[#13842] Better introspection for Frames, Thread, and enhancing binding. — "Rocky Bernstein" <rocky.bernstein@...>

The below is a little long. So here's a summary.

11 messages 2007/12/01

[#13851] Array#flatten works quadratic time on length of resulting array. It could be linear — "Voroztsov Artem" <artem.voroztsov@...>

I encountered problem with Array#flatten slowness (it can be much

19 messages 2007/12/03
[#13863] Re: Array#flatten works quadratic time on length of resulting array. It could be linear — Charles Oliver Nutter <charles.nutter@...> 2007/12/03

Voroztsov Artem wrote:

[#13867] Re: Array#flatten works quadratic time on length of resulting array. It could be linear — "Voroztsov Artem" <artem.voroztsov@...> 2007/12/03

2007/12/3, Charles Oliver Nutter <charles.nutter@sun.com>:

[#13868] Re: Array#flatten works quadratic time on length of resulting array. It could be linear — "Voroztsov Artem" <artem.voroztsov@...> 2007/12/03

2007/12/3, Voroztsov Artem <artem.voroztsov@gmail.com>:

[#13870] Re: Array#flatten works quadratic time on length of resulting array. It could be linear — "Yusuke ENDOH" <mame@...> 2007/12/03

Hi,

[#13903] Clarification of retry change — Charles Oliver Nutter <charles.nutter@...>

Matz confirmed that retry-outside-rescue will no longer work, but I

14 messages 2007/12/07
[#13905] Re: Clarification of retry change — SASADA Koichi <ko1@...> 2007/12/07

Hi,

[#13908] What's the status of compiler/compiling on windows? — Gonzalo Garramu <ggarra@...>

20 messages 2007/12/07
[#13913] Re: What's the status of compiler/compiling on windows? — Nobuyoshi Nakada <nobu@...> 2007/12/07

Hi,

[#13914] Re: [Spam] Re: What's the status of compiler/compiling on windows? — Gonzalo Garramu <ggarra@...> 2007/12/07

Nobuyoshi Nakada wrote:

[#13926] Re: [Spam] Re: What's the status of compiler/compiling on windows? — "Luis Lavena" <luislavena@...> 2007/12/07

T24gRGVjIDcsIDIwMDcgODoyMSBBTSwgR29uemFsbyBHYXJyYW11w7FvIDxnZ2FycmFAYWR2YW5j

[#14038] Re: [Spam] Re: What's the status of compiler/compiling on windows? — "Joe Swatosh" <joe.swatosh@...> 2007/12/12

Hi Luis

[#14039] Re: [Spam] Re: What's the status of compiler/compiling on windows? — "Luis Lavena" <luislavena@...> 2007/12/12

On Dec 12, 2007 4:05 PM, Joe Swatosh <joe.swatosh@gmail.com> wrote:

[#14040] Re: [Spam] Re: What's the status of compiler/compiling on windows? — "Austin Ziegler" <halostatue@...> 2007/12/12

> This was discussed in other thread in ruby-talk, but just to summarize:

[#13969] redefineable not operator — David Flanagan <david@...>

Matz,

37 messages 2007/12/10
[#13971] Re: redefineable not operator — murphy <murphy@...> 2007/12/10

David Flanagan wrote:

[#13972] Re: redefineable not operator — Yukihiro Matsumoto <matz@...> 2007/12/10

Hi,

[#14007] Re: redefineable not operator — murphy <murphy@...> 2007/12/11

Yukihiro Matsumoto wrote:

[#14011] Re: redefineable not operator — Yukihiro Matsumoto <matz@...> 2007/12/11

Hi,

[#14013] Re: redefineable not operator — murphy <murphy@...> 2007/12/12

Yukihiro Matsumoto wrote:

[#14016] Re: redefineable not operator — David Flanagan <david@...> 2007/12/12

murphy wrote:

[#14019] Re: redefineable not operator — Yukihiro Matsumoto <matz@...> 2007/12/12

Hi,

[#14024] Re: redefineable not operator — Gary Wright <gwtmp01@...> 2007/12/12

[#14029] Re: redefineable not operator — Dave Thomas <dave@...> 2007/12/12

[#14042] Fix e2mmap.rb for 1.9 — Eric Hodel <drbrain@...7.net>

E2MM.Raise complains about $! being read-only now, and E2MM is used by

22 messages 2007/12/13
[#14043] Re: Fix e2mmap.rb for 1.9 — Dave Thomas <dave@...> 2007/12/13

