Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
tmm1
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
10 ms
·
31.
▲
by
tmm1
16y ago
This is a great article. I came across it while hacking on posix-spawn and it was quite useful.
32.
▲
by
tmm1
16y ago
The reason fork is so slow on Linux is because the default page size is 4k. This means for a 300MB process, 76800 page table entries have to be copied during the fork. There are ways to enable HugeTBL on linux to increase the page size, whi
33.
▲
by
tmm1
16y ago
Here's the full announcement: https://gist.github.com/e0d36c881e6efc57514b
34.
▲
posix-spawn: a faster fork+exec for ruby
(github.com)
56 points
by
tmm1
16y ago
|
12 comments
35.
▲
by
tmm1
16y ago
A FIFO might work. Do you know how big the buffers associated with FIFOs on OSX are?
36.
▲
by
tmm1
16y ago
rbtrace is built on ruby's event hook api, which provides events when ruby and C extension methods are called, and when they return. python offers a very similar event hook api: http://www.doughellmann.com/PyMOTW/sys/tracing.html
37.
▲
by
tmm1
16y ago
If you tried it, you would understand. Ruby programs make a _LOT_ of method calls. Unfortunately the firehose doesn't work too well on OSX, because the unix message queue API I'm using (msgget(2)) is capped in the kernel to 2048 bytes and 4
38.
▲
rbtrace: like strace, but for ruby code
(github.com)
101 points
by
tmm1
16y ago
|
11 comments
39.
▲
by
tmm1
16y ago
Yes it is.
40.
▲
Memprof: a memory profiler for ruby [scribd]
(scribd.com)
33 points
by
tmm1
16y ago
|
4 comments
41.
▲
Debugging Ruby with MongoDB
(scribd.com)
63 points
by
tmm1
16y ago
|
3 comments
42.
▲
EventMachine: scalable non-blocking i/o in ruby
(timetobleed.com)
82 points
by
tmm1
17y ago
|
37 comments
43.
▲
by
tmm1
17y ago
http://dl.dropbox.com/u/635/gc-export.pdf
44.
▲
by
tmm1
17y ago
GC tuning is easy with REE, you can set environment variables to change the behavior of the mark and sweep collector. See slides 50-53 for more info.
45.
▲
by
tmm1
17y ago
"Now you know, and knowing is half the battle."
46.
▲
by
tmm1
17y ago
You're right, there's about 90k NEWLINE nodes on the heap, and at 40 bytes a piece that's taking up about 3.6mb of slots on the heap that could be used by other ruby objects instead. Unfortunately, removing newlines from your codebase will
47.
▲
by
tmm1
17y ago
This a new presentation specifically about the MRI GC, but it covers some of the same tools mentioned in our Debugging Ruby and Threaded Awesome talks
48.
▲
Garbage Collection and the Ruby Heap
(timetobleed.com)
99 points
by
tmm1
17y ago
|
27 comments
49.
▲
Touchscreen analysis shows iPhone accuracy lead over Droid
(appleinsider.com)
2 points
by
tmm1
17y ago
|
1 comments
50.
▲
by
tmm1
17y ago
That commit is for YARV, not MRI
51.
▲
by
tmm1
17y ago
>> if false >> a = 1 >> end >> defined? a => "local-variable" >> a => nil
52.
▲
What is a ruby object?
(timetobleed.com)
41 points
by
tmm1
17y ago
|
0 comments
53.
▲
Pollution in China, in pictures
(chinahush.com)
5 points
by
tmm1
17y ago
|
0 comments
54.
▲
by
tmm1
17y ago
Same here, can't RSVP
55.
▲
by
tmm1
17y ago
Plus, you can trace a lot more than mysql queries.. commonly used libraries like libMagick, libxml, libcurl, libmemcached, and even internal VM functions like garbage_collect(), PyEval_SafeThread(), etc.
56.
▲
by
tmm1
17y ago
Using strace to trace mysql traffic is painful. You have to know which file descriptors to trace (check via lsof first), then trace all network traffic on those fds in verbose mode to see the queries themselves (doesn't work if the mysql pr
57.
▲
by
tmm1
17y ago
Reverse debugging, awesome! http://www.gnu.org/software/gdb/news/reversible.html
58.
▲
Threaded Awesome: Understanding Ruby 1.8 Threads
(timetobleed.com)
1 points
by
tmm1
17y ago
|
0 comments
59.
▲
by
tmm1
18y ago
It's cheaper than the certs at Verisign. Where do you buy your SSL certificates?
60.
▲
by
tmm1
18y ago
http://goog-perftools.sourceforge.net/doc/tcmalloc-opspersec...
More ›