January 2011 - MacRuby-devel - lists.macosforge.org

Concurrency bug in ControlTower
by Charles Oliver Nutter 23 Oct '11

23 Oct '11

Apologies reporting this here; I'm not sure where I should report bugs in ControlTower. I believe there's a bug in ControlTower around line 34: def open @status = :open while (@status == :open) connection = @socket.accept # << here @request_queue.async(@request_group) do The "connection" local variable is used in the async block below to parse the request and send the response. Unless MacRuby's blocks behave differently than regular Ruby, this variable is shared across all activations of that block. As a result, it's possible that two concurrent requests will end up using each others' connection, and usually just blowing up as a result. The fix I've come up with is to wrap the @request_queue.async call in a tap call: def open @status = :open while (@status == :open) conn = @socket.accept conn.tap do |connection| @request_queue.async(@request_group) do Since async can't accept any explicitly-passed state, this seems like the safest way to ensure the connection reference is not shared by separate invocations. I'm not sure if it's possible in GCD, but having async take an optional argument with explicitly-passed state might also be a good way to fix this. - Charlie

4 8

How to detect Cylons with MacRuby
by Martin Hawkins 05 Feb '11

05 Feb '11

I was just looking for info relating to debugging when I re-discovered this post. In it, there's a suggestion that somebody might work on XCode integration. Has there been any development around debugging?

2 1

is_a? method
by Robert Rice 04 Feb '11

04 Feb '11

Hi: I notice that the obj.is_a? method does not work on a class constant. It only works on an instance of the class. Is this the way it should work? Do I need to create a loop checking using the superclass method to check the inheritance of a class or is there an easier way? Thanks, Bob Rice

5 8

MacRuby Book discount
by Matt Aimonetti 04 Feb '11

04 Feb '11

O'Reilly is running a special today and my MacRuby: Definitive Guide - Early Release - ebook (PDF, mobi & ePub) is for sale for only $15.99 with the following code: DDMRD http://oreil.ly/eij2QS The book is also available in HTML format for free. If you've enjoyed the free version, please consider leaving a review on the page above to show your support to myself & to my publisher. Thanks, - Matt Sent from my iPhone

14 27

IDE and TextEditor
by Daniel Lopes 03 Feb '11

03 Feb '11

Hello everybody, I'm spending more and more time with macruby but I'm not very happy with XCode yet. The IDE is pretty good but I didn't found a good theme that colorizes Ruby syntax in the right way. I don't consider TextMate and Vim because for frameworks like Cocoa I think autocomplete and integration with docs is really important. What do you guys use for edit MacRuby code?

5 4

Ruby sort algorithm
by Robert Rice 01 Feb '11

01 Feb '11

Hi: Does the Ruby Array sort algorithm maintain the relative position for children returning the same value for the comparison? I had an instance where two children having the compare value were interchanged. Thanks, Bob Rice

3 9

Subject: Re: Strange NSDate behavior building 32 bit v 64 bit
by Richard Sepulveda 01 Feb '11

01 Feb '11

That makes perfectly good sense but i unfortunately started selling a MacRuby app on the App Store for i386 and 64 bit machines. And a few people are experiencing this issue. I was just hoping for a quick workaround to make them happy. And I would discontinue selling the 32 bit version on the next release. But i can't see anything obvious other than rewriting all of my NSDate based code in Objective-C or waiting for a fix. i include the MacRuby framework in my Pkg so that is possible. Richard > Message: 2 > Date: Sun, 30 Jan 2011 22:08:38 -0800 > From: "Jordan K. Hubbard" <jkh(a)apple.com> > To: "MacRuby development discussions." > <macruby-devel(a)lists.macosforge.org> > Subject: Re: [MacRuby-devel] Strange NSDate behavior building 32 bit v > 64 bit > Message-ID: <D31EF44C-06F8-45B1-83B4-7977A32BD53F(a)apple.com> > Content-Type: text/plain; charset="us-ascii" > > I suppose this begs the question: Does anyone really *require* 32 bit support for MacRuby at this point? SnowLeopard is already the minimum supported config, and the only Intel 32 bit-only platforms (very early MacBook and Mac Mini configurations) are several years old now. I don't want to sound like an unfeeling ogre to anyone who actually has such a configuration, mind you, but how big of an installed base does this really represent? > > - Jordan > > On Jan 30, 2011, at 8:49 PM, Vincent Isambart wrote: > >>> 1. Modified the Valid Archetectures to "i386 x86_64" >> >> There's a simple way to run macruby (or any other program) on the >> command line in 32 bits: just add "arch -i386" before the name of the >> program to execute: >> $ macruby -v >> MacRuby 0.9 (ruby 1.9.2) [universal-darwin10.0, x86_64] >> $ arch -i386 macruby -v >> MacRuby 0.9 (ruby 1.9.2) [universal-darwin10.0, i386]

