December 2009 - MacRuby-devel - lists.macosforge.org

[MacRuby] #364: priority queue program crashes MacRuby
by MacRuby 26 Mar '10

26 Mar '10

#364: priority queue program crashes MacRuby -------------------------------------+-------------------------------------- Reporter: jordan.breeding@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: MacRuby 0.5 Component: MacRuby | Keywords: -------------------------------------+-------------------------------------- So, I have been trying to track down a crash in MacRuby for some time. The crash is inside of a priority queue program I wrote earlier this year for an artificial intelligence class. In MacRuby 0.4 the program was non- deterministic but would run. Under MacRuby 0.5 (trunk) it currently doesn't run at all, crashing instead. crash text: {{{ 126 jordan@thetourist ...tificial Intelligence/Assignment No. 1/priority_queue_bug > ./search.cocoa -m -c a-star input_files/input13_7.txt uncaught Objective-C/C++ exception... terminate called after throwing an instance of 'RoxorReturnFromBlockException*' zsh: abort ./search.cocoa -m -c a-star input_files/input13_7.txt }}} The current search.cocoa file will only run under MacRuby, but it could probably be made to work on RubyCocoa. search.pqueue will work under Ruby 1.8.6 and Ruby 1.9, but crash with the same error under MacRuby. I originally wanted to narrow this down to a smaller, more management test case that could easily be turned into a test for MacRuby's test suite, but everything I try works as a small test even though the original larger program crashes. Since I have had no luck tracking this down for sometime I am just uploading it as a bug. -- Ticket URL: <http://www.macruby.org/trac/ticket/364> MacRuby <http://macruby.org/>

1 3

[MacRuby] #431: Seg fault evaluating an escaped unicode string
by MacRuby 25 Mar '10

25 Mar '10

#431: Seg fault evaluating an escaped unicode string ----------------------------+----------------------------------------------- Reporter: jazzbox@… | Owner: lsansonetti@… Type: defect | Status: new Priority: minor | Milestone: MacRuby 0.5 Component: MacRuby | Keywords: ----------------------------+----------------------------------------------- {{{ $ macruby -e '"\xFC\u20AC"' Segmentation fault }}} the € only string does not crash, nevertheless the output is not correct: {{{ $ macruby -e 'p "\u20AC"' "¬" }}} while ruby1.9 is OK: {{{ $ ruby1.9 -e 'p "\u20AC"' "€" }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/431> MacRuby <http://macruby.org/>

1 3

[MacRuby] #454: String#getbyte not working?
by MacRuby 25 Mar '10

25 Mar '10

#454: String#getbyte not working? ------------------------------+--------------------------------------------- Reporter: paul@… | Owner: lsansonetti@… Type: defect | Status: new Priority: minor | Milestone: MacRuby 0.5 Component: MacRuby | Keywords: ------------------------------+--------------------------------------------- It appears the new String method called getbyte for Ruby 1.9 is not working? Results in an "Abort trap" message. ~ $ macruby --version MacRuby version 0.5 (ruby 1.9.0) [universal-darwin10.0, x86_64] ~ $ macruby -e ' "abc".getbyte(0);' Abort trap -- Ticket URL: <http://www.macruby.org/trac/ticket/454> MacRuby <http://macruby.org/>

2 5

[MacRuby] #192: String#length and String#size do not work correctly
by MacRuby 25 Mar '10

25 Mar '10

#192: String#length and String#size do not work correctly ----------------------------------------+----------------------------------- Reporter: vincent.isambart@… | Owner: lsansonetti@… Type: defect | Status: new Priority: critical | Milestone: Component: MacRuby | Keywords: ----------------------------------------+----------------------------------- String#length and String#size return the byte size instead of the number of characters in the string. {{{ # coding: utf-8 p 'あ'.length p 'あ'.size p 'あ'.bytesize }}} Ruby 1.9: {{{ 1 1 3 }}} MacRuby: {{{ 3 3 3 }}} For #size, the function is declared normally in string.c but for #length, you have to remove it from opt_length in insns.def. -- Ticket URL: <http://www.macruby.org/trac/ticket/192> MacRuby <http://macruby.org/>

1 3

[MacRuby] #116: "%s" inside of format statements in macruby doesn't work with multibyte characters
by MacRuby 25 Mar '10

25 Mar '10

#116: "%s" inside of format statements in macruby doesn't work with multibyte characters -------------------------------------+-------------------------------------- Reporter: jordan.breeding(a)mac.com | Owner: lsansonetti(a)apple.com Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -------------------------------------+-------------------------------------- this: {{{ printf("straight puts: ") puts(p["name"]) printf("straight printf: ") printf("%s\n", p["name"]) printf("string formatted puts: ") puts("%s" % [p["name"]]) printf("nsstring printf: ") printf("%@\n", p["name"]) printf("nsstring formatted puts: ") puts("%@" % [p["name"]]) printf("format using nsstring, then puts: ") puts(NSString.stringWithFormat("%@", p["name"])) }}} will produce this: {{{ straight puts: Surgeon’s Girl straight printf: Surgeon‚Äôs Girl string formatted puts: Surgeon‚Äôs Girl nsstring printf: Surgeon’s Girl nsstring formatted puts: Surgeon’s Girl format using nsstring, then puts: Surgeon’s Girl }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/116> MacRuby <http://macruby.org/>

