski [Fri, 26 Apr 2019 07:23:16 +0000 (10:23 +0300)]
Fix the stack bottom variable name in README

* README.md (The C Interface to the Allocator): Fix a typo
("GC_stackbottom").

5 years agoRefine do_blocking() documentation in gc.h
Ivan Maidanski [Thu, 25 Apr 2019 07:28:40 +0000 (10:28 +0300)]
Refine do_blocking() documentation in gc.h

* include/gc.h (GC_do_blocking): Refine comment (copy part of comment
from misc.c, add information that GC should be initialized and the
current thread should be registered).
* misc.c (GC_do_blocking): Remove title comment (as it duplicates the
one in gc.h).

5 years agoAdd assertion that GC is initialized to base incremental_protection_needs
Ivan Maidanski [Thu, 25 Apr 2019 07:16:00 +0000 (10:16 +0300)]
Add assertion that GC is initialized to base incremental_protection_needs

* os_dep.c [!HAVE_INCREMENTAL_PROTECTION_NEEDS]
(GC_incremental_protection_needs): Call GC_ASSERT(GC_is_initialized).

5 years agoExplicitly zero-initialize trace_buf
Ivan Maidanski [Thu, 25 Apr 2019 07:04:02 +0000 (10:04 +0300)]
Explicitly zero-initialize trace_buf

* mark.c [TRACE_BUF] (GC_trace_buf): Initialize to all zeros.

5 years agoCheck real-symbols are already initialized in pthread_join/detach
Ivan Maidanski [Thu, 25 Apr 2019 06:54:13 +0000 (09:54 +0300)]
Check real-symbols are already initialized in pthread_join/detach

* pthread_support.c (ASSERT_SYMS_INITIALIZED): New macro (checks either
GC_syms_initialized or parallel_initialized depending on
GC_USE_DLOPEN_WRAP).
* pthread_support.c [!SN_TARGET_ORBIS && !SN_TARGET_PSP2] (pthread_join,
pthread_detach): Call ASSERT_SYMS_INITIALIZED() instead of
INIT_REAL_SYMS().

5 years agoWorkaround 'duplicate expression for condition and assignment' cppcheck FP
Ivan Maidanski [Tue, 23 Apr 2019 19:21:54 +0000 (22:21 +0300)]
Workaround 'duplicate expression for condition and assignment' cppcheck FP

The entitled cppcheck warnings are false positives.

* alloc.c (GC_should_collect): Move reassignment of the global variable
which is checked in the conditional expression to the end of the
conditional block statement.
* finalize.c (GC_notify_or_invoke_finalizers): Likewise.
* mark.c [WRAP_MARK_SOME] (GC_mark_some): Likewise.

5 years agoFix save_callers for multi-threaded case if built-in backtrace unavailable
Ivan Maidanski [Mon, 22 Apr 2019 21:44:52 +0000 (00:44 +0300)]
Fix save_callers for multi-threaded case if built-in backtrace unavailable

GC_stackbottom could be used only in the single-threaded case.

* os_dep.c [SAVE_CALL_CHAIN && (NARGS!=0 || NFRAMES%2!=0
|| !GC_HAVE_BUILTIN_BACKTRACE) && THREADS] (GC_save_callers): Do not
compare fp to GC_stackbottom.
* os_dep.c [SAVE_CALL_CHAIN && (NARGS!=0 || NFRAMES%2!=0
|| !GC_HAVE_BUILTIN_BACKTRACE) && THREADS && STACK_GROWS_UP]
(GC_save_callers): Break the loop if fp is NULL.

5 years agoRemove Win32 main_thread static variable if threads discovery is disabled
Ivan Maidanski [Fri, 19 Apr 2019 05:07:52 +0000 (08:07 +0300)]
Remove Win32 main_thread static variable if threads discovery is disabled
(code refactoring)

* win32_threads.c (GC_main_thread): Do not define static variable if
GC_NO_THREADS_DISCOVERY.
* win32_threads.c [!GC_NO_THREADS_DISCOVERY] (GC_attached_thread):
Reformat comment; move the definition to be near GC_please_stop.
* win32_threads.c [GC_NO_THREADS_DISCOVERY] (GC_thr_init): Define
GC_main_thread as a macro (and undefine it at the end of the function).