4 6

Array and NSMutableArray mutability
by Mark Rada 31 Jan '11

31 Jan '11

Hey, I was trying to implement a breadth first search where I would mutate the search queue while iterating through it. Simplified, it was something like this: numbers = [1, 2, 3, 4] numbers.each { |number| numbers << (1+number) unless number > 3 puts number } This works fine for arrays that are being made in the MacRuby VM, but if I get an array back from an Objective-C method and try to iterate over that array then I would get an exception (which segfaulted the VM) when I tried to mutate the array during iteration. Simplified, it would be something like this: numbers = NSMutableArray.alloc.init numbers.addObjectsFromArray [1, 2, 3, 4] numbers.each { |number| numbers << (1+number) unless number > 3 puts number } Should both types of arrays work the same way? An instance of NSMutableArray says its class is Array, so I thought that they would. Also, is this type of exception one that should be caught by the VM and propagated as a normal Ruby Exception that I could rescue? Mark Rada mrada(a)marketcircle.com

4 4

Strange NSDate behavior building 32 bit v 64 bit
by Richard Sepulveda 31 Jan '11

31 Jan '11

After further investigation, I have found that the "to_f" method in my 32 bit build has problems. My simple test case, I created a new MacRuby project in the latest XCode. 1. Modified the Valid Archetectures to "i386 x86_64" 2. Added the following snippet of code to rb_main.rb NSLog( "In rb_main.rb") i = 1296445544 puts "i = %d class = %s" % [i, i.class.to_s] f = i.to_f puts "f = %f class = %@ DIG = %d" % [ f, f.class, Float::DIG.to_s] if i == f puts "Is equal" else puts "Not equal" end time = Time.new puts "time = %s" % [time.to_s] puts "time to_i = %d" % [time.to_i] puts "time to_f = %f" % [time.to_f] 3. Modified the Architecture to "64 Bit Intel" or "Standard (32/64bit Universal)". Ran the code and got this. 2011-01-30 22:17:19.694 TestFloat[15654:a0f] In rb_main.rb i = 1296445544 class = Fixnum f = 1296445544.000000 class = Float DIG = 15 Is equal time = 2011-01-30 22:17:19 -0600 time to_i = 1296447439 time to_f = 1296447439.697984 3. Modified the Architecture to "32 Bit Universal". Ran the code and got this. 2011-01-30 22:20:47.891 TestFloat[15685:a0f] +[NSATSGlyphGenerator initialize] invocation. The class is deprecated. 2011-01-30 22:20:49.317 TestFloat[15685:a0f] In rb_main.rb i = 1296445544 class = Bignum f = 1296445440.000000 class = Float DIG = 15 Not equal time = 2011-01-30 22:20:49 -0600 time to_i = 1296447649 time to_f = 1296447488.000000 You will notice how the floating point representation has lost a lot of precision in the 32 bit version which is understandable in a much larger number but this number is 10 digits and not 15 digits. This creates big problems when dealing with anything that is dealing with NSDate such as NSCalendar, NSDatePicker, NSTimer, etc You can also reproduce the problem by click the 32 bit button on the Get Info of the 32/64 bit universal version I hope that i am missing something obvious. Please help! On Jan 28, 2011, at 9:00 AM, macruby-devel-request(a)lists.macosforge.org wrote: > > Message: 3 > Date: Fri, 28 Jan 2011 06:02:23 -0600 > From: Richard Sepulveda <rsepulveda2(a)gmail.com> > To: macruby-devel(a)lists.macosforge.org > Subject: [MacRuby-devel] Strange NSDate behavior building 32 bit v 64 > bit > Message-ID: <1258938A-1F77-4150-BEAB-05CD2045DE45(a)gmail.com> > Content-Type: text/plain; charset="us-ascii" > > I am seeing some strange behavior with NSDate, here is my basic code in a new MacRuby project. > > I am running on a MacBook 2009 model, Mac OSX 10.6.6 > > I am using the latest XCode and the MacRuby Framework 0.7 in XCode > > in rbmain.rb: > > NSLog("In rb_main.rb") > puts "NSDate = " + NSDate.date.to_s > NSLog( "NSDate = " + NSDate.date.to_s) > > CCode.printNSDate > > in CCode.m: > > @implementation CCode > > + (void) printNSDate > { > NSLog(@"In printNSDate, NSDate = %@", NSDate.date); > } > > @end > > When compiled: 64-bit Universal, the output is > > 2011-01-28 05:54:16.972 TestRuby[90066:a0f] In rb_main.rb > NSDate = 2011-01-28 05:54:16 -0600 > 2011-01-28 05:54:16.976 TestRuby[90066:a0f] NSDate = 2011-01-28 05:54:16 -0600 > 2011-01-28 05:54:16.979 TestRuby[90066:a0f] In printNSDate, NSDate = 2011-01-28 05:54:16 -0600 > > When compiled: 32-bit Universal, the output is > > 2011-01-28 05:51:23.941 TestRuby[89998:a0f] In rb_main.rb > NSDate = 2011-01-28 05:44:00 -0600 > 2011-01-28 05:51:23.958 TestRuby[89998:a0f] NSDate = 2011-01-28 05:44:00 -0600 > 2011-01-28 05:51:23.961 TestRuby[89998:a0f] In printNSDate, NSDate = 2011-01-28 05:51:23 -0600 > 2011-01-28 05:51:24.565 TestRuby[89998:a0f] Error loading /Library/InputManagers/Smart Crash Reports/Smart Crash Reports.bundle/Contents/MacOS/Smart Crash Reports: dlopen(/Library/InputManagers/Smart Crash Reports/Smart Crash Reports.bundle/Contents/MacOS/Smart Crash Reports, 265): no suitable image found. Did find: > /Library/InputManagers/Smart Crash Reports/Smart Crash Reports.bundle/Contents/MacOS/Smart Crash Reports: GC capability mismatch > 2011-01-28 05:51:25.409 TestRuby[89998:a0f] Error loading /Library/InputManagers/SIMBL/SIMBL.bundle/Contents/MacOS/SIMBL: dlopen(/Library/InputManagers/SIMBL/SIMBL.bundle/Contents/MacOS/SIMBL, 265): no suitable image found. Did find: > /Library/InputManagers/SIMBL/SIMBL.bundle/Contents/MacOS/SIMBL: GC capability mismatch > > > The Obj-C routine prints the correct time. 5:51:23 in both cases. > For 32 bit, the Ruby code prints some time in the past that never changes. > For 64 bit, the Ruby code prints the correct time. > > Any idea's what could be going on?

