macdice

14 ms·

91.
by macdice 8y ago
>> illegal things in _PG_init() > just curious, can you point to some of these? i want to do my own extension audits and considering that plv8 does/did a lot of things based on postgres/src/pl/* it would be goo
92.
by macdice 8y ago
Hi Jerry, Yeah, parallelism is hard... but speaking as someone involved in several bug fixes in that area, I can tell you that we put a lot of effort into chasing those bugs down, and even made changes to allow extensions that were doing
93.
by macdice 8y ago
I think English speakers just find it a bit strange because our own language is such a mess. In fact we don't even agree on how to spell basic words (maybe due to a deliberate political act by Noah Webster? not sure), and there is zer
94.
by macdice 8y ago
The situation in English is a bit more complicated though. We have words of Norman French or Latin origin that contain gender information, like actor vs actress, and then we have the ending -er that is of Germanic origin, and we can use th
95.
by macdice 8y ago
I can report that the touchpad on a Dell XPS 13 under FreeBSD is absolutely horrendous (at least by default, and I haven't figure out how to tune it better yet). I hope all this talk about X and libinput really means "the open so
96.
by macdice 8y ago
Thank you!
97.
by macdice 8y ago
I really enjoyed Dark Pools. The asm + FoxPro source code for the Island ECN matcher described in the book has been released on the net and that amused me greatly when I was working on an order matching engine a few years ago. I love excha
98.
by macdice 8y ago
Anecdote: Yeah I was a bit surprised that a kernel/filesystem bug I logged back in November complete with a reproducer didn't even get acknowledged. I hope they'll eventually get to it, as it causes spurious build farm failu
99.
by macdice 8y ago
FreeBSD has had different semantics here for ~20 years: https://github.com/freebsd/freebsd/commit/e4e8fec98ae986357c... I wish I could find the discussion around that commit. It'll only start throwing aw
100.
by macdice 8y ago
I think it's debatable whether that's a bug in PostgreSQL, an underspecified interface, or a bug somewhere else. A more interesting question is how we get it fixed. The new Linux errseq_t design makes sure that every fd that was
101.
by macdice 8y ago
Well, I wouldn't. But people do. It makes more sense to use a SAN IMHO. I'm told it's not uncommon to use NFS for Oracle. One interesting thing is that they have their own NFS client implementation instead of trusting the
102.
by macdice 8y ago
I don't know anything about Windows, but that seems like the right sort of approach; also is there such a thing as a network block layer you could temporarily break by disconnecting it? On other OSes there are fault-injecting drivers
103.
by macdice 8y ago
FWIW here is a standalone test that shows Linux NFS exhibiting behaviour that would corrupt a PostgreSQL database: https://www.postgresql.org/message-id/CAEepm=1FGo=ACPKRmAxvb... You can also tweak that test so that EN
104.
by macdice 8y ago
I am also curious. I tried asking Microsoft via Twitter and they asked me to copy the information from that Wiki page into Microsoft One Drive so they could read it, and then suggested I try asking on Stack Overflow instead! https:/
105.
by macdice 8y ago
First, note that PostgreSQL uses SSI, an optimistic strategy for implementing SERIALIZABLE, so it neeeds to be able to nuke transactions that cannot be proven to be safe. This is a trade-off; other systems use (pessimistic) strict two-phas
106.
by macdice 8y ago
Edit: TL;DR: No, not with replication in the picture. As for whether replication shows fresh data, it sort of can, but not in a very practical way. synchronous_commit = on, when synchronous_standbys has been configured, does indeed make CO
107.
by macdice 8y ago
But the spec doesn't imply in any way that you'll need to write() again if fsync() fails. And dropping dirty flags seems way out of spec because now you can read data from cache that is not on disk and never will be even if futur
108.
by macdice 8y ago
I agree, the initial report on pgsql-hackers had a subject that blamed PostgreSQL too strongly and I think that carried through to articles and blogs and Twitter etc. You are right, this affects other things too. And indeed several project
109.
by macdice 8y ago
The original report from Craig Ringer was based on a real system that experienced a failure like this on thin provisioning. MySQL in buffered IO mode surely has the same problem, and has implemented the same solution: PANIC rather than retr
110.
by macdice 8y ago
I worked on PostgreSQL's investigation and response to this stuff and I agree with you FWIW. But apparently only FreeBSD (any filesystem) and ZFS (any OS) agree with us. Other systems I looked at throw away buffers and/or mark t
111.
by macdice 8y ago
For the record and not mentioned in Tomas's talk: the PostgreSQL release due out next week will add a PANIC after any fsync failure (in other words, no longer retry). The same thing has been done by MySQLand MongoDB and probably every
112.
by macdice 8y ago
I have a 9360 with coil whine. So far I like everything about it (size, weight, power) except that noise. Luckily, FreeBSD powerd_flag="-a adaptive" seems to get rid of most of it. I find that it is silent when idle with low CP
113.
by macdice 8y ago
Take a look at ARRAY LAYOUTS FOR COMPARISON-BASED SEARCHING. TL;DR there are schemes like this called "Eytzinger" and "van Emde Boas". https://arxiv.org/ftp/arxiv/papers/1509/1509.050
114.
by macdice 8y ago
PostgreSQL has the beginnings of a solution to this problem: version tracking for collations. I think it should be extended to track versions in a more fine grained way so that it can detect and reject this scenario (and related scenarios)
115.
by macdice 8y ago
The whole concept that "Linux is just a kernel" sounds like some kind of whining pedantic hair splitting, unless you understand that almost all the stuff that ships in a Linux distribution including all the FOSS stuff we call &quo
116.
by macdice 8y ago
I once wrote a program to send data for free anywhere in the world, by dialing another machine and then hanging up. The remote machine's modem would say RING. With the right excruciatingly adjusted and slow timing you could send bina
117.
by macdice 8y ago
An easier option than 2 for most Australians: come help build the software industry in New Zealand. It's a 3 hour flight and requires zero paperwork for Aussies to work here (as long as you have no criminal record; so you probably can
118.
by macdice 8y ago
Suppose there is an Australian who works on important and widely used open source infrastructural software, has commit rights, and is compelled to insert a back door by this new law. Well, obviously that would be really bad. But it seems
119.
by macdice 8y ago
One difficulty with French is that their system of stress makes it very hard for a beginner to know when one word ends and another begins. I lived in France for a few years, and I was initially surprised that after a few months of working
120.
by macdice 8y ago
Can you elaborate on the cons?

More ›