* [ruby-core:72523] [Ruby trunk - Bug #11901] [Open] Performance Issue with OpenStruct
[not found] <redmine.issue-11901.20151227153720@ruby-lang.org>
@ 2015-12-27 15:37 ` ariel.caplan
2015-12-27 15:48 ` [ruby-core:72524] [Ruby trunk - Bug #11901] " ariel.caplan
` (7 subsequent siblings)
8 siblings, 0 replies; 9+ messages in thread
From: ariel.caplan @ 2015-12-27 15:37 UTC (permalink / raw)
To: ruby-core
Issue #11901 has been reported by Ariel Caplan.
----------------------------------------
Bug #11901: Performance Issue with OpenStruct
https://bugs.ruby-lang.org/issues/11901
* Author: Ariel Caplan
* Status: Open
* Priority: Normal
* Assignee:
* ruby -v: ruby 2.3.0p0 (2015-12-25 revision 53290) [x86_64-darwin13]
* Backport:
----------------------------------------
After recent changes to define OpenStruct getter/setter methods lazily, there is a heavy performance impact for the use case where an attribute is assigned at initialization time (i.e. `Openstruct.new(foo: :bar)`). Once an attribute is stored in the internal hash, the appropriate singleton methods will never be defined, due to the recent changes to OpenStruct's `#respond_to_missing?` - meaning that every time I call `#foo` or `#foo=` it relies on `#method_missing`. Benchmark using benchmark-ips is attached.
I'm primarily concerned about the case of configuration objects, which may be populated at initialization time and then accessed many times throughout the life of the program.
---Files--------------------------------
openstruct-regression-benchmark.rb (1.36 KB)
--
https://bugs.ruby-lang.org/
^ permalink raw reply [flat|nested] 9+ messages in thread
* [ruby-core:72524] [Ruby trunk - Bug #11901] Performance Issue with OpenStruct
[not found] <redmine.issue-11901.20151227153720@ruby-lang.org>
2015-12-27 15:37 ` [ruby-core:72523] [Ruby trunk - Bug #11901] [Open] Performance Issue with OpenStruct ariel.caplan
@ 2015-12-27 15:48 ` ariel.caplan
2015-12-27 16:02 ` [ruby-core:72525] " ariel.caplan
` (6 subsequent siblings)
8 siblings, 0 replies; 9+ messages in thread
From: ariel.caplan @ 2015-12-27 15:48 UTC (permalink / raw)
To: ruby-core
Issue #11901 has been updated by Ariel Caplan.
Assignee set to Marc-Andre Lafortune
To be more specific (but not clog up the description), the problem can be traced to https://github.com/ruby/ruby/blob/b8d9770b6c699af6e63dab727621777fbfbf7b44/lib/ostruct.rb#L166 where the methods are only defined if `#respond_to?` is `false` for that method name. Since `#repond_to_missing?` was overridden to report `true` when the key is in the table (https://github.com/ruby/ruby/blob/b8d9770b6c699af6e63dab727621777fbfbf7b44/lib/ostruct.rb#L176), the methods are never defined.
----------------------------------------
Bug #11901: Performance Issue with OpenStruct
https://bugs.ruby-lang.org/issues/11901#change-55788
* Author: Ariel Caplan
* Status: Open
* Priority: Normal
* Assignee: Marc-Andre Lafortune
* ruby -v: ruby 2.3.0p0 (2015-12-25 revision 53290) [x86_64-darwin13]
* Backport:
----------------------------------------
After recent changes to define OpenStruct getter/setter methods lazily, there is a heavy performance impact for the use case where an attribute is assigned at initialization time (i.e. `Openstruct.new(foo: :bar)`). Once an attribute is stored in the internal hash, the appropriate singleton methods will never be defined, due to the recent changes to OpenStruct's `#respond_to_missing?` - meaning that every time I call `#foo` or `#foo=` it relies on `#method_missing`. Benchmark using benchmark-ips is attached.
I'm primarily concerned about the case of configuration objects, which may be populated at initialization time and then accessed many times throughout the life of the program.