3 2

Gems again; signing records
by Martin Hawkins 29 Jan '11

29 Jan '11

I know this has been asked before but his is driving me nuts.It' s been a frustrating day; I've been trying to use the UUID gem and have still not been able to 'require' it successfully. I have installed uuid using macgem and have unpacked it to a vendor directory, so I now have : vendor -- macaddr-1.0.0 -- lib -- macaddr.rb -- uuid-2.3.1 -- lib -- uuid.rb Obviously, there's more, but those are important bits. I have tried the following, after googling, with no success, on the basis that the files are 'require'd and once loaded, can be 'require'd again by simple reference: In rb_main.rb $:.unshift File.join(File.dirname(__FILE__), 'Vendor/uuid-2.3.1/lib') $:.unshift File.join(File.dirname(__FILE__), 'Vendor/macaddr-1.0.0/ lib') require 'macaddr' require 'uuid' However, when I try to require 'macaddr' or 'uuid' from another class definition file, I get 'no such file to load'. I've tried setting ENV['GEM_HOME']='/Users/martin/work/macruby/onWeb/ PeepOpen/Vendor' in rb_main.rb; that seemed to make no difference. I've tried using the full path: require '/Users/martin/work/macruby/onWeb/xxxxxx/Vendor/macaddr-1.0.0/ lib/macaddr' require '/Users/martin/work/macruby/onWeb/xxxxxx/Vendor/uuid-2.3.1/lib/ uuid' and this produces a 'no such file to load -- fileutils' - neither gem depends on fileutils ! But hey, I tried it and it refused to build - 'checking for Magick- config... no' So I have two questions: 1. What is the *proper* way of incorporating gems into a MacRuby project using XCode to build and run; and 2. All I want to do is sign records for later comparison. Can anybody suggest an alternative method, other than using UUID? thanks

6 11