[#14049] RDoc + irb (Was: Fix e2mmap.rb for 1.9) — Eric Hodel <drbrain@...7.net> 2007/12/13

On Dec 12, 2007, at 16:19 PM, Dave Thomas wrote:

[#14052] Re: RDoc + irb (Was: Fix e2mmap.rb for 1.9) — Dave Thomas <dave@...> 2007/12/13

[#14056] Re: RDoc + irb (Was: Fix e2mmap.rb for 1.9) — Charles Oliver Nutter <charles.nutter@...> 2007/12/13

Dave Thomas wrote:

[#14123] Some Ruby 1.9 loose ends to tie up — David Flanagan <david@...>

Matz,

20 messages 2007/12/17
[#14220] Re: Some Ruby 1.9 loose ends to tie up — Yukihiro Matsumoto <matz@...> 2007/12/21

Hi,

[#14238] Re: Some Ruby 1.9 loose ends to tie up — Charles Oliver Nutter <charles.nutter@...> 2007/12/22

Yukihiro Matsumoto wrote:

[#14147] named captures assigning to local variables — David Flanagan <david@...>

I've just been browsing ruby-dev. For an english speaker, it is kind of

26 messages 2007/12/19
[#14150] Re: named captures assigning to local variables — Tanaka Akira <akr@...> 2007/12/19

In article <47686B87.7050609@davidflanagan.com>,

[#14158] Re: named captures assigning to local variables — david@... 2007/12/19

Thank you for the clarification, akr. I'm embarassed to say that it didn't

[#14161] Re: named captures assigning to local variables — David Flanagan <david@...> 2007/12/20

If I may, have a proposal. My apologies if this has already been

[#14170] Re: named captures assigning to local variables — Tanaka Akira <akr@...> 2007/12/20

In article <476A087E.3070000@davidflanagan.com>,

[#14172] Re: named captures assigning to local variables — David Flanagan <david@...> 2007/12/20

How about making the return value an array of the captured strings, or nil

[#14149] Experimental PATCH to improve thread performance — Brent Roman <brent@...>

The attached patch against Ruby 1.8.6-p110 improves the performance of

38 messages 2007/12/19
[#14202] Re: Experimental PATCH to improve thread performance — Charles Oliver Nutter <charles.nutter@...> 2007/12/21

Brent Roman wrote:

[#14257] Re: Experimental PATCH to improve thread performance — Brent Roman <brent@...> 2007/12/22

[#14266] Re: Experimental PATCH to improve thread performance — Charles Oliver Nutter <charles.nutter@...> 2007/12/22

Brent Roman wrote:

[#14274] Re: Experimental PATCH to improve thread performance — Sylvain Joyeux <sylvain.joyeux@...4x.org> 2007/12/22

On Sat, Dec 22, 2007 at 11:07:25PM +0900, Charles Oliver Nutter wrote:

[#14186] Rake 0.8.0 added to Ruby — Jim Weirich <jim.weirich@...>

Just a heads up here. I've added Rake (version 0.8.0 ... the latest)

34 messages 2007/12/21
[#14210] Re: [Spam] Rake 0.8.0 added to Ruby — Gonzalo Garramu <ggarra@...> 2007/12/21

Jim Weirich wrote:

[#14213] Re: [Spam] Rake 0.8.0 added to Ruby — "Rick DeNatale" <rick.denatale@...> 2007/12/21

On Dec 21, 2007 10:24 AM, Gonzalo Garramu=F1o <ggarra@advancedsl.com.ar> wr=

[#14215] Re: [Spam] Rake 0.8.0 added to Ruby — Jim Weirich <jim.weirich@...> 2007/12/21

[#14303] IRHG - GC Memory Fragmentation? — Charles Thornton <ceo@...>

While working on Chapter 05 and referencing various works

23 messages 2007/12/23
[#14308] Re: IRHG - GC Memory Fragmentation? — Ryan Davis <ryand-ruby@...> 2007/12/23

[#14335] Many external symbols _without_ prefix in libruby object file — Tadashi Saito <shiba@...2.accsnet.ne.jp>

