[#85940] [Ruby trunk Bug#14578] Forking a child process inside of a mutex crashes the ruby interpreter — ben.govero@...
Issue #14578 has been reported by bengovero (Ben Govero).
3 messages
2018/03/05
[#86205] [Ruby trunk Feature#14618] Add display width method to String for CLI — aycabta@...
Issue #14618 has been reported by aycabta (aycabta .).
3 messages
2018/03/19
[#86366] Re: [ruby-cvs:70102] usa:r63008 (trunk): get rid of test error/failure on Windows introduced at r62955 — Eric Wong <normalperson@...>
usa@ruby-lang.org wrote:
3 messages
2018/03/28
[ruby-core:86369] Re: [ruby-cvs:70102] usa:r63008 (trunk): get rid of test error/failure on Windows introduced at r62955
From:
Eric Wong <normalperson@...>
Date:
2018-03-28 16:41:37 UTC
List:
ruby-core #86369
"U.NAKAMURA" <usa@garbagecollect.jp> wrote: > Hi, Eric > > In message "Re: [ruby-cvs:70102] usa:r63008 (trunk): get rid of test error/failure on Windows introduced at r62955" > on Wed, 28 Mar 2018 16:21:06 +0000, normalperson@yhbt.net wrote: > > I guess Windows does not have pread? > > Yes. > > > > For current C Ruby, lock/unlock can use GVL. So I think that > > workaround can be implemented for Windows in IO.copy_stream > > itself, and not WEBrick. > > The document of IO.copy_stream says: > > IO.copy_stream doesn't move the current file offset. > > I don't have any idea to implement such behavior for Windows. Sorry, I left out one syscall (ENOCOFFEE): > > # only reliable facsimile for pread iff FDs are not shared across fork > > lock > > lseek > > read lseek back to original position > > unlock Unsubscribe: <mailto:ruby-core-request@ruby-lang.org?subject=unsubscribe> <http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>