Revert "Revert "Revert "[ruby/fiddle] Use ffi_closure_free by
default. (#20)"""
This reverts commit 87f6154bb4.
It turned out that the change fails to build on macOS
https://rubyci.org/logs/rubyci.s3.amazonaws.com/osx1014/ruby-master/log/20200304T074503Z.fail.html.gz
```
+ make 'TESTS=--hide-skip -v fiddle' RUBYOPT=-w test-all
dyld: lazy symbol binding failed: Symbol not found: _ffi_closure_alloc
Referenced from: /Users/hsbt/Documents/cb/tmp/build/20200304T074503Z/ruby/.ext/x86_64-darwin18/fiddle.bundle
Expected in: flat namespace
dyld: Symbol not found: _ffi_closure_alloc
Referenced from: /Users/hsbt/Documents/cb/tmp/build/20200304T074503Z/ruby/.ext/x86_64-darwin18/fiddle.bundle
Expected in: flat namespace
make: *** [yes-test-all] Abort trap: 6
```
[ruby/fiddle] Use ffi_closure_free if available
[ruby/fiddle] ffi_closure_free is available in the bundled libffi
[ruby/fiddle] use ffi_closure_alloc only with 3.2 or later
[ruby/fiddle] always use ffi_closure_alloc on Windows
Fixed a typo
Show libffi version only if set
ext/fiddle/extconf.rb: check if ffi_closure_alloc is available
to define HAVE_FFI_CLOSURE_ALLOC.
The macro is used in closure.c, so have_func check is needed.
If pkg-config is not installed, extconf.rb fails to detect the version
of libffi, and does not add "-DUSE_FFI_CLOSURE_ALLOC=1" even when system
libffi version is >= 3.2.
If USE_FFI_CLOSURE_ALLOC is not defined, closure.c attempts to check if
HAVE_FFI_CLOSURE_ALLOC is defined or not, but have_func was removed with
528a3a1797, so the macro is always not
defined.
This resulted in this deprecation warning:
https://rubyci.org/logs/rubyci.s3.amazonaws.com/ubuntu2004/ruby-master/log/20200512T123003Z.log.html.gz
```
compiling closure.c
closure.c: In function 'initialize':
closure.c:265:5: warning: 'ffi_prep_closure' is deprecated: use ffi_prep_closure_loc instead [-Wdeprecated-declarations]
265 | result = ffi_prep_closure(pcl, cif, callback, (void *)self);
| ^~~~~~
In file included from ./fiddle.h:42,
from closure.c:1:
/usr/include/x86_64-linux-gnu/ffi.h:334:1: note: declared here
334 | ffi_prep_closure (ffi_closure*,
| ^~~~~~~~~~~~~~~~
```
Do not try ffi_closure_alloc if libffi is <= 3.1
Maybe due to e1855100e4, CentOS, RHEL, and
Fedora CIs have started failing with SEGV. Try to avoid
ffi_closure_alloc on those environments.
https://rubyci.org/logs/rubyci.s3.amazonaws.com/centos8/ruby-master/log/20200512T183004Z.fail.html.gzhttps://rubyci.org/logs/rubyci.s3.amazonaws.com/fedora32/ruby-master/log/20200512T183004Z.fail.html.gzhttps://rubyci.org/logs/rubyci.s3.amazonaws.com/rhel8/ruby-master/log/20200512T183003Z.fail.html.gz
ext/fiddle/extconf.rb: Fix the condition of libffi <= 3.1
ver is [3, 1, 0] which is not less then or equal to [3, 1]
* Use ffi_closure_free unconditionally.
The current conditionals reflect historic heritage of FFI. Usage of
ffi_closure_free should be better default nowadays, because libffi 3.0.5
fixing issues of ffi_closure_free should be widely available.
* RUBY_LIBFFI_MODVERSION is not used anymore.
Because `ffi_closure_free()` is not used unconditionally, there is no
other use for RUBY_LIBFFI_MODVERSION define, so drop its usage.
* Use more meaningful variable name.
`ver` variable used to be used to pupulate RUBY_LIBFFI_MODVERSION
define. Since the define was removed, the `libffi_dir` variable name
should better describe the remaining usage of the variable.
https://github.com/ruby/fiddle/commit/c49cc79eb8
If the first parameter to Fiddle::Function is a closure object (rather
than an interger), `rb_Integer` will cast it to an integer but not
maintain a reference to the closure. Then if the closure gets GC'd, we
have a segv. This commit keeps a reference to the original parameter to
initialize so that the object will not be GC'd.