Hi all,

12 messages 2007/12/23

[#14364] RDoc: [FATAL] failed to allocate memory — Martin Duerst <duerst@...>

With revision 14590, I suddenly get an error when I do "make install"

13 messages 2007/12/24

[#14367] replace csv.rb with fastercsv.rb — "NAKAMURA, Hiroshi" <nakahiro@...>

-----BEGIN PGP SIGNED MESSAGE-----

15 messages 2007/12/24
[#14390] Re: replace csv.rb with fastercsv.rb — James Gray <james@...> 2007/12/24

On Dec 24, 2007, at 3:34 AM, NAKAMURA, Hiroshi wrote:

[#14418] Base64 not there makes Rails 2.0.2 fail to load in 1.9.0 — Richard Kilmer <rich@...>

Just re-built latest svn of 1.9.0 and base64.rb is removed. Its

51 messages 2007/12/25
[#14420] Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — Eric Hodel <drbrain@...7.net> 2007/12/25

On Dec 25, 2007, at 07:03 AM, Richard Kilmer wrote:

[#14427] Re: Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — "M. Edward (Ed) Borasky" <znmeb@...> 2007/12/25

Eric Hodel wrote:

[#14431] Re: Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — Dave Thomas <dave@...> 2007/12/25

[#14446] Re: Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — Eric Hodel <drbrain@...7.net> 2007/12/26

On Dec 25, 2007, at 13:35 PM, Dave Thomas wrote:

[#14452] Re: Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — Dave Thomas <dave@...> 2007/12/26

[#14492] Re: Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — Eric Hodel <drbrain@...7.net> 2007/12/27

On Dec 26, 2007, at 06:16 AM, Dave Thomas wrote:

[#14494] Re: Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — Dave Thomas <dave@...> 2007/12/27

[#14503] Re: Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — Richard Kilmer <rich@...> 2007/12/27

[#14505] Re: Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — Charles Oliver Nutter <charles.nutter@...> 2007/12/27

Richard Kilmer wrote:

[#14429] Re: Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — hemant <gethemant@...> 2007/12/25

On Dec 26, 2007 1:01 AM, M. Edward (Ed) Borasky <znmeb@cesmail.net> wrote:

[#14430] Re: Legacy support (Was: Base64 not there makes Rails 2.0.2 fail to load in 1.9.0) — "M. Edward (Ed) Borasky" <znmeb@...> 2007/12/25

hemant wrote:

[#14517] Invalid use of mktime() in Ruby 1.8/1.9 results in incorrect Time objects — Dirkjan Bussink <d.bussink@...>

Hello,

12 messages 2007/12/27

[#14549] multibyte strings & bucket-of-bytes efficiency under 1.9.0 — khaines@...

Like everyone else, I've been testing my stuff under 1.9.0. In general,

38 messages 2007/12/28
[#14560] Re: multibyte strings & bucket-of-bytes efficiency under 1.9.0 — Brent Roman <brent@...> 2007/12/28

[#14573] Re: multibyte strings & bucket-of-bytes efficiency under 1.9.0 — murphy <murphy@...> 2007/12/29

Brent Roman wrote:

[#14603] Re: multibyte strings & bucket-of-bytes efficiency under 1.9.0 — Brent Roman <brent@...> 2007/12/30

[#14617] Re: multibyte strings & bucket-of-bytes efficiency under 1.9.0 — Tanaka Akira <akr@...> 2007/12/31

In article <14544702.post@talk.nabble.com>,

[#14568] Layout of includes in ruby 1.9 — Gonzalo Garramu <ggarra@...>

19 messages 2007/12/29
[#14576] Re: Layout of includes in ruby 1.9 — "Rick DeNatale" <rick.denatale@...> 2007/12/29

On Dec 29, 2007 2:39 AM, Gonzalo Garramu=F1o <ggarra@advancedsl.com.ar> wro=

[#14569] Wide strings to ruby strings — Gonzalo Garramu <ggarra@...>

11 messages 2007/12/29

[#14602] RCR allow indexing last n items — "Michal Suchanek" <hramrach@...>

Hello

15 messages 2007/12/30
[#14609] Re: RCR allow indexing last n items — "David A. Black" <dblack@...> 2007/12/30

Hi --

[#14610] Re: RCR allow indexing last n items — "Michal Suchanek" <hramrach@...> 2007/12/30

On 30/12/2007, David A. Black <dblack@rubypal.com> wrote:

[#14616] Re: RCR allow indexing last n items — "David A. Black" <dblack@...> 2007/12/30

Hi --

[#14621] Module.new(&block) in Ruby 1.9 — murphy <murphy@...>

Hello!