---Files--------------------------------
openstruct-regression-benchmark.rb (1.36 KB)
--
https://bugs.ruby-lang.org/
^ permalink raw reply [flat|nested] 9+ messages in thread
* [ruby-core:72525] [Ruby trunk - Bug #11901] Performance Issue with OpenStruct
[not found] <redmine.issue-11901.20151227153720@ruby-lang.org>
2015-12-27 15:37 ` [ruby-core:72523] [Ruby trunk - Bug #11901] [Open] Performance Issue with OpenStruct ariel.caplan
2015-12-27 15:48 ` [ruby-core:72524] [Ruby trunk - Bug #11901] " ariel.caplan
@ 2015-12-27 16:02 ` ariel.caplan
2015-12-31 5:40 ` [ruby-core:72631] " ruby-core
` (5 subsequent siblings)
8 siblings, 0 replies; 9+ messages in thread
From: ariel.caplan @ 2015-12-27 16:02 UTC (permalink / raw)
To: ruby-core
Issue #11901 has been updated by Ariel Caplan.
Now, to throw in my own opinion: probably the simplest fix would be to circumvent the `#respond_to?` check if we hit `#method_missing?` already - the check is both unnecessary and inaccurate. So probably we'd want the method defining methods to be its own method, and then have both `#method_missing?` and `#new_ostruct_member` rely on that. Something like:
``` ruby
class OpenStruct
def new_ostruct_member(name)
name = name.to_sym
unless respond_to?(name)
define_openstruct_methods(name)
end
name
end
def define_openstruct_methods(name)
define_singleton_method(name) { @table[name] }
define_singleton_method("#{name}=") { |x| modifiable[name] = x }
name
end
def method_missing(mid, *args) # :nodoc:
len = args.length
if mname = mid[/.*(?==\z)/m]
if len != 1
raise ArgumentError, "wrong number of arguments (#{len} for 1)", caller(1)
end
modifiable[define_openstruct_methods(mname)] = args[0]
elsif len == 0
if @table.key?(mid)
define_openstruct_methods(mid)
@table[mid]
end
else
err = NoMethodError.new "undefined method `#{mid}' for #{self}", mid, args
err.set_backtrace caller(1)
raise err
end
end
end
```
Running the previously attached benchmark prefaced with this code, the "assigned on initialization" benchmark outperforms the "assigned after initialization" one.
However, it does raise the question: Should calling `#foo=` define the methods actively, or should we only try to define on `#foo` to match lazy behavior on the initializer?
----------------------------------------
Bug #11901: Performance Issue with OpenStruct
https://bugs.ruby-lang.org/issues/11901#change-55789
* Author: Ariel Caplan
* Status: Open
* Priority: Normal
* Assignee: Marc-Andre Lafortune
* ruby -v: ruby 2.3.0p0 (2015-12-25 revision 53290) [x86_64-darwin13]
* Backport:
----------------------------------------
After recent changes to define OpenStruct getter/setter methods lazily, there is a heavy performance impact for the use case where an attribute is assigned at initialization time (i.e. `Openstruct.new(foo: :bar)`). Once an attribute is stored in the internal hash, the appropriate singleton methods will never be defined, due to the recent changes to OpenStruct's `#respond_to_missing?` - meaning that every time I call `#foo` or `#foo=` it relies on `#method_missing`. Benchmark using benchmark-ips is attached.
I'm primarily concerned about the case of configuration objects, which may be populated at initialization time and then accessed many times throughout the life of the program.
---Files--------------------------------
openstruct-regression-benchmark.rb (1.36 KB)
--
https://bugs.ruby-lang.org/
^ permalink raw reply [flat|nested] 9+ messages in thread
* [ruby-core:72631] [Ruby trunk - Bug #11901] Performance Issue with OpenStruct
[not found] <redmine.issue-11901.20151227153720@ruby-lang.org>
` (2 preceding siblings ...)
2015-12-27 16:02 ` [ruby-core:72525] " ariel.caplan
@ 2015-12-31 5:40 ` ruby-core
2015-12-31 5:43 ` [ruby-core:72632] " ruby-core
` (4 subsequent siblings)
8 siblings, 0 replies; 9+ messages in thread
From: ruby-core @ 2015-12-31 5:40 UTC (permalink / raw)
To: ruby-core
Issue #11901 has been updated by Marc-Andre Lafortune.
