--001517476a78095bba04ae41dd79 Content-Type: text/plain; charset=ISO-8859-1 > I made a questionnaire "What do you want to introduce in 2.0?" in my > RubyConf2011 presentation. > > Followings are results from my memo. My memo should be incomplete, so > please complement them. Thank you for summarizing and requesting feedback. It's nice for those who didn't attend your presentation to see the current results and be able to provide feedback. > - Cleanup syntax > - Bytecode export > - Symbol GC > - Remove Proc binding > - Macros > - Getting parse tree > - Getting source code > - Cross thread Fiber migration > - Standard Gem > - Review all standard libraries > - Remove obsolete one standard libraries > - Improve Proc#curry > - Non-blocking I/O > - Dtrace > - GC API (replacable GC) One issue that continually fails to gain sustainable traction is identifying the root causes and fixing the 1.9.x IO performance regressions on Windows. Until this issue is added to high profile lists like yours, prioritized, and viewed as important to MRI's future by Matz and ruby-core, I'm concerned whether it will ever get fixed. And yes, I'm aware of Luis' recent activities in investigating refactoring opportunities in io.c, file.c, etc, and others past efforts. From my perspective, it's simple. Either it's important to MRI's future (and by extension Ruby the language) to be a great multi-platform implementation, or it is not. If best-of-breed multi-platform support [1] is important, prioritize fixing the current MRI-on-Windows performance regressions in a timely manner given the current resourcing realities. I request that you add "Fix Windows 1.9.x performance regressions" to your 2.0 list. Jon [1] As an example, see the great multi-platform efforts being done in https://github.com/joyent/libuv for node and other usages. See how easy it is to build and test the library on Windows. See that node has a prebuilt exe available for Windows at http://nodejs.org/ at v0.5.8 not 1.x or 2.x --- blog: http://jonforums.github.com/ twitter: @jonforums Most people die of a sort of creeping common sense, and discover when it is too late that the only things one never regrets are one's mistakes. - Oscar Wilde --001517476a78095bba04ae41dd79 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable <div>> I made a questionnaire "What do you want to introduce in 2.0?" in my</div><div>> RubyConf2011 presentation.</div><div>>/div><div>> Followings are results from my memo. ¨Âíåíï óèïõìä âå éîãïíðìåôåóＯäéö¾ <div>> please complement them.</div><div><br></div><div><br></div><div>Thank you for summarizing and requesting feedback. It's nice for those who didn't attend your presentation to see the current results and be able to provide feedback.</div> <div><br></div><div><br></div><div>> - Cleanup syntax</div><div>> - Bytecode export</div><div>> - Symbol GC</div><div>> - Remove Proc binding</div><div>> - Macros</div><div>> - Getting parse tree</div><div> > - Getting source code</div><div>> - Cross thread Fiber migration</div><div>> - Standard Gem</div><div>> - Review all standard libraries</div><div>> - Remove obsolete one standard libraries</div><div>> - Improve Proc#curry</div> <div>> - Non-blocking I/O</div><div>> - Dtrace</div><div>> - GC API (replacable GC)</div><div><br></div><div><br></div><div>One issue that continually fails to gain sustainable traction is identifying the root causesnd fixing the 1.9.x IO performance regressions on Windows.</div> <div><br></div><div>Until this issue is added to high profile lists like yours, prioritized, and viewed as important to MRI's future by Matz and ruby-core, I'm concerned whether it will ever get fixed. And yes, I'm aware of Luis' recent activities in investigating refactoring opportunities in io.c, file.c, etc, and others past efforts.</div> <div><br></div><div>From my perspective, it's simple. Either it's important to MRI's future (and by extension Ruby the language) to be a great multi-platform implementation, or it is not. If best-of-breed multi-platform support [1] is important, prioritize fixing the current MRI-on-Windows performance regressions in a timely manner given the current resourcing realities.</div> <div><br></div><div>I request that you add "Fix Windows 1.9.x performance regressions" to your 2.0 list.</div><div><br></div><div>Jon</div><div><br></div><div><br></div><div>[1] As an example, see the great multi-platform efforts being done in https://github.com/joyent/libuv</div> <div>for node and other usages. See how easy it is to build and test the library on Windows. See that</div><div>node has a prebuilt exe available for Windows at http://nodejs.org/ at v0.5.8 not 1.x or 2.x</div> <div><br></div><div>---</div><div>blog: http://jonforums.github.com/</div><div>twitter: @jonforums</div><div><br></div><div>Most people die of a sort of creeping common sense, and discover when it</div> <div>is too late that the only things one never regrets are one's mistakes.</div><div> Oscar Wilde</div><div><br></div> --001517476a78095bba04ae41dd79--