ruby-core

Hi,

On Wed, Aug 21, 2002 at 02:00:17PM -0400, James F. Hranicky wrote:
> 
> How's ThreadSafeRuby coming along?
> 

	My idea was to bring real threads to Ruby (POSIX, Windows, or SW
	emul (as it is now)).

	For native threads, you definitely need some locking.

	The problem is what to lock.
	====================================
	My idea was to add RW-Lock into every ruby object (to RBasic
	struct).
	
	So: Every binding method should call rb_lock_wr, or
	rb_lock_rd. I think that we cannot use POSIX rwlock, because we
	need reentrant write lock for the same thread (or rewrite
	methods not to call each other that could be definitely a big
	pain) - so we have to write our own RWlock)

	As I calculated the space needed for this new lock, it could add
	about 5 pointers size to every Ruby object, that means: doubles
	memory needs! (Current std. Ruby object is about 5 pointers as
	well) I'm not sure wether it is acceptable.

	Also this could bring a lot of deadlocks and I'm not sure if it
	is suitable for eval and parsing part.
	====================================

	If you will have just 1 global lock, I don't see any benefit of
	bringing native threads in.
	
	So the project has stopped because of lack of discussion.

		Michal

-- 
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Michal Rokos                         Czech Technical University, Prague
E-mail:m.rokos@sh.cvut.cz      ICQ:36118339      Jabber:majkl@jabber.cz
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Thread

Prev Next

In This Thread

Prev Next

[#290] — Florian Frank <flori@...>

[#291] Kernel Thread implementation — Matthew Bloch <mattbee@...>

[#297] GC longjmp macros — Michal Rokos <m.rokos@...>

[#298] Is there a bug in Matrix or Complex? — Jim Freeze <jim@...>

[#308] Q: OSSL in std. distr? — Michal Rokos <m.rokos@...>

[#309] Progress on incremental GC? — "Timothy J. Wood" <tjw@...>

[#314] resolv.rb on 1.6.7 — Dave Thomas <Dave@...>

[#316] Embeding Ruby in C++ code. — Kostas Katsamakas <kmak@...>

[#318] C and Ruby... — Kostas Katsamakas <kmak@...>

[#321] Ruby module, Warn(ing) interface — Michal Rokos <m.rokos@...>

[#326] Implications of a #force_free method in Object? — Matthew Bloch <mattbee@...>

[#328] Int vs Long — Michal Rokos <m.rokos@...>

[#329] Int vs Long #2 — Michal Rokos <m.rokos@...>

[#337] Int vs Long (2nd part) — Michal Rokos <m.rokos@...>

[#340] Int vs Long #3 — Michal Rokos <m.rokos@...>

[#345] Bug in gc.c or misunderstanding on my part? — "Timothy J. Wood" <tjw@...>

[#349] Array#select — Michal Rokos <m.rokos@...>

[#353] File (struct stat handling) — Michal Rokos <m.rokos@...>

[#355] Array and .*ALLOC.* changes — Michal Rokos <m.rokos@...>

[#358] node.h for eval.c — Michal Rokos <m.rokos@...>

[#363] Ruby embedding limitations — "Timothy J. Wood" <tjw@...>

[#367] RE: Welcome to our (ruby-core ML) You are added automatically — Girish Sharangpani <girishs@...>

[#372] rb_class_path — Michal Rokos <m.rokos@...>

[#376] Virtual host and port number in CGI — Raymond Blum <raymond@...>

[#382] Port match to new dup, clone framework — Michal Rokos <m.rokos@...>

[#387] rb_str_replace — Michal Rokos <m.rokos@...>

[#392] io.c read_all — Michal Rokos <m.rokos@...>

[#393] in dln.c — Michal Rokos <m.rokos@...>

[#394] in file.c — Michal Rokos <m.rokos@...>

[#395] CVS version and win32 naming — "Rich Kilmer" <rich@...>

Re: [Cleanup?] Int vs Long

Thread

In This Thread

[#355] Array and .ALLOC. changes — Michal Rokos <m.rokos@...>