Re: [racket] Problems trying to do non-blocking I/O and a reactor loop

2012-08-04 Thread Neil Van Dyke
Matthew Flatt wrote at 08/04/2012 12:33 PM: At Fri, 3 Aug 2012 20:51:37 -0400, Greg Hendershott wrote: 3. Bad news: I just got it to happen again, with 5.3.0.16: I see how that could happen, and I've pushed a repair --- mostly by improving the guarantees about progress evts. (I think

Re: [racket] Problems trying to do non-blocking I/O and a reactor loop

2012-08-04 Thread Neil Van Dyke
Greg Hendershott wrote at 08/03/2012 08:51 PM: I wonder can anyone else elicit these problems trying to run it, too? Your test seems to fail for me (on Debian Stable GNU/Linux, 32-bit x86) after around 26K to 27K requests (with "-c 10"), using a pre-release Racket from several days ago.

Re: [racket] Problems trying to do non-blocking I/O and a reactor loop

2012-08-04 Thread Matthew Flatt
At Sat, 4 Aug 2012 09:33:09 -0400, Sam Tobin-Hochstadt wrote: > On Sat, Aug 4, 2012 at 6:24 AM, Neil Van Dyke wrote: > > Matthew Flatt wrote at 08/03/2012 07:56 PM: > > > >> I see that you're using `read-bytes-avail!-evt', which has problems in > >> v5.2.1 that are fixed for v5.3. The problems inc

Re: [racket] Problems trying to do non-blocking I/O and a reactor loop

2012-08-04 Thread Matthew Flatt
At Fri, 3 Aug 2012 20:51:37 -0400, Greg Hendershott wrote: > 1. Bad news: The reqs/sec seem much lower than before. i.e. Does the > bug fix have a performance hit? That's not expected. It's possible that the broken version took shortcuts that happened to work out much of the time, but I'm not sure

Re: [racket] Problems trying to do non-blocking I/O and a reactor loop

2012-08-04 Thread Sam Tobin-Hochstadt
On Sat, Aug 4, 2012 at 6:24 AM, Neil Van Dyke wrote: > Matthew Flatt wrote at 08/03/2012 07:56 PM: > >> I see that you're using `read-bytes-avail!-evt', which has problems in >> v5.2.1 that are fixed for v5.3. The problems include triggering a bug >> in `sync', which is also fixed for v5.3. >> > >

Re: [racket] Problems trying to do non-blocking I/O and a reactor loop

2012-08-04 Thread Neil Van Dyke
Matthew Flatt wrote at 08/03/2012 07:56 PM: I see that you're using `read-bytes-avail!-evt', which has problems in v5.2.1 that are fixed for v5.3. The problems include triggering a bug in `sync', which is also fixed for v5.3. I am extremely interested in these bugs. How do I find more info

Re: [racket] Problems trying to do non-blocking I/O and a reactor loop

2012-08-03 Thread Greg Hendershott
The results so far: 1. Bad news: The reqs/sec seem much lower than before. i.e. Does the bug fix have a performance hit? 2. Good news: The "input port closed" error seemed to be gone with 5.3.0.16. Except ... 3. Bad news: I just got it to happen again, with 5.3.0.16: $ /Applications/Racket_v5.3

Re: [racket] Problems trying to do non-blocking I/O and a reactor loop

2012-08-03 Thread Greg Hendershott
Thank you for such a quick reply! > I see that you're using `read-bytes-avail!-evt', which has problems in > v5.2.1 that are fixed for v5.3. The problems include triggering a bug > in `sync', which is also fixed for v5.3. Well, I only tried read-bytes-avail!-evt after getting similar errors with

Re: [racket] Problems trying to do non-blocking I/O and a reactor loop

2012-08-03 Thread Matthew Flatt
I see that you're using `read-bytes-avail!-evt', which has problems in v5.2.1 that are fixed for v5.3. The problems include triggering a bug in `sync', which is also fixed for v5.3. Does a nightly build behave any differently? At Fri, 3 Aug 2012 19:11:44 -0400, Greg Hendershott wrote: > Even the