Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
bifrost
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
69 ms
·
841.
▲
by
bifrost
8y ago
I'm def not in the camp of being PO'ed by that but I understand why some people are. One thing I learned long ago was that there were "dieters" and "ethical vegans" and only sometimes did they overlap. The avai
842.
▲
by
bifrost
8y ago
I'm referring to the process of cooking it. They don't separate the food and you get cross-contact with meat and dairy proteins - making it non-vegetarian and definately not vegan. This is basically all marketing driven (which is
843.
▲
by
bifrost
8y ago
Not if it has meat juice on it due to how its prepared. I realize some people are willing to be handwavey about it...
844.
▲
by
bifrost
8y ago
Its kinda misleading, there's still animal products in this because of how the food is prepared...
845.
▲
by
bifrost
8y ago
If you clone it, are you really doing something you want to be doing?
846.
▲
by
bifrost
8y ago
This is pretty cool! Your last article was great and really helped provide some insight into how hard (and easy) it is to do this.
847.
▲
by
bifrost
8y ago
Probably everything you have stored in Docker Hub.
848.
▲
by
bifrost
8y ago
That is unlikely to ever happen...
849.
▲
The Best Way to Avoid Data Leaks and Privacy Scandals: Don't Own Consumer Data
(forbes.com)
2 points
by
bifrost
8y ago
|
0 comments
850.
▲
by
bifrost
8y ago
I'm a total convert. I won't buy GHE again.
851.
▲
by
bifrost
8y ago
Because they get first pick of the good interns and they're still better than a couple of the other big places to work for.
852.
▲
by
bifrost
8y ago
I don't think thats true TBH. I've worked on some pretty huge projects with both and I think it depends on the underlying server. SVN does seem to be a bit more "chatty" to the server than git so its more sensitive to la
853.
▲
by
bifrost
8y ago
I think you're probably right. I've used rcs,cvs, p4, svn and git. I don't particularly have a favorite but I think git's security model is basically nonexistent BUT that lack of security basically lead to the genesis of
854.
▲
by
bifrost
8y ago
Uh, its always been like this. AWS has never had a truly "standards compliant" platform. When you go AWS, you are vendor locked.
855.
▲
Subversion vs. Git: Myths and Facts
(svnvsgit.com)
6 points
by
bifrost
8y ago
|
4 comments
856.
▲
by
bifrost
8y ago
There are a bunch of brokerages available, but you can also find sites that will pay you to park your domain with them. I used estibot to get an appraisal, it was a bit low but reasonable for a free service. FWIW the domain is fairly specif
857.
▲
by
bifrost
8y ago
Someday we will have tools that save us from human problems!
858.
▲
by
bifrost
8y ago
I agree with you 100% here, I would not deploy any CI publicly unless its heavily fenced off into "read only" territory.
859.
▲
by
bifrost
8y ago
Very Cool! I'll check it out!
860.
▲
by
bifrost
8y ago
I've never claimed it was a "one stop", but it certainly keeps the random internet users to a minimum. And yes, using GHE or self hosted GitLab doesn't make up for storing secrets, but it at least keeps them out of the p
861.
▲
by
bifrost
8y ago
I should clarify this comment a bit since it seems to be the most controversial. When I say the attacker was persisted via CM, I'm pointing at his own notes, nodding to broken CM, the requirements of supporting the CM and availability
862.
▲
by
bifrost
8y ago
Not quite. In this case Ansible is enabled by a user logging into the system; The user should not be allowed to login to the system in the first place. Ideally you want your configuration system indepenent of user logins. Ansible has its pl
863.
▲
by
bifrost
8y ago
Correct, the minion runs as root and doesn't require interactive ssh access. Its controlled by a remote master that you ideally properly protect.
864.
▲
by
bifrost
8y ago
baroffoos: Thats pretty close to what I'm suggesting, no public repo access. It works.
865.
▲
by
bifrost
8y ago
I'll vote yes for this one :)
866.
▲
by
bifrost
8y ago
It wasn't CloudFlare's API not revoking a token, they just didn't revoke all the tokens. Basically human error. "The API key was known compromised in the original attack, and during the rebuild the key was theoretically
867.
▲
by
bifrost
8y ago
I've mostly seen this reported in forums and during discussion, if you Google around you'll find some pretty useful hits. Here's a good one from reddit: https://www.reddit.com/r/github/comments/
868.
▲
by
bifrost
8y ago
My 5 second lazy summaries of the CVEs: CVE-2019-1003001, CVE-2019-1003002 -> Anyone with read access to Jenkins can own the build environment. CVE-2019-1003000 -> I didn't get a lot of the details on this but it basically looks
869.
▲
by
bifrost
8y ago
Ever seen anyone working in a Coffee shop using one? Me neither. Good security technology exists, the problem is that people don't want to use it because its easier to ignore it.
870.
▲
by
bifrost
8y ago
I guess I technically glossed over that but I did say "One of the more interesting pieces of this was how Ansible was used to keep the attacker in the system". The attacker was persisted via CM and their public repo, I'm actu
More ›