ruby-dev

小松です．

drb-1.2.1を使っているときに出た問題なのですが，Ruby 1.6.0で
Solaris 2.5.1だけsocketとmd5をこの順でrequireすると，
MD5の値が間違っていたり毎回変わったりするという現象がおきていて，
悩んでいます．

しかも同じバイナリをSolaris 2.6で動かすとOKで，
他のプラットフォームでは再現しません．

コンパイラかと思って，gcc 2.8.1，egcs 1.1.2，gcc 2.95.2で"-g -O2"と
"-g"を，WorkShop Compiler 4.2 (2世代前，最新は6.0) で"-g"と"-O"を
試してみましたが，毎回値が変化するかどうかが変わるぐらいで間違って
いることは変わりありませんでした．

Ruby 1.4.6でも同様でした．

うーん，きっとSolarisのシェアードライブラリのローダーのバグ
なんだろうな．
しょうがないのでmd5をスタティックリンクして回避しています．

参考まで．

□Solaris 2.5.1, egcs 1.1.2, -g -O2
% uname -mrs
SunOS 5.5.1 sun4u
% ruby-1.6.0 -v -rsocket -rmd5 -e '5.times { p MD5.new("0").hexdigest }'
ruby 1.6.0 (2000-09-19) [sparc-solaris2.5.1] ← egcs 1.1.2で-g -O2で作成
"000001c0000000009bf3a3bb9b0a90a0" ← 値がおかしい上に毎回変わる
"000003c00000000089042cb03434dbf5"
"000005c0000000002fe81e55e2d8729b"
"000007c0000000000b90924f87a0f3d9"
"000009c000000000c0a253b7a1f11083"
% ruby-1.6.0 -v -rsocket -rmd5 -e '5.times { p MD5.new("0").hexdigest }'
ruby 1.6.0 (2000-09-19) [sparc-solaris2.5.1]
"000001c0000000009bf3a3bb9b0a90a0" ← 同一マシンとバイナリでは1回目のRubyの
"000003c00000000089042cb03434dbf5"    出力パターンと2回目のRubyの出力の
"000005c0000000002fe81e55e2d8729b"    パターンは同じ
"000007c0000000000b90924f87a0f3d9"
"000009c000000000c0a253b7a1f11083"
% ruby-1.6.0 -v -rmd5 -rsocket -e '5.times { p MD5.new("0").hexdigest }'
ruby 1.6.0 (2000-09-19) [sparc-solaris2.5.1] ← egcs 1.1.2で-g -O2で作成
"cfcd208495d565ef66e7dff9f98764da" ← -rmd5を先にもっていくと-rsocketが
"cfcd208495d565ef66e7dff9f98764da"    あっても正常
"cfcd208495d565ef66e7dff9f98764da"
"cfcd208495d565ef66e7dff9f98764da"
"cfcd208495d565ef66e7dff9f98764da"
% 

□Solaris 2.6, egcs 1.1.2, -g -O2
% uname -mrs
SunOS 5.6 sun4u
% ruby-1.6.0 -v -rsocket -rmd5 -e '5.times { p MD5.new("0").hexdigest }'
ruby 1.6.0 (2000-09-19) [sparc-solaris2.5.1] ← バイナリは上と同じもの
"cfcd208495d565ef66e7dff9f98764da" ← Solaris 2.6では-rsocketが先に
"cfcd208495d565ef66e7dff9f98764da"    あっても正常
"cfcd208495d565ef66e7dff9f98764da"
"cfcd208495d565ef66e7dff9f98764da"
"cfcd208495d565ef66e7dff9f98764da"
% 

--
小松克行 / Katsuyuki Komatsu <komatsu@sarion.co.jp>

Thread

Prev Next

In This Thread

Prev Next

[#10781] MatchingData#[] — Shugo Maeda <shugo@...>

[#10793] 今度こそ (patch of the ruby-1.4.6 for NT4.0&VC4.0 on DEC Alpha.) — kou@...1609.sip.eee.yamaguchi-u.ac.jp (Koichi Okada)

[#10812] empty rescue clause — Takaaki Tateishi <ttate@...>

[#10814] Re: ruby 1.6.0 preview2 — tenmei@... (Noriaki Harada)

