Re: [Cake] flent crash in head

[moeller0 <moeller0@gmx.de>]

Hi Toke,


> On Dec 21, 2015, at 21:56 , Toke Høiland-Jørgensen <toke@toke.dk> wrote:
> 
> moeller0 <moeller0@gmx.de> writes:
> 
>>> Well, just fixed the rrul_50_down and a bunch of other tests that had
>>> broken use of the 'duplicates' parameter. Shouldn't crash now, but won't
>>> plot the individual flows
>> 
>> After a git pull the "crashes" are gone, sorry, I was not precise
>> enough, python crashed once, but the biggest issue was a failure to
>> load the files.
> 
> Cool.
> 
>>> (doing, essentially, `sed s/::::/::/g` on the
>>> contents of the data file should fix that; new tests will be fine).
>> 
>> 	Ah, I guess I need to unzip the files first?
> 
> Yup. Something like this:
> 
> zcat rrul_50_down-2015-12-21T054402.343858.cake_flows.flent.gz | sed s/::::/::/g | gzip > rrul_50_down-2015-12-21T054402.343858.cake_flows.fixed.flent.gz

	Ah, sorry for not writing back, I just went and used gz to unzip them, applied the sed line you gave and used gz to zip them again (actually I used mc to select the bunch and at least do the unzipping and zipping for the whole batch). It just worked, and I now see way more lines than I actually can visually parse ;)

> 
>>> 
>>> Can't reproduce the crash Sebastian reported…
>> 
>> This only happened once, and I would blame it on my macports installed
>> python/matplotlib…
> 
> Yes, it did seem to be something related to PyQt that I'm hoping we can
> just write off as a fluke :)

	Sure, I could not make it crash again, so noting really to work on ;)

Best Regards
	Sebastian

> 
> -Toke