Indeed, the test in `new_ostruct_member` was incorrect now that `respond_to_missing?` has been changed.
Instead, if we look for the actual method I believe this will fix all issues.
----------------------------------------
Bug #11901: Performance Issue with OpenStruct
https://bugs.ruby-lang.org/issues/11901#change-55883
* Author: Ariel Caplan
* Status: Open
* Priority: Normal
* Assignee: Marc-Andre Lafortune
* ruby -v: ruby 2.3.0p0 (2015-12-25 revision 53290) [x86_64-darwin13]
* Backport:
----------------------------------------
After recent changes to define OpenStruct getter/setter methods lazily, there is a heavy performance impact for the use case where an attribute is assigned at initialization time (i.e. `Openstruct.new(foo: :bar)`). Once an attribute is stored in the internal hash, the appropriate singleton methods will never be defined, due to the recent changes to OpenStruct's `#respond_to_missing?` - meaning that every time I call `#foo` or `#foo=` it relies on `#method_missing`. Benchmark using benchmark-ips is attached.
I'm primarily concerned about the case of configuration objects, which may be populated at initialization time and then accessed many times throughout the life of the program.
---Files--------------------------------
openstruct-regression-benchmark.rb (1.36 KB)
--
https://bugs.ruby-lang.org/
^ permalink raw reply [flat|nested] 9+ messages in thread
* [ruby-core:72632] [Ruby trunk - Bug #11901] Performance Issue with OpenStruct
[not found] <redmine.issue-11901.20151227153720@ruby-lang.org>
` (3 preceding siblings ...)
2015-12-31 5:40 ` [ruby-core:72631] " ruby-core
@ 2015-12-31 5:43 ` ruby-core
2015-12-31 17:10 ` [ruby-core:72639] [Ruby trunk - Bug #11901] [Closed] " naruse
` (3 subsequent siblings)
8 siblings, 0 replies; 9+ messages in thread
From: ruby-core @ 2015-12-31 5:43 UTC (permalink / raw)
To: ruby-core
Issue #11901 has been updated by Marc-Andre Lafortune.
Backport set to 2.3: REQUIRED
----------------------------------------
Bug #11901: Performance Issue with OpenStruct
https://bugs.ruby-lang.org/issues/11901#change-55884
* Author: Ariel Caplan
* Status: Open
* Priority: Normal
* Assignee: Marc-Andre Lafortune
* ruby -v: ruby 2.3.0p0 (2015-12-25 revision 53290) [x86_64-darwin13]
* Backport: 2.3: REQUIRED
----------------------------------------
After recent changes to define OpenStruct getter/setter methods lazily, there is a heavy performance impact for the use case where an attribute is assigned at initialization time (i.e. `Openstruct.new(foo: :bar)`). Once an attribute is stored in the internal hash, the appropriate singleton methods will never be defined, due to the recent changes to OpenStruct's `#respond_to_missing?` - meaning that every time I call `#foo` or `#foo=` it relies on `#method_missing`. Benchmark using benchmark-ips is attached.
I'm primarily concerned about the case of configuration objects, which may be populated at initialization time and then accessed many times throughout the life of the program.
---Files--------------------------------
openstruct-regression-benchmark.rb (1.36 KB)
--
https://bugs.ruby-lang.org/
^ permalink raw reply [flat|nested] 9+ messages in thread
* [ruby-core:72639] [Ruby trunk - Bug #11901] [Closed] Performance Issue with OpenStruct
[not found] <redmine.issue-11901.20151227153720@ruby-lang.org>
` (4 preceding siblings ...)
2015-12-31 5:43 ` [ruby-core:72632] " ruby-core
@ 2015-12-31 17:10 ` naruse
2016-01-01 17:35 ` [ruby-core:72658] [Ruby trunk - Bug #11901] " ruby-core
` (2 subsequent siblings)
8 siblings, 0 replies; 9+ messages in thread
From: naruse @ 2015-12-31 17:10 UTC (permalink / raw)
To: ruby-core
Issue #11901 has been updated by Yui NARUSE.
Status changed from Open to Closed
Could you fix rubyspec?