[#10829] [BUG] re_search with range — Minero Aoki <aamine@...>

[#10856] typo in finalize.rb — "Nobuyoshi.Nakada" <nobu.nakada@...>

[#10861] MatchingData#last_group — Shugo Maeda <shugo@...>

[#10864] a = []; p a.sort! — "Akinori -Aki- MUSHA" <knu@...>

[#10866] Class Variable (sono 2) — Shin-ichiro HARA <sinara@...>

[#10879] recursive require / preload — Yasushi Shoji <yashi@...>

[#10882] setenv,unsetenv — Takaaki Tateishi <ttate@...>

[#10889] Re: mod_ruby error? — Wakou Aoyama <wakou@...>

[#10890] Re: Telnet#preprocess — Wakou Aoyama <wakou@...>

[#10891] Re: net/telnet.rb — Wakou Aoyama <wakou@...>

[#10892] Re: lib/cgi.rb — Wakou Aoyama <wakou@...>

[#10893] Re: lib/cgi.rb — Wakou Aoyama <wakou@...>

[#10894] Forward: [ruby-ext:01274] Re: 多重代入での to_a (Re: Ruby/Python doesn't work with Ruby 1.6) — matz@... (Yukihiro Matsumoto)

[#10897] Re: [CVS 266] "matz: 1.6.0 final (hopufully)" — WATANABE Hirofumi <Hirofumi.Watanabe@...>

[#10903] Class Variable と継承 — Kazuhiro NISHIYAMA <zn@...>

[#10904] Module#class_variables — Yasushi Shoji <yashi@...>

[#10907] attr_tester — Yasushi Shoji <yashi@...>

[#10910] tsvr.rb broken ? — Katsuyuki Komatsu <komatsu@...>

[#10911] time.rb — Katsuyuki Komatsu <komatsu@...>

[#10920] SIGINT on windows — "Nobuyoshi.Nakada" <nobu.nakada@...>

[#10921] proper tail recursion — Shugo Maeda <shugo@...>

[#10924] instance methods of Module/Class — Hiroshi IGARASHI <igarashi@...>

[#10928] man ruby — Takaaki Tateishi <ttate@...>

[#10933] Meta-Level API — Hiroshi IGARASHI <igarashi@...>

[#10943] Passing []/String as argument — Jun Adachi <adachi@...>

[#10944] dummy DLL on Windows — "Nobuyoshi.Nakada" <nobu.nakada@...>

[#10946] require "marshal" failure — Katsuyuki Komatsu <komatsu@...>

[#10948] File::Constants — Yasushi Shoji <yashi@...>

[#10952] negative signal number — GOTOU YUUZOU <gotoyuzo@...>

[#10956] server maintenance — Shugo Maeda <shugo@...>

[#10959] match data — 助田 雅紀 <masaki.suketa@...>

[#10968] jcode.rb bug — "Okada Jun" <yun@...>

[#10978] [PATCH] require in require — "Nobuyoshi.Nakada" <nobu.nakada@...>

[#10985] httphead.rb proxy version problem — Katsuyuki Komatsu <komatsu@...>

[#10988] [patch] String#index( sub, pos, range ) — Minero Aoki <aamine@...>

[#11006] lstat — WATANABE Hirofumi <eban@...>

[#11028] regex.c warning of alpha — Katsuyuki Komatsu <komatsu@...>

[#11037] RUBYLIB_PRIFIX at emx/gcc — kenn@...

[#11049] [PATCH] RTEST() value as boolean — nobu.nakada@...

[#11052] ruby-1.6.1-20000921-i386-cygwin and Thread — Shin-ichiro HARA <sinara@...>

[#11061] [PATCH] prefix for sitedir — "Nobuyoshi.Nakada" <nobu.nakada@...>

[#11064] jcode.rb bug — KANEKO Naoshi <wbs01621@...>

[#11066] [PATCH] gcc extension in regex.c — "Nobuyoshi.Nakada" <nobu.nakada@...>

[#11073] expect.rb — Kazuhiro NISHIYAMA <zn@...>

[#11074] 定数のネスト — Kazuhiro NISHIYAMA <zn@...>

[#11078] test — Shugo Maeda <shugo@...>

[#11082] LIBPATH patch for emx/gcc — kenn@...

[#11083] Application Error in 1.6.0 on Win2K — matz@... (Yukihiro Matsumoto)

[#11090] Ruby CVS Repository — "Akinori -Aki- MUSHA" <knu@...>

[#11091] ruby 1.6.1 mingw32 bug? — kenn@...

[#11093] gsub! : not substitute if block returns nil — Masahiro Tanaka <masa@...>

[#11098] mod_ruby error? — akira yamada / やまだあきら <akira@...>

[ruby-dev:10972] require socket and md5

Thread

In This Thread

[#10959] match data — 助田　雅紀 <masaki.suketa@...>