From: Eric Wong Date: 2017-10-18T23:30:17+00:00 Subject: [ruby-core:83359] Re: [Ruby trunk Bug#14013] [PATCH] Webrick 60172 fix ``` Greg.mpls@gmail.com wrote: > Bad news. Appveyor just failed with: Oops, try r60210; I moved the return outside of the timeout block. > Re the patch you included above, can you please bracket it > with triple backticks (\`\`\`), with the first being > \`\`\`diff ? May need a blank line before the first set. I > view this on the web; the markdown parser makes a mess of diff > listings without the backticks. I'll try to remember in the future. I remember there was a time when redmine didn't try to assume emails were markdown and always treaded them as preformatted (I either use email or w3m, so no variable-width fonts). > Re the `#nil?`, the docs/comments do state returns 'nil on > timeout'. Maybe we should remove the `WEBrick::Utils.timeout` > block and put the timeout value in the `#wait_*` methods? I've considered it, but it requires accounting for total time spent in case we need to call accept_nonblock multiple times; so it's extra math + logic which I don't want... I actually want to improve 'timeout' in stdlib and implement it in the core VM so it affects all methods like IO.select and IO#wait*able, and even IO#read transparently. ``` Unsubscribe: