aseidl

Twitter thread announcement: https://twitter.com/BigCodeProject/status/158563117635379609...

You can get by with one designed for 2.4GHz wifi if those are easier to acquire. They were < $50 on Amazon when I did my build a year or so ago.

Just have to flatten the front a bit and push it out by about an inch. I used a short piece of pvc pipe and some zipties.

Direct link to the repo: https://github.com/github/roadmap / https://github.com/github/roadmap/projects/1

Announcement: http://googlecloudplatform.blogspot.com/2015/12/Google-Cloud...

Announcement: http://devblogs.nvidia.com/parallelforall/digits-deep-learni...

Correct, the speedups are directly due to bundle caching (which previously had only been available for private/paid repos). Indirectly, I would assume this caching is now implemented via Docker's use of filesystem layers.

Some rough numbers: bundle install has gone from ~50s to 3s for my Jekyll + s3_website setup. The overall build has dropped from ~120s to 45s thanks to that and using a cached version of Ruby.

One gotcha for anyone with a similar setup: make sure Jekyll's _config.yml is set to exclude/ignore the vendor directory. Oops.

If anyone else is interested, it sounds like HDMI-CEC should be working within the next week or so.

http://forum.odroid.com/viewtopic.php?f=61&t=7479#p58825

It looks like prices are around the bulk D or E rates from Gandi (their upstream registrar). One notable exception is .io, which is currently $78 on AWS, but only $39 from Gandi (A rate).

Edit: Pricing list is available at: https://d32ze2gidvkk54.cloudfront.net/Amazon_Route_53_Domain...

The only problem with -W is that it is not available everywhere yet: RHEL6 defaults to OpenSSH 5.3, while -W was introduced in 5.4. Thankfully EL7 comes with OpenSSH 6.4 (and a kernel newer than 2.6.32!).

The data might be coming from the .com/.net/.org zonefiles, where a domain will only show up if it has an NS record configured.

I'm not sure if there's a free and authoritative source for registered domains, short of grep'ing zonefiles and checking whois.

Have you looked into Beets (http://beets.radbox.org/) or the classic Picard (http://musicbrainz.org/doc/MusicBrainz_Picard)?

A few suggestions:

SCP'ing the public key directly to ~/.ssh/authorized_keys could overwrite the file if it already exists. It's better to use 'ssh-copy-id', which takes care of everything for you. Most distros should have it already.

Autossh can take care of setting up, monitoring, and restarting the tunnels for you. Arguments are almost exactly the same as ssh, just toss it into your crontab and have it run @reboot.

I've rerun the tests on some slightly more recent hardware: a Tesla K20c and Xeon E5-2630. See the 'new benchmarks' section on: https://github.com/aseidl/Sort_data_parallel

I'm working on running the tests on 2x Xeon E5-2630 and a Tesla K20c. Will post the results here in a couple hours.

I've made it a habit to ping and dig my servers before signing in to a new wireless network. Of the hundreds of coffee shops, trains, planes, airports, etc that I've done this at, there has only been one network that blocked/rerouted both ICMP and DNS requests: a university that shared its network security team with a nearby national lab.

However, exploiting these holes on a wireless network is incredibly easy to detect and block for an admin worth their salt. It's quite likely that at least a few of the networks I've been on would start blocking traffic from an unregistered device making tons of DNS requests.

That being said, just pay the few dollars they charge for access. If your time is money, this small fee won't be noticed. If you just want to be able to refresh Reddit/lurk on HN, maybe you should take this opportunity to get away from technology for a few hours (while sitting in an aluminum tube hurdling through the skies).

A number of the sponsoring orgs will grant you access to their zonefiles if you're willing to do a little paperwork (HelloFax is awesome) and have a valid reason for access. Granted it's not the complete list of registered domains (those without nameservers on file won't show up), but it's pretty close.

The list of sponsoring orgs for each TLD is at: http://www.iana.org/domains/root/db

The issue may have to do with NoScript or some other JavaScript blocker as the method appears to rely upon JS. (Tested in Lynx and Firefox with JS disabled.)

Jhsto: you might want to include a noscript tag warning visitors about this. Of course it might be better to have this done server side as anything client side will eventually get worked around.

For those who may have missed it (I certainly did), this came only a few days after the 1.5.0 release.

http://rhodecode.org/news/2012/rhodecode-150-released/

Makes sense, you want to make it as clear as possible to see that they're different. My thinking was that if you're already differentiating by size and weight, why introduce another variation that will take an extra fraction of a second to process?

Going beyond the not so great font, is it really a good idea to mix serif fonts with sans serif on the same page, specifically having (sub)headings one and main content the other? I've noticed this in a few places lately, but it doesn't seem very clean to me.

I'm not sure what the current best stack is for email (last time I hosted my own, the setup involved UUCP), but my ideal setup would be as follows:

- continue using Gmail as the main account

- have a catch-all on my server which simply forwards to the Gmail account

- for sending mail, have a regexp that when matched, will rewrite the 'from' username to that

As an example of the last point, if my subject has &hn& at the beginning, remove that string and change the from address to 'hn@example.com'.

I believe this should be possible with postfix. What I'm not sure of is if Gmail will flag my servers for the steady stream of spam being forwarded (I'd prefer to let Gmail's filters handle all spam).

Jekyll/Octopress+Git http://octopress.org/