----------------------------------------
Bug #11901: Performance Issue with OpenStruct
https://bugs.ruby-lang.org/issues/11901#change-55890
* Author: Ariel Caplan
* Status: Closed
* Priority: Normal
* Assignee: Marc-Andre Lafortune
* ruby -v: ruby 2.3.0p0 (2015-12-25 revision 53290) [x86_64-darwin13]
* Backport: 2.3: REQUIRED
----------------------------------------
After recent changes to define OpenStruct getter/setter methods lazily, there is a heavy performance impact for the use case where an attribute is assigned at initialization time (i.e. `Openstruct.new(foo: :bar)`). Once an attribute is stored in the internal hash, the appropriate singleton methods will never be defined, due to the recent changes to OpenStruct's `#respond_to_missing?` - meaning that every time I call `#foo` or `#foo=` it relies on `#method_missing`. Benchmark using benchmark-ips is attached.
I'm primarily concerned about the case of configuration objects, which may be populated at initialization time and then accessed many times throughout the life of the program.
---Files--------------------------------
openstruct-regression-benchmark.rb (1.36 KB)
--
https://bugs.ruby-lang.org/
^ permalink raw reply [flat|nested] 9+ messages in thread
* [ruby-core:72658] [Ruby trunk - Bug #11901] Performance Issue with OpenStruct
[not found] <redmine.issue-11901.20151227153720@ruby-lang.org>
` (5 preceding siblings ...)
2015-12-31 17:10 ` [ruby-core:72639] [Ruby trunk - Bug #11901] [Closed] " naruse
@ 2016-01-01 17:35 ` ruby-core
2016-01-01 17:38 ` [ruby-core:72659] " eregontp
2016-03-29 9:52 ` [ruby-core:74680] [Ruby trunk Bug#11901] " naruse
8 siblings, 0 replies; 9+ messages in thread
From: ruby-core @ 2016-01-01 17:35 UTC (permalink / raw)
To: ruby-core
Issue #11901 has been updated by Marc-Andre Lafortune.
Yui NARUSE wrote:
> Could you fix rubyspec?
Great, that was a bug
I made an easy fix so `method_missing` doesn't add a method if frozen. I hesitated to redefine `freeze` to add all the needed method definitions, but it's difficult to know if it would be preferable.
----------------------------------------
Bug #11901: Performance Issue with OpenStruct
https://bugs.ruby-lang.org/issues/11901#change-55909
* Author: Ariel Caplan
* Status: Closed
* Priority: Normal
* Assignee: Marc-Andre Lafortune
* ruby -v: ruby 2.3.0p0 (2015-12-25 revision 53290) [x86_64-darwin13]
* Backport: 2.3: REQUIRED
----------------------------------------
After recent changes to define OpenStruct getter/setter methods lazily, there is a heavy performance impact for the use case where an attribute is assigned at initialization time (i.e. `Openstruct.new(foo: :bar)`). Once an attribute is stored in the internal hash, the appropriate singleton methods will never be defined, due to the recent changes to OpenStruct's `#respond_to_missing?` - meaning that every time I call `#foo` or `#foo=` it relies on `#method_missing`. Benchmark using benchmark-ips is attached.
I'm primarily concerned about the case of configuration objects, which may be populated at initialization time and then accessed many times throughout the life of the program.
---Files--------------------------------
openstruct-regression-benchmark.rb (1.36 KB)
--
https://bugs.ruby-lang.org/
^ permalink raw reply [flat|nested] 9+ messages in thread
* [ruby-core:72659] [Ruby trunk - Bug #11901] Performance Issue with OpenStruct
[not found] <redmine.issue-11901.20151227153720@ruby-lang.org>
` (6 preceding siblings ...)
2016-01-01 17:35 ` [ruby-core:72658] [Ruby trunk - Bug #11901] " ruby-core
@ 2016-01-01 17:38 ` eregontp
2016-03-29 9:52 ` [ruby-core:74680] [Ruby trunk Bug#11901] " naruse
8 siblings, 0 replies; 9+ messages in thread
From: eregontp @ 2016-01-01 17:38 UTC (permalink / raw)
To: ruby-core
Issue #11901 has been updated by Benoit Daloze.
Marc-Andre Lafortune wrote:
> Yui NARUSE wrote:
> > Could you fix rubyspec?