21 messages 2007/12/31
[#14622] Re: Module.new(&block) in Ruby 1.9 — "Cheah Chu Yeow" <chuyeow@...> 2007/12/31

This looks like a related bug with passing block arguments to

[#14633] Re: Module.new(&block) in Ruby 1.9 — murphy <murphy@...> 2007/12/31

Cheah Chu Yeow wrote:

[#14716] Re: Module.new(&block) in Ruby 1.9 — SASADA Koichi <ko1@...> 2008/01/03

Hi,

[#14726] Re: Module.new(&block) in Ruby 1.9 — ts <decoux@...> 2008/01/03

>>>>> "S" == SASADA Koichi <ko1@atdot.net> writes:

[#14728] Re: Module.new(&block) in Ruby 1.9 — SASADA Koichi <ko1@...> 2008/01/03

Hi,

[#16093] Re: Module.new(&block) in Ruby 1.9 — "Jeremy Kemper" <jeremy@...> 2008/04/01

Hi,

Re: Experimental PATCH to improve thread performance

From: Brent Roman <brent@...>
Date: 2007-12-23 06:13:09 UTC
List: ruby-core #14301
Charlie,

Thank you for those references.  
I always understood why kill!/exit!/terminate! methods
were doomed to be unsafe.  However, now I see why
even the "non-bang" versions of these methods 
and thread#raise are also unsafe. 
There is no practical way to interrupt a thread
at any arbitrary point and expect it to preserve
and/or repair data all invariants.

Even so,
I'm not convinced that Ruby's Thread.stop is unsafe,
because, unlike Java's, Ruby's Thread.stop method
can only act on the currently executing thread.
One cannot stop another Ruby thread.  
A Ruby thread can only stop itself.

Further, I believe that thread#raise (and even the non-bang
variants exit, kill and terminate) can be made
safe by adding an exception queue 
and an "interruptible" flag to each Ruby thread.
This may provide Ruby thread safety without 
breaking most existing threaded applications.

When a Ruby thread is interruptible, it acts upon
exceptions raised in it by other threads.  When it
is not interruptible, exceptions raised by other threads
are queued to be acted upon (in order received)
when next the thread becomes interruptible.

When a Ruby thread is first created, it would not
be interruptible until it either explicitly sets
its interruptible flag or blocks for any reason.
A thread would always be interruptible while blocked.
It would revert back to its previous interruptible status
when it unblocked.  

A thread would always become uninterruptible whenever
it acted upon an exception raised in it by Thread#raise.
But, exceptions raised by Kernel.raise would not
change the thread's interruptible status and they
would always bypass the thread's exception queue.

What this attempts to do is to turn Ruby's
(very flexible) exception mechanism into a safe
vehicle for messaging. 

Do you think there's any merit to this approach?
Any obvious (or not to obvious) holes in it?

- brent



Charles Oliver Nutter-2 wrote:
> 
> 
> See my description of threading in Ruby here:
> 
> http://spec.ruby-doc.org/wiki/Ruby_Threading
> 
> and JRuby here:
> 
> http://spec.ruby-doc.org/wiki/JRuby_Threading
> 
> ... snip ...
> 
> Provably unsafe thread operations Ruby supports and should not:
> 
> stop
> exit/exit!/kill/kill!/terminate/terminate!
> raise
> global abort_on_exception
> 
> Operations which are largely incompatible with concurrent threading 
> (which Ruby can't support if it wants to support concurrent threading
> well):
> 
> critical=
> 
> I'll enumerate the reasons why these are unsafe if anyone wants to know. 
> I think Java's threading is a good model here; it provides the thread 
> operations that can be safely supported and does not provide the others.
> 
> Here's an article that explains some of the reasons:
> 
> http://java.sun.com/j2se/1.5.0/docs/guide/misc/threadPrimitiveDeprecation.html
> 
> - Charlie
> 
> 
> 

-- 
View this message in context: http://www.nabble.com/Experimental-PATCH-to-improve-thread-performance-tp14409855p14476198.html
Sent from the ruby-core mailing list archive at Nabble.com.


In This Thread