5 years agoSimplify logged message in scratch_recycle
Ivan Maidanski [Wed, 17 Apr 2019 19:43:16 +0000 (22:43 +0300)]
Simplify logged message in scratch_recycle

* mark.c (GC_scratch_recycle_inner): Do not pass bytes variable to
GC_COND_LOG_PRINTF().

5 years agoEnable true incremental collection even if parallel marker is on
Ivan Maidanski [Tue, 16 Apr 2019 21:52:59 +0000 (00:52 +0300)]
Enable true incremental collection even if parallel marker is on

Issue #151 (bdwgc).

Because of the current limitation of the parallel marker implementation,
it is not possible to interrupt the collection when performed by the
parallel marker.  This change allows to have the true incremental mode
at the expense of disabling the parallel marker during most collection
phases.  By default, the old behavior (a generational collection with
the parallel marker enabled) is preserved unless the client sets
GC_time_limit to a value other than GC_TIME_UNLIMITED.

* alloc.c [(!GC_TIME_LIMIT || CPPCHECK) && PARALLEL_MARK]
(GC_time_limit): Set to GC_TIME_UNLIMITED; add comment.
* alloc.c [PARALLEL_MARK] (GC_collect_a_little_inner): Temporarily
set GC_parallel_mark_disabled to TRUE before GC_mark_some repeated
invocation if GC_time_limit is not GC_TIME_UNLIMITED.
* alloc.c [PARALLEL_MARK] (GC_stopped_mark): Temporarily set
GC_parallel_mark_disabled to TRUE before GC_mark_some repeated
invocation if stop_func is not GC_never_stop_func; add verbose logging
if parallel marking is disabled temporarily.
* doc/README.environment (GC_PAUSE_TIME_TARGET): Update the description
(remove the limitation for the case when parallel marking is on).
* doc/scale.md (The Parallel Marking Algorithm): Update the
documentation regarding incremental mode.
* include/gc.h (GC_parallel, GC_enable_incremental): Update the comment
(remove the limitation on the incremental mode when parallel marking
is on).
* include/private/gc_priv.h [PARALLEL_MARK] (GC_parallel_mark_disabled):
Declare global variable.
* mark.c [PARALLEL_MARK] (GC_parallel_mark_disabled): Define.
* mark.c [PARALLEL_MARK] (GC_mark_some_inner): Do not call
GC_do_parallel_mark() if GC_parallel_mark_disabled; update comment.
* pthread_support.c [PARALLEL_MARK] (GC_thr_init): Do not set
GC_time_limit to GC_TIME_UNLIMITED if available_markers_m1 > 0; remove
comment.
* win32_threads.c [PARALLEL_MARK] (GC_thr_init): Likewise.

5 years agoComments reformatting in mark.c to properly delimit sentences
Ivan Maidanski [Tue, 16 Apr 2019 07:19:53 +0000 (10:19 +0300)]
Comments reformatting in mark.c to properly delimit sentences
(code refactoring)

* mark.c (GC_clear_hdr_marks, clear_marks_for_block, GC_set_mark_bit,
GC_clear_marks, GC_initiate_gc, GC_mark_from, alloc_mark_stack,
GC_push_all, GC_push_all_eager, GC_push_marked1, GC_push_marked2,
GC_push_marked4): Add training dot to comment sentences (where missing);
capitalize the first letter of comment sentences (where needed);
reformat some comments.
* mark.c [WRAP_MARK_SOME] (GC_mark_some): Likewise.
* mark.c [PARALLEL_MARK] (GC_do_parallel_mark, GC_help_wanted,
GC_helper_count, GC_active_count, GC_steal_mark_stack, GC_mark_local):
Likewise.
* mark.c [!GC_DISABLE_INCREMENTAL] (GC_push_next_marked_dirty):
Likewise.

