sabrina dreid:~> ps aux | grep “(Finder)”
dreid 245 0.0 -0.0 0 0 ?? E Mon11AM 0:00.00 (Finder)
Finder has been “Exiting” since monday. And no kill -9 doesn’t work.
#!/usr/bin/env pythonfrom ctypes import *import ctypes.utillibc = cdll.LoadLibrary(ctypes.util.find_library('libc'))def getCPUnums(): counts = {'logicalcpu': c_int(0), 'ncpu': c_int(0), 'physicalcpu': c_int(0)} for k, v in counts.iteritems(): size = c_int(sizeof(v)) libc.sysctlbyname('hw.%s' % (k,), c_voidp(addressof(v)), addressof(size), None, 0) return (int(counts['ncpu'].value), int(counts['physicalcpu'].value), int(counts['logicalcpu'].value))if __name__ == “__main__”: print “”"CPU Counts:N:\t\t%dPhysical:\t%dLogical:\t%d”"” % getCPUnums()
Of course my Intel Core 2 Duo gives me the following.
sabrina dreid:~> python cpun.pyCPU Counts:N: 2Physical: 2Logical: 2
You could have just as easily done sysctl hw.ncpu hw.physicalcpu hw.logicalcpu with commands.getoutput. But ctypes is cooler and almost certainly several times faster.
I’m considering starting a San Francisco Hacking Society chapter. I’ve been thinking about this for a few weeks ever since I was at Ritual Roasters with Donovan Preston, Steve Dekorte, and some other people and Steve expressed the desire for a shorter more frequent SuperHappyDevHouse in San Francisco. This is of course exactly what a Hacking Society would be. So for the past couple of weeks I’ve been considering trying to organize one, and most importantly I’ve been reflecting on why the ill-fated Davis Hacking Society failed.First of all it was a poorly publicized event. Really the only way people found out about it was via the Davis Wiki. We lacked the LUG affiliation which most Hacking Societies have, this was by design but clearly there weren’t enough programmers in the Davis Wiki audience.Second aside from the 3 regular attendees (Me, Zac, and Philip) people were showing up who either 1) Didn’t code and were therefor nearly ignored. 2) Didn’t have laptops 3) Were distracting.So what have I learned from the fall of DHS?I have to tell people about it. People need to know about it before they show up. I need to go to BayPiggies, and SuperHappyDevHouse, and even work and scream from the roof tops about this great thing. I need to have a dedicated wiki, and a mailing list, and all those things.I have to listen to other people. I organized DHS in more of a vacuum then I would have liked. It was just hard to get input from people other than “Hey, cool.” So the time and place selections were completely arbitrary. Partly this was because it wasn’t a pre-existing group of people I was trying to organize this with. I was making a new thing, and like most new things it never works right the first time. I intentionally avoided the LUG affiliation because I didn’t want people to think it was all Linux all the time, and we’d just be jabbering on about which distro was better all the time. I didn’t want to do that, I didn’t want it to be that. I wanted it to be open, free of prejudices against OS, Language, or Editor. Now of course, there are more organizations from which I can draw attendance, SuperHappyDevHouse is exactly the kind of thing I want, except on a smaller scale so that it can happen more frequently.A weeknight evening might not be the best timeslot. DHS was on thursdays from 7 until the cafe closed. I think people tend to have stuff to do at night. After work or just don’t want to do anything. I think weekend Saturday or Sunday afternoon might be better. The time should be predictable, though I do like the idea of adhoc sprinting it only works after you’ve built up an expansive network of participants, and are able to communicate with the participants.A cafe that was popular because it wasn’t popular might not have been the best choice. I alwasy liked Cafe Roma because of it’s flexibility. It didn’t have a bar, or any tables bolted to the floor, everything was very easy to reconfigure and no one minded when you did reconfigure. It was also virtually empty during our timeslot.I think that about covers it. Now it’s just about the details and not repeating my past mistakes.
After figuring out how to filter my mail I set about figuring out howto use imapfilter to archive old mail in my INBOX. This is actuallyquite simple.
oldmail = { 'old', 'not_since ' .. date_before(7)}results = match(myAccount, 'INBOX', oldmail)move(myAccount, 'INBOX', myAccount, 'INBOX-archive', results)
First off this code matches any old mail (read, seen, whatever) thatarrived more than 7 days ago. Then it moves this mail off to the sideto a folder called ‘INBOX-archive’. It’s put off to the side becauseof some deficiencies with mutt. But I don’t sync -archive foldersanyway so it’s not a big deal, it’s just there if I ever need it.Of course, this still lets the mail in my local folders for all mymailing lists (which constitute most of my mail) pile up. Thesolution? Well I’m subscribed to all the mailing list mailboxes, andit turns out that imapfilter gives me a really nice list of all mysubscribed mailboxes.
print("Archiving mail recieved before " .. date_before(7) .. ".")subscriptions = lsub(myAccount, '')for n in subscriptionsdo print('Archiving in ' .. subscriptions[n]) mailbox = subscriptions[n] results = match(myAccount, mailbox, oldmail) move(myAccount, mailbox, myAccount, mailbox .. ‘-archive’, results)end
Of course this doesn’t work. Atleast not out of the box on imapfilter1.2.2. At one point the following essentially occurs
prefix = NULL;/* read a bunch of stuff off of the network */strlen(prefix);EXC_BAD_ACCESS
Apparently imapfilter doesn’t exactly know how to do deal with myimap server, and prefix stays NULL, causing a not very nice Bus Error.The below patch fixes the problem, though not in the nicest waypossible. It merely does what the original author did and sprinklternary operators liberally.
--- old-namespace.c 2006-09-13 22:41:20.000000000 -0700+++ namespace.c 2006-09-13 22:42:15.000000000 -0700@@ -60,8 +60,8 @@ buffer_reset(&nbuf); o = 0;- if (!strncasecmp(mbox, prefix, strlen(prefix)))- o = strlen(prefix);+ if (!strncasecmp(mbox, prefix, strlen((prefix ? prefix : ""))))+ o = strlen((prefix ? prefix : "")); n = snprintf(nbuf.data, nbuf.size + 1, "%s", mbox + o); if (n > (int)nbuf.size) {
I’ll probably submit this patch to the upstream, or at the very leastnotify them of the bug. imapfilter is a pretty handy little app. Sohandy that it got me to actualy debug C code for the first time inyears.
When I switched to my minimalist mail environment (which takes noless than 4 command lines to read my mail) I suddenly found myselfwith a filtering problem. First some background on my environmentI read my mail in mutt. It doesn’t suck, very much. It alsosatisfies my absolute biggest requirement in a mail reader which isthe exact same behavior on Linux as on OS X. Being able to read mymail over SSH isn’t a requirement it’s just a perk. However I do_not_ use mutt’s built in IMAP support, to actually download mymail I use offlineimap to sync my remote mailboxes with a localMaildir format directory. So now I have all my email in~/Mail/dreid.org/INBOX but having all my mail in one place is notreally an optimal solution, and so this brings me to my filteringproblem. I’d used procmail in the past, but frankly at this pointafter so many years of staring at nice clean python code for avariety of purposes I’d rather eat my own excrement than writeanother .procmailrc. Luckily for me there is a nice little toolcalled imapfilter. Written in C but scriptable in Lua so installedit then immediately set about writing a filter for my mailinglists.
My Lua code handles the two most common List-Id headers that I’veseen, and works across all the mailing lists I’mcurrently subscribed to.First there is the domain only header:
List-Id: baypiggies.python.org
Then there is the Description
List-Id: Twisted Python Discussion
In my case I wanted to use the first segment of the domain as themailbox name. I use a very simple algorithm that makes thefollowing assumptions:
Given the 5 or 6 mailing lists I’m subscribed to and that they allfollow these rules, the obvious deficiencies in the code don’t bother meall that much, however your mileage may vary.
So without further ado here is the code so you too can be on yourway to mailing list filtering goodness
function parseListId(header) baseheader = string.sub(header, string.find(header, ':')+1, nil) destname = '' for i=1,string.len(baseheader) do c = string.sub(baseheader, i, i) if c == '<' then -- reset destname on any < brackets destname = '' elseif c == '.' then -- stop on the first . return destname elseif c ~= ' ' then -- add any non-space character to the destination name destname = destname .. c end endend
mailinglist = { 'header "list-id" ""'}results = match(myAccount, 'INBOX', mailinglist)listids = fetchfields(myAccount, 'INBOX', {'list-id'}, results) or {}mailboxes = {}for message, header in pairs(listids)do mailbox = parseListId(header) if not mailboxes[mailbox] then mailboxes[mailbox] = {} end table.insert(mailboxes[mailbox], message)endfor mailbox, messages in pairs(mailboxes)do move(myAccount, ‘INBOX’, myAccount, mailbox, messages)end
Posting this I can see some room for improvement, I could cache andcompare the actual unparsed header (it shouldn’t vary from message tomessage) which would reduce the number of calls to parseListId (I’mnot sure it’s slow enough to worry about though.So now you have what I think is a pretty decent solution, eventuallyimapfilter will get run by a cronjob on my server (cron seems to notactually work on OS X tiger, which is another blog post all together.)Then I’ll only have to run 3 commands to check my mail. Next timeI’ll teach you how to create archives for all subscribed mailboxes.Which is a ridiculously easy bit of lua though I had to patchimapfilter to not get an EXC_BAD_ACCESS when reading folder lists frommy server.
It struck me the other day that I’m payed up through the year of hosting on dreid.org (as well as various other domains.) And I really don’t _use_ it for anything. So here it is. My new “blog” I’ve had many blogs over the years, I’ve written a lot of blog software. Mostly I’ve stuck with livejournal because it was easy, now however I’ve grown tired with it, for multiple reasons the most important of which is that I have no control over what is syndicated to planet twisted. So hopefully wordpress solves this problem. (Why wordpress? Because it only took me 5 minutes to get to this point.)