>
> Great, that was a bug
>
> I made an easy fix so `method_missing` doesn't add a method if frozen. I hesitated to redefine `freeze` to add all the needed method definitions, but it's difficult to know if it would be preferable.
nobu did the latter in r53396.
----------------------------------------
Bug #11901: Performance Issue with OpenStruct
https://bugs.ruby-lang.org/issues/11901#change-55910
* Author: Ariel Caplan
* Status: Closed
* Priority: Normal
* Assignee: Marc-Andre Lafortune
* ruby -v: ruby 2.3.0p0 (2015-12-25 revision 53290) [x86_64-darwin13]
* Backport: 2.3: REQUIRED
----------------------------------------
After recent changes to define OpenStruct getter/setter methods lazily, there is a heavy performance impact for the use case where an attribute is assigned at initialization time (i.e. `Openstruct.new(foo: :bar)`). Once an attribute is stored in the internal hash, the appropriate singleton methods will never be defined, due to the recent changes to OpenStruct's `#respond_to_missing?` - meaning that every time I call `#foo` or `#foo=` it relies on `#method_missing`. Benchmark using benchmark-ips is attached.
I'm primarily concerned about the case of configuration objects, which may be populated at initialization time and then accessed many times throughout the life of the program.
---Files--------------------------------
openstruct-regression-benchmark.rb (1.36 KB)
--
https://bugs.ruby-lang.org/
^ permalink raw reply [flat|nested] 9+ messages in thread
* [ruby-core:74680] [Ruby trunk Bug#11901] Performance Issue with OpenStruct
[not found] <redmine.issue-11901.20151227153720@ruby-lang.org>
` (7 preceding siblings ...)
2016-01-01 17:38 ` [ruby-core:72659] " eregontp
@ 2016-03-29 9:52 ` naruse
8 siblings, 0 replies; 9+ messages in thread
From: naruse @ 2016-03-29 9:52 UTC (permalink / raw)
To: ruby-core
Issue #11901 has been updated by Yui NARUSE.
Backport changed from 2.3: REQUIRED to 2.3: DONE
ruby_2_3 r54388 merged revision(s) 53395,53396.
----------------------------------------
Bug #11901: Performance Issue with OpenStruct
https://bugs.ruby-lang.org/issues/11901#change-57808
* Author: Ariel Caplan
* Status: Closed
* Priority: Normal
* Assignee: Marc-Andre Lafortune
* ruby -v: ruby 2.3.0p0 (2015-12-25 revision 53290) [x86_64-darwin13]
* Backport: 2.3: DONE
----------------------------------------
After recent changes to define OpenStruct getter/setter methods lazily, there is a heavy performance impact for the use case where an attribute is assigned at initialization time (i.e. `Openstruct.new(foo: :bar)`). Once an attribute is stored in the internal hash, the appropriate singleton methods will never be defined, due to the recent changes to OpenStruct's `#respond_to_missing?` - meaning that every time I call `#foo` or `#foo=` it relies on `#method_missing`. Benchmark using benchmark-ips is attached.
I'm primarily concerned about the case of configuration objects, which may be populated at initialization time and then accessed many times throughout the life of the program.
---Files--------------------------------
openstruct-regression-benchmark.rb (1.36 KB)
--
https://bugs.ruby-lang.org/
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2016-03-29 9:16 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <redmine.issue-11901.20151227153720@ruby-lang.org>
2015-12-27 15:37 ` [ruby-core:72523] [Ruby trunk - Bug #11901] [Open] Performance Issue with OpenStruct ariel.caplan
2015-12-27 15:48 ` [ruby-core:72524] [Ruby trunk - Bug #11901] " ariel.caplan
2015-12-27 16:02 ` [ruby-core:72525] " ariel.caplan
2015-12-31 5:40 ` [ruby-core:72631] " ruby-core
2015-12-31 5:43 ` [ruby-core:72632] " ruby-core
2015-12-31 17:10 ` [ruby-core:72639] [Ruby trunk - Bug #11901] [Closed] " naruse
2016-01-01 17:35 ` [ruby-core:72658] [Ruby trunk - Bug #11901] " ruby-core
2016-01-01 17:38 ` [ruby-core:72659] " eregontp
2016-03-29 9:52 ` [ruby-core:74680] [Ruby trunk Bug#11901] " naruse
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).