5 years agoSimplify loops of collect_a_little/stopped_mark invoking mark_some
Ivan Maidanski [Tue, 16 Apr 2019 06:29:42 +0000 (09:29 +0300)]
Simplify loops of collect_a_little/stopped_mark invoking mark_some
(code refactoring)

* alloc.c (GC_collect_a_little_inner, GC_stopped_mark): If
GC_mark_some() returns true then just break to quit the loop; move the
logic that is needed to funish the collection outside the loop.
* alloc.c (GC_stopped_mark): Change i local variable type from unsigned
to int.

5 years agoAllow GC_PAUSE_TIME_TARGET environment variable values smaller than 5 ms
Ivan Maidanski [Fri, 12 Apr 2019 17:25:47 +0000 (20:25 +0300)]
Allow GC_PAUSE_TIME_TARGET environment variable values smaller than 5 ms

All time limit values starting from 1 ms are now accepted.

* misc.c [!GC_DISABLE_INCREMENTAL] (GC_init): Do not set GC_time_limit
if NO_CLOCK; allow all non-zero time_limit values (instead of >4);
do not call WARN() if time_limit value is incorrect.

5 years agoRefine flags field comment in pthread_support.h
Ivan Maidanski [Fri, 12 Apr 2019 16:51:02 +0000 (19:51 +0300)]
Refine flags field comment in pthread_support.h

* include/private/pthread_support.h (GC_Thread_Rep.flags): Add comment
that the field is accessed only by the GC lock owner.

5 years agoFix 'USE_MUNMAP requires USE_MMAP' compilation error on Cygwin
Ivan Maidanski [Wed, 10 Apr 2019 20:42:11 +0000 (23:42 +0300)]
Fix 'USE_MUNMAP requires USE_MMAP' compilation error on Cygwin
(fix of commit fdb655334)

* include/private/gcconfig.h [CYGWIN32] (MMAP_SUPPORTED): Do not define
if USE_WINALLOC.

5 years agoFix MPROTECT_VDB definition for single-threaded GC builds
Ivan Maidanski [Wed, 10 Apr 2019 20:33:39 +0000 (23:33 +0300)]
Fix MPROTECT_VDB definition for single-threaded GC builds
(fix of commit 6d9eec852)

* include/private/gcconfig.h [MPROTECT_VDB && !MSWIN32 && !MSWINCE]:
Include signal.h.

5 years agoFix 'un-mprotect failed' gctest error on Cygwin
Ivan Maidanski [Wed, 10 Apr 2019 20:25:50 +0000 (23:25 +0300)]
Fix 'un-mprotect failed' gctest error on Cygwin
(fix of commit fa7f7a72b)

Issue #272 (bdwgc).

* include/private/gcconfig.h [(I386 || X86_64) && CYGWIN32]
(MPROTECT_VDB): Do not define if USE_WINALLOC.
* os_dep.c [CYGWIN32 && MPROTECT_VDB] (GC_setpagesize): Set
GC_page_size to dwAllocationGranularity.

5 years agoTravis CI: Add a test of configure --enable-mmap (Linux)
Ivan Maidanski [Tue, 9 Apr 2019 21:49:54 +0000 (00:49 +0300)]
Travis CI: Add a test of configure --enable-mmap (Linux)

5 years agoAdjust macro def/usage for AVR32, CRIS, NETBSD, OPENBSD, SH4 in gcconfig.h
Ivan Maidanski [Tue, 9 Apr 2019 21:42:23 +0000 (00:42 +0300)]
Adjust macro def/usage for AVR32, CRIS, NETBSD, OPENBSD, SH4 in gcconfig.h
(code refactoring)