1 3

[MacRuby] #436: Strange pair(?) of bugs
by MacRuby 25 Mar '10

25 Mar '10

#436: Strange pair(?) of bugs ------------------------------------+--------------------------------------- Reporter: macruby@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: MacRuby 0.5 Component: MacRuby | Keywords: ------------------------------------+--------------------------------------- I am writing some code, see attached, in ruby and have run into a strange pair of bugs, I think. I'm using 'MacRuby version 0.5 (ruby 1.9.0) [universal-darwin10.0, x86_64]' When I go to execute the attached code I have come across two problems which I think are bugs. First when I run this line [line 20] of code 'd = Test.new(9.1)' I get the following response '##-<Test:0x200274f20 @min=9.09999999999999>' which is wrong. Second, when I execute this line [line 24] of code 'c << d' I get '[##-<Test:0x2000e5e60 @min=10.0>, ##-<Test:0x20020d2e0 @min=10.1>, ##-<Test:0x200274f20 @min=9.09999999999999>]' which is also wrong. It should look like this '[##-<Test:0x200274f20 @min=9.09999999999999>, ##-<Test:0x2000e5e60 @min=10.0>, ##-<Test:0x20020d2e0 @min=10.1>]' which it does if I execute this line of code 'c.sort!'. Executing all of this in 'ruby 1.8.7 (2008-08-11 patchlevel 72) [universal-darwin10.0]' returns the expected results. - AYAL -- Ticket URL: <http://www.macruby.org/trac/ticket/436> MacRuby <http://macruby.org/>

1 1

[MacRuby] #126: bytestrings, nsstring, and unicode/multibyte characters do not get along
by MacRuby 25 Mar '10

25 Mar '10

#126: bytestrings, nsstring, and unicode/multibyte characters do not get along -------------------------------------+-------------------------------------- Reporter: jordan.breeding(a)mac.com | Owner: lsansonetti(a)apple.com Type: defect | Status: new Priority: blocker | Milestone: MacRuby 0.4 Component: MacRuby | Keywords: -------------------------------------+-------------------------------------- I will try to publish some example later, but I have seen problems trying to print bytestrings. Unicode has issues with "%s" so I tried using "%@" in most places. That worked until I tried using a bytestring, bytestrings don't get along with "%@". Most of this was outside of the bug tracker in emails with Laurent. -- Ticket URL: <http://www.macruby.org/trac/ticket/126> MacRuby <http://macruby.org/>

1 1

[MacRuby] #308: "too short multibyte code string" when using Pathname
by MacRuby 25 Mar '10

25 Mar '10

#308: "too short multibyte code string" when using Pathname -------------------------------+-------------------------------------------- Reporter: tjw@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -------------------------------+-------------------------------------------- Possibly related, but simpler case of https://www.macruby.org/trac/ticket/303 See the attached script, which hits a regex parse error when run. Digging around, it looks like this might start from lib/pathname.rb in #initialize: if /\0/ =~ @path raise ArgumentError, "pathname contains \\0: #{(a)path.inspect}" end The /\0/ fails to parse and a RegexpError error is thrown. -- Ticket URL: <http://www.macruby.org/trac/ticket/308> MacRuby <http://macruby.org/>

1 2

[MacRuby] #179: Calling Regexp#match from ObjC in a background thread hangs
by MacRuby 25 Mar '10

25 Mar '10

#179: Calling Regexp#match from ObjC in a background thread hangs -------------------------------+-------------------------------------------- Reporter: tjw@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -------------------------------+-------------------------------------------- Calling this from a foreground thread works fine, but if I call Regexp#match from ObjC in a background thread, it hangs. Attaching example. -- Ticket URL: <http://www.macruby.org/trac/ticket/179> MacRuby <http://macruby.org/>

1 1

[MacRuby] #244: String.name should return "String" and not "NSMutableString"
by MacRuby 25 Mar '10

25 Mar '10

#244: String.name should return "String" and not "NSMutableString" -------------------------------+-------------------------------------------- Reporter: carllerche@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -------------------------------+-------------------------------------------- macirb >> String.name # => "NSMutableString" I believe that it is a ruby expectation that core classes will have the same name on any ruby implementation. Extlib and other libraries rely on this fact. -- Ticket URL: <http://www.macruby.org/trac/ticket/244> MacRuby <http://macruby.org/>

1 3