[#86755] [Ruby trunk Feature#14723] [WIP] sleepy GC — normalperson@...
Issue #14723 has been reported by normalperson (Eric Wong).
6 messages
2018/04/29
[ruby-core:86423] [Ruby trunk Bug#14642] Fiber make crash on Windows - webrick/httpproxy.rb ?
From:
Greg.mpls@...
Date:
2018-04-01 23:39:42 UTC
List:
ruby-core #86423
Issue #14642 has been updated by MSP-Greg (Greg L). I have tried to isolate the issue with Fiber, but haven't succeeded. Please look at the two below (old) issues, which also involve cont.c (sorry for the duplication): https://bugs.ruby-lang.org/issues/13298 https://bugs.ruby-lang.org/issues/13485 Using a current MinGW build, the following still generates a SEGV (it is one of two patches used in ruby-loco to allow it to pass test-all, the other is readline 'meta key' issue): ruby -W --disable-gems runner.rb -I./lib ruby/test_enum.rb -ntest_callcc The test is passing on Ruby Appveyor & Travis, and RubyCI.org, and also passes on a recent, locally created vc14 build, ruby 2.6.0dev (2018-03-31 trunk 63043) [x64-mswin64_140]. Given that, and given that MinGW is somewhat of a hybrid between *nix and Windows, might cont.c be 'misclassifying' MinGW somewhere in its code? Just a guess, as I don't know c... Finally, for the time being, I have reverted 62966 (see https://github.com/MSP-Greg/ruby-loco/blob/master/patches/gte20600/windows_fiber_webrick_proxy.patch), so ruby-loco can be used for testing more recent commits and CI testing in other repositories. Thanks, Greg ---------------------------------------- Bug #14642: Fiber make crash on Windows - webrick/httpproxy.rb ? https://bugs.ruby-lang.org/issues/14642#change-71357 * Author: usa (Usaku NAKAMURA) * Status: Assigned * Priority: Normal * Assignee: ko1 (Koichi Sasada) * Target version: * ruby -v: trunk r62966 * Backport: 2.3: DONTNEED, 2.4: DONTNEED, 2.5: DONTNEED ---------------------------------------- We found that r62966 causes crash on Windows. Of course, the patch seems to be no problem, so I guess that this code just happened to reveal a potential bug. FYI, AppVayor and mswinci both reported the crash, but they didn't show any details. Stable versions don't contain the patch, but will be backported later, maybe. -- https://bugs.ruby-lang.org/ Unsubscribe: <mailto:[email protected]?subject=unsubscribe> <http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>