* include/private/gcconfig.h [__vax__ || __MIPSEL__ || __sparc__
|| i386 || __i386__ || __x86_64__ || sparc && unix && !sun && !linux]:
Replace __NetBSD__ to NETBSD.
* include/private/gcconfig.h [sparc && unix && !sun && !linux]: Replace
__FreeBSD__ || __DragonFly__ to FREEBSD.
* include/private/gcconfig.h [sparc && unix && !sun && !linux
|| __powerpc__ || i386 || __i386__]: Replace __OpenBSD__ to OPENBSD;
do not redefine OPENBSD macro.
* include/private/gcconfig.h [CRIS] (OS_TYPE, DYNAMIC_LOADING,
LINUX_STACKBOTTOM, SEARCH_FOR_DATA_START, DATAEND): Define only if
LINUX.
* include/private/gcconfig.h [AVR32] (OS_TYPE, DYNAMIC_LOADING,
LINUX_STACKBOTTOM, SEARCH_FOR_DATA_START, DATAEND): Likewise.
* include/private/gcconfig.h [SH4] (OS_TYPE, DATAEND): Define only if
MSWINCE.

5 years agoMove QNX and Emscripten macro definitions to proper place in gcconfig.h
Ivan Maidanski [Tue, 9 Apr 2019 19:50:09 +0000 (22:50 +0300)]
Move QNX and Emscripten macro definitions to proper place in gcconfig.h
(code refactoring)

* include/private/gcconfig.h [__QNX__] (I386, mach_type_known): Move
macro definition down (to be in a group where mach_type_known is
defined for most platforms).
* include/private/gcconfig.h [__QNX__] (OS_TYPE, SA_RESTART,
HEURISTIC1, DATASTART, DATAEND): Move macro definition down; define
only if I386.
* include/private/gcconfig.h [__EMSCRIPTEN__] (OS_TYPE, CPP_WORDSZ,
ALIGNMENT, DATASTART, DATAEND, STACK_NOT_SCANNED): Likewise.

5 years agoEnable sbrk-to-mmap fallback on major supported Unix-like platforms
Ivan Maidanski [Tue, 9 Apr 2019 06:38:44 +0000 (09:38 +0300)]
Enable sbrk-to-mmap fallback on major supported Unix-like platforms

This only affects --disable-munmap configuration.

* include/private/gcconfig.h [CYGWIN32 || DARWIN || HAIKU || NETBSD
|| OPENBSD] (MMAP_SUPPORTED): Define macro; add comment.

5 years agoUpdate ChangeLog file
Ivan Maidanski [Sat, 6 Apr 2019 18:07:19 +0000 (21:07 +0300)]
Update ChangeLog file

5 years agoUpdate ChangeLog file (v8.0 changes)
Ivan Maidanski [Sat, 6 Apr 2019 17:58:10 +0000 (20:58 +0300)]
Update ChangeLog file (v8.0 changes)

5 years agoUpdate ChangeLog file (v7.6 changes)
Ivan Maidanski [Sat, 6 Apr 2019 07:05:33 +0000 (10:05 +0300)]
Update ChangeLog file (v7.6 changes)

5 years agoUpdate ChangeLog file (v7.2 - v7.4 changes only)
Ivan Maidanski [Sat, 6 Apr 2019 07:01:41 +0000 (10:01 +0300)]
Update ChangeLog file (v7.2 - v7.4 changes only)

5 years agoAppveyor/Travis CI: Test with GC time limit set to 10 ms
Ivan Maidanski [Fri, 5 Apr 2019 18:27:20 +0000 (21:27 +0300)]
Appveyor/Travis CI: Test with GC time limit set to 10 ms

5 years agoFix typos in documentation (regarding 'non-incremental')
Ivan Maidanski [Thu, 4 Apr 2019 21:47:37 +0000 (00:47 +0300)]
Fix typos in documentation (regarding 'non-incremental')

* doc/gcdescr.md (Introduction): Fix typos ("non-incremental",
"this operation").
* mark.c (GC_collection_in_progress): Fix a typo in comment
("non-incremental").