Fixes: https://bugs.ruby-lang.org/issues/13286https://github.com/ruby/fiddle/commit/0fc697bbc5
Moving public headers was 12-years ago, no depend files would
expect ruby.h in the top source directory now.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67033 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/fiddle/extconf.rb: It supports to build libffi with standalone gem.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65210 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/fiddle/function.c (initialize): use RARRAY_AREF() instead of
using RARRAY_PTR().
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64989 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Fixes: r62450 (commit 1aaeeb326e)
("long long is a C99ism")
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@62453 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
so SIZEOF_LONG_LONG is not always available.
We have to check its defined?-ness before using.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@62450 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Added following gemspecs.
* extensions: date, dbm, etc, fiddle, gdbm, sdbm, stringio, strscan, zlib
* pure ruby libraries: cmath, csv, fileutils, scanf, webrick
psych and rdoc is out of scope of this commit. I will merge after
upstream was change to `frozen_string_literal: true`.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59115 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Onigumo 6 (r57045) introduced new onigumo.h header file, which is
required from quite much everywhere. This commit adds necessary
dependencies.
Note: ruby/oniguruma.h now includes onigumo.h,
ruby/io.h includes oniguruma.h,
ruby/encoding.h also includes oniguruma.h,
and internal.h includes encoding.h.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58054 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/fiddle/depend (build-libffi): get rid of making $(LIBFFI_A)
a sole target, which may be empty when installed libffi is
found.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@57601 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/fiddle/extconf.rb: fix $(SUBMAKE_PRE) to chdir to
$(LIBFFI_DIR) instead of $(@D), since $(LIBFFI_A) is not
underneath libffi but under .libs.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@57593 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/fiddle/depend (libffi): use $(MAKE) explicitly for recursive
make, so that jobserver works.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@57583 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Patch provided by tenderlove and nobu.
* test/fiddle/test_handle.rb (class TestHandle): add test for above.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@53153 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/fiddle/function.c (initialize): fix typo "ary" to "args",
and adjust type of variables to suppress warnings.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@53090 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/fiddle/function.c (initialize): check all arguments first.
reported by Marcin 'Icewall' Noga of Cisco Talos.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@53082 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/fiddle/conversions.h (PTR2NUM): use signed integer to make
Fixnum for negative values.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@53081 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Some external functions I wish to call may take a long time
and unnecessarily block other threads. This may lead to performance
regressions for fast functions as releasing/acquiring the GVL is not
cheap, but can improve performance for long-running functions
in multi-threaded applications.
This also means we must reacquire the GVL when calling Ruby-defined
callbacks for Fiddle::Closure, meaning we must detect whether the
current thread has the GVL by exporting ruby_thread_has_gvl_p
in internal.h
* ext/fiddle/function.c (struct nogvl_ffi_call_args):
new struct for GVL release
(nogvl_ffi_call): new function
(function_call): adjust for GVL release
[ruby-core:71642] [Feature #11607]
* ext/fiddle/closure.c (struct callback_args):
new struct for GVL acquire
(with_gvl_callback): adjusted original callback function
(callback): wrapper for conditional GVL acquire
* ext/fiddle/depend: add dependencies
* ext/fiddle/extconf.rb: include top_srcdir for internal.h
* internal.h (ruby_thread_has_gvl_p): expose for fiddle
* vm_core.h (ruby_thread_has_gvl_p): moved to internal.h
* test/fiddle/test_function.rb (test_nogvl_poll): new test
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52723 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/bigdecimal/bigdecimal.c: use RARRAY_CONST_PTR just fore
reference instead of RARRAY_PTR, to keep the array WB-protected.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52448 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext: use long for index instead of int and RARRAY_LENINT.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52447 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext: use rb_check_arity and rb_error_arity to raise
ArgumentError. [Feature #9025]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52275 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
The `callback' function is not used outside of closure.c and
should not be able to cause namespace conflicts in compilers
without visibility pragmas.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52198 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/fiddle/extconf.rb: revert part of r51401 and expand
$LOCAL_LIBS as it will be extracted and copied to exts.mk.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51429 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
NAKAMURA Usaku