>> On Sun, Dec 20, 2009 at 5:31 PM, Nigel Cunningham
>> <ncunningham@xxxxxxxxxxx> wrote:
>>> Hi.
>>>
>>> Michael Evans wrote:
>>>> On Sun, Dec 20, 2009 at 5:19 PM, Nigel Cunningham
>>>> <ncunningham@xxxxxxxxxxx> wrote:
>>>>> Ooooooooh.
>>>>>
>>>>> I'm glad you did all that recording, because you hit a bug I've been
>>>>> trying to find the cause of for aaaaages. I'm going to be staring at
>>>>> this email veeeeeeery carefully! Must stop hitting those vowel keys
>>>>> repeatedly, though. How easily can you reproduce that?
>>>>>
>>>>> Nigel
>>>>>
>>>> I really am not sure. This was the first time I've noticed that
>>>> particular failure. Any suggestions on configuration or code
>>>> bisections?
>>> Not yet - still reading carefully. And I should be getting on with other
>>> things, so I might be a while (sorry!). I wouldn't worry too much about
>>> reproducing it. I'm hoping I have enough here to figure it out.
>>>
>>> The double logging just makes it a bit harder to read.
>>>
>>> Thanks!
>>>
>>> Nigel
>>>
On the very next try it happened again. I'm using this perl script to
filter out most of the more complicated duplicate cases.
I'll look at this later, I really need to take care of something else.
./plfilter.pl < cycle2.log | sed -e '/Writing: device/d;/Reading:
device/d' > cycle2.txt
#!/usr/bin/perl
use strict;
use warnings;
my @used;
my $ii;
for ($ii=8; $ii > 0; $ii--) {
$used[$ii - 1] = "";
}
outer: foreach (<main::stdin>) {
for ($ii=0; $ii < 8; $ii++) {
next outer if ($_ eq $used[$ii]);
}
for ($ii=8; $ii > 0; $ii--) {
$used[$ii] = $used[$ii - 1];
}
$used[0] = $_;
print $_;
}
Attachment:
cycle2.txt.lzma
Description: application/lzma