5 years agoUpdate documentation about the incremental collector support
Ivan Maidanski [Thu, 4 Apr 2019 21:34:04 +0000 (00:34 +0300)]
Update documentation about the incremental collector support

* doc/README.darwin (Important Usage Notes): Remove outdated note about
flaky incremental mode.
* doc/README.darwin (Implementation Information): Remove information
about old incremental collector implementation.
* doc/README.win32 (Watcom compiler): Refine information about the
incremental mode support.
* doc/scale.md (Options for enhanced scalability): Refine when the
dedicated marker threads are created.

5 years agoRefine GC_parallel documentation in gc.h
Ivan Maidanski [Wed, 3 Apr 2019 08:09:37 +0000 (11:09 +0300)]
Refine GC_parallel documentation in gc.h

* include/gc.h [GC_THREADS] (GC_parallel): Refine comment about the
case when it is non-zero.

5 years agoFix a typo (GC_NPROCS spelling) in a comment in gc.h
Ivan Maidanski [Wed, 3 Apr 2019 08:00:00 +0000 (11:00 +0300)]
Fix a typo (GC_NPROCS spelling) in a comment in gc.h

* include/gc.h [GC_THREADS] (GC_parallel): Fix a typo in comment
("GC_NPROCS").

5 years agoRemove a misleading comment about Solaris in gc.h
Ivan Maidanski [Wed, 3 Apr 2019 07:49:59 +0000 (10:49 +0300)]
Remove a misleading comment about Solaris in gc.h

* include/gc.h: Remove a comment about Solaris.

5 years agoAllow memory unmapping in case of MPROTECT_VDB
Ivan Maidanski [Wed, 3 Apr 2019 06:47:14 +0000 (09:47 +0300)]
Allow memory unmapping in case of MPROTECT_VDB

Issue #271 (bdwgc).

* allchblk.c [USE_MUNMAP && MPROTECT_VDB] (GC_has_unmapped_memory):
Remove function definition.
* os_dep.c [MPROTECT_VDB && USE_MUNMAP] (GC_dirty_init): Likewise.
* doc/README.macros (USE_MUNMAP): Update the description.
* include/private/gc_priv.h [MPROTECT_VDB && USE_MUNMAP]
(GC_mprotect_dirty_init, GC_has_unmapped_memory): Remove prototype.
* include/private/gcconfig.h [USE_MUNMAP && GWW_VDB] (MPROTECT_VDB):
Do not undefine; remove TODO; remove comment.
* include/private/gcconfig.h: Remove comment that PARALLEL_MARK does
not cause undefine MPROTECT_VDB.
* os_dep.c [MPROTECT_VDB && !USE_MUNMAP] (GC_mprotect_dirty_init):
Remove macro.
* os_dep.c [MPROTECT_VDB] (GC_mprotect_dirty_init): Rename to
GC_dirty_init.
* tests/test.c [!GC_DISABLE_INCREMENTAL && (TEST_DEFAULT_VDB
|| !DEFAULT_VDB)] (main): Call GC_enable_incremental() even in case of
both MPROTECT_VDB and USE_MUNMAP are defined.

5 years agoAppveyor CI: Do not use GC_PREFER_MPROTECT_VDB together with USE_MUNMAP
Ivan Maidanski [Wed, 3 Apr 2019 06:12:11 +0000 (09:12 +0300)]
Appveyor CI: Do not use GC_PREFER_MPROTECT_VDB together with USE_MUNMAP
(fix of commit b7d83aaf6)

5 years agoAllow incremental GC on Cygwin
Ivan Maidanski [Tue, 2 Apr 2019 22:00:57 +0000 (01:00 +0300)]
Allow incremental GC on Cygwin

Both GWW- and mprotect-based implementations should work. The first one
works only if the collector is compiled with -D USE_WINALLOC passed to
CFLAGS.

* include/private/gcconfig.h [(I386 || X86_64) && CYGWIN32]
(MPROTECT_VDB): Define.
* include/private/gcconfig.h [(I386 || X86_64) && CYGWIN32
&& USE_WINALLOC] (GWW_VDB): Define.
* include/private/gcconfig.h [(I386 || X86_64) && CYGWIN32 && USE_MMAP]
(NEED_FIND_LIMIT, USE_MMAP_ANON): Do not define if USE_WINALLOC.
* include/private/gcconfig.h [GWW_VDB && !USE_WINALLOC && !CPPCHECK]:
Produce an invalid config error.

5 years agoRemove unused STACK_GRAN macro definitions
Ivan Maidanski [Tue, 2 Apr 2019 20:37:33 +0000 (23:37 +0300)]
Remove unused STACK_GRAN macro definitions
(code refactoring)

* include/private/gcconfig.h [I386 && CYGWIN32] (STACK_GRAN): Remove
macro redefinition.
* include/private/gcconfig.h [(ARM32 || M32R) && LINUX
&& LINUX_STACKBOTTOM] (STACK_GRAN): Likewise.
* os_dep.c [!STACKBOTTOM] (GC_get_main_stack_base): Define
STACKBOTTOM_ALIGNMENT_M1 macro only if HEURISTIC1.

5 years agoFix OS_TYPE and USE_MMAP_ANON definitions for Cygwin/x64
Ivan Maidanski [Tue, 2 Apr 2019 20:11:41 +0000 (23:11 +0300)]
Fix OS_TYPE and USE_MMAP_ANON definitions for Cygwin/x64
(fix of commit 6c535e8d6)

* include/private/gcconfig.h [IA64 && CYGWIN32] (OS_TYPE, DATASTART,
DATAEND, STACK_GRAN, NEED_FIND_LIMIT, USE_MMAP_ANON): Do not define.
* include/private/gcconfig.h [X86_64 && CYGWIN32] (OS_TYPE): Define.
* include/private/gcconfig.h [X86_64 && CYGWIN32 && USE_MMAP]
(USE_MMAP_ANON): Define.

5 years agoAvoid rare 'Unexpected heap growth' in multi-threaded gctest on Cygwin/x64
Ivan Maidanski [Tue, 2 Apr 2019 19:59:54 +0000 (22:59 +0300)]
Avoid rare 'Unexpected heap growth' in multi-threaded gctest on Cygwin/x64

* tests/test.c [!VERY_SMALL_CONFIG && CPP_WORDSZ==64]
(check_heap_stats): Change 25000000 (max_heap_sz value) to 26000000.

5 years agoAppveyor CI: Test MinGW with -D GC_PREFER_MPROTECT_VDB
Ivan Maidanski [Tue, 2 Apr 2019 08:53:23 +0000 (11:53 +0300)]
Appveyor CI: Test MinGW with -D GC_PREFER_MPROTECT_VDB

5 years agoAvoid compiler warning about unused d in GC_CALLOC/MALLOC_EXPLICITLY_TYPED
Ivan Maidanski [Mon, 1 Apr 2019 19:21:54 +0000 (22:21 +0300)]
Avoid compiler warning about unused d in GC_CALLOC/MALLOC_EXPLICITLY_TYPED

Even if GC_DEBUG is defined, "d" argument of GC_CALLOC_EXPLICITLY_TYPED
and GC_MALLOC_EXPLICITLY_TYPED should be used anyway.

* include/gc_typed.h [GC_DEBUG] (GC_MALLOC_EXPLICITLY_TYPED,
GC_CALLOC_EXPLICITLY_TYPED): Cast d to void.

5 years agoAdd testing of GC_CALLOC/MALLOC_EXPLICITLY_TYPED (gctest)
Ivan Maidanski [Mon, 1 Apr 2019 19:13:18 +0000 (22:13 +0300)]
Add testing of GC_CALLOC/MALLOC_EXPLICITLY_TYPED (gctest)

* tests/test.c (typed_test): Call GC_MALLOC_EXPLICITLY_TYPED()
instead of GC_malloc_explicitly_typed(); call
GC_CALLOC_EXPLICITLY_TYPED() instead of GC_calloc_explicitly_typed().

5 years agoUpdate ChangeLog file
Ivan Maidanski [Wed, 27 Mar 2019 18:29:19 +0000 (21:29 +0300)]
Update ChangeLog file

5 years agoUpdate ChangeLog file (v8.0 changes)
Ivan Maidanski [Wed, 27 Mar 2019 18:13:55 +0000 (21:13 +0300)]
Update ChangeLog file (v8.0 changes)
(a cherry-pick of commit f74ff85f from 'release-8_0')

5 years agoUpdate ChangeLog file (v7.6 changes)
Ivan Maidanski [Wed, 27 Mar 2019 17:59:20 +0000 (20:59 +0300)]
Update ChangeLog file (v7.6 changes)
(a cherry-pick of commit 3b645a40 from 'release-7_6')

5 years agoUpdate ChangeLog file (v7.2 - v7.4 changes only)
Ivan Maidanski [Wed, 27 Mar 2019 17:44:00 +0000 (20:44 +0300)]
Update ChangeLog file (v7.2 - v7.4 changes only)
(a cherry-pick of commit 0c7bb67e from 'release-7_4')

5 years agoAdjust highlighting of API prototypes in gcinterface.md
Ivan Maidanski [Tue, 26 Mar 2019 21:31:25 +0000 (00:31 +0300)]
Adjust highlighting of API prototypes in gcinterface.md

* doc/gcinterface.md: Properly highlight listed API functions
(GC_MALLOC, GC_MALLOC_ATOMIC, GC_MALLOC_UNCOLLECTABLE, GC_REALLOC,
GC_FREE, GC_MALLOC_IGNORE_OFF_PAGE, GC_MALLOC_ATOMIC_IGNORE_OFF_PAGE,
GC_INIT, GC_gcollect, GC_enable_incremental, GC_set_warn_proc,
GC_REGISTER_FINALIZER).

5 years agoFix a typo in gcinterface.md after recent documentation update
Ivan Maidanski [Tue, 26 Mar 2019 20:28:18 +0000 (23:28 +0300)]
Fix a typo in gcinterface.md after recent documentation update
(fix of commit 6f8f39af5)

* doc/gcinterface.md (GC_MALLOC_IGNORE_OFF_PAGE): Fix a typo
("inheritance").

5 years agoFix printf format specifier in simple_example.md
Ivan Maidanski [Tue, 26 Mar 2019 20:09:49 +0000 (23:09 +0300)]
Fix printf format specifier in simple_example.md

* doc/simple_example.md (Writing the program): Adjust printf format
specifier to match the argument type.

5 years agoReformat gc.man (wrap long lines)
Ivan Maidanski [Tue, 26 Mar 2019 06:27:22 +0000 (09:27 +0300)]
Reformat gc.man (wrap long lines)

* doc/gc.man: Wrap lines longer than 78 characters.

5 years agoUpdate the documentation to match the current GC implementation
Ivan Maidanski [Tue, 26 Mar 2019 05:37:11 +0000 (08:37 +0300)]
Update the documentation to match the current GC implementation

* README.md: Update the documentation to match the current
implementation of the collector.
* doc/README.Mac: Likewise.
* doc/README.autoconf: Likewise.
* doc/README.darwin: Likewise.
* doc/README.hp: Likewise.
* doc/README.linux: Likewise.
* doc/README.macros: Likewise.
* doc/README.solaris2: Likewise.
* doc/README.win32: Likewise.
* doc/debugging.md: Likewise.
* doc/finalization.md: Likewise.
* doc/gc.man: Likewise.
* doc/gcdescr.md: Likewise.
* doc/gcinterface.md: Likewise.
* doc/leak.md: Likewise.
* doc/overview.md: Likewise.
* doc/porting.md: Likewise.
* doc/scale.md: Likewise.

5 years agoAdjust formatting of numbered lists in README.md to match other .md files
Ivan Maidanski [Mon, 25 Mar 2019 15:50:26 +0000 (18:50 +0300)]
Adjust formatting of numbered lists in README.md to match other .md files

* README.md (Installation and Portability,
The C Interface to the Allocator, Incremental/Generational Collection):
Reformat numbered lists.