After a chat on #gentoo-haskell, it turns out that compiling on ramfs and Hardened do not go together.
ramfs does not support XATTRs (tmpfs does), which paxctl-ng uses to store PaX markings, so you end up with a GHC that has no -m (no MPROTECT, i.e. no W^X) marking set. And one of the very first things GHC does is a RWX mmap()...
1. Minor GHC breakage due to grsecurity
2. Recompile everything Haskell
3. ???
4. Complete GHC breakage due to RWX mmap() vs grsecurity
Not what I wanted.
@randomgeek Yeah, that's a good idea. I had a class on it in school at some point, you can recreate the contents by forcing yourself to type by memory/touch and doing typing exercises, which usually start off with 2-char combinations and then extend to longer words. Shifting keys around isn't that much of a problem when you're forced to use your memory, just a bit annoying when you have to use other computers ;)
@randomgeek Good idea :)
I also learned it somewhat late (age 17 or so?), it's still worth it, especially for coding and heated arguments on IRC and such.
If you want to force yourself to not look at the keyboard, don't forget that there are bumps on most keyboards where your index fingers go. As soon as you know the 'base position', your memory can do the rest. (can also cover the keyboard with some elevated piece of wood that prevents you from seeing the keys)
@CobaltVelvet @Siphonay will light up to review papers because the introduction is already too boring on its own
@jalefkowit America's Got Talent is your #Eurovision, though :P
This is a good repo (if you want to pull a little CSS prank on someone): https://github.com/queueRAM/aprilFools.css
@Elizafox bash.org recently started moderating again, qdb.us sometimes do (but the queue is where the real fun is at if you know all the rest).
@Elizafox Read http://bash.org/?latest or http://qdb.us/queue instead, at least it's selected bullshit instead of random bullshit.
[my impression of the state of AI in pop culture, snarky, sorry]
RESEARCHER: I have made yet another somewhat interesting way to transform one spreadsheet into another spreadsheet
JOURNALISTS: we're all doomed by robots lol
[fin]
CDs scare birds because they have good enough vision to see all the pits and troughs and theyre smart enough to decode a ISO 9660 filesystem on-the-fly (literally)
"/tmp: write failed, file system is full", buried under 1500 lines of other stderr. Slightly suboptimal.
Bah, OpenBSD's tar(1) is insane. Incredibly short filename limit (not even 128 chars, as it seems), simply breaks with "tar: End of archive volume 1 reached" for some as-of-yet unknown reason without any error message...not the kind of reliability I want :(
@argumatronic @chris_martin I feel that `do` is useful at first because it (usually) provides a neat structure to code and enables you to not worry about what actually happens; a PITA a few weeks/months later because it obscures stuff; and a mixed blessing in the end that should still be used in some situations, but not all.
I'd been wanting to turn some of my conversations with @chris_martin into blog posts, because over the course of conversation we've covered some topics pretty thoroughly, this is how we're writing The Joy of Haskell, and also it allows for presentation of two perspectives on a subject naturally. First one is about do-notation:
https://joyofhaskell.com/posts/2017-05-07-do-notation.html
@selenized Yeah, that's true. But sadly, I already accepted that as given - I see it in my own behaviour, too much social media and the like has instilled a mild version of ADD in me where I get bored of books way too quickly and start distracting myself. The lack of instant gratification (compared to the hyperactivity of current mass media) doesn't make it easy.
So "hijacking" (for lack of a better word) online media to get people to read books is still valuable, imho.
Oh. That's probably worth remembering. /\d/ in Perl finds Unicode numerals, which gets you much more than /[0-9]/ - unless you use the /a modifier for restricting to ASCII text.
http://blogs.perl.org/users/ben_bullock/2017/05/d-does-not-validate-numbers.html
So, things on my mind right now:
1) the computing platform I really want (and I do not have),
2) how to grab all that unhappy scientific talent in the US and bring it to Europe,
3) introducing liability in software "the right way",
4) a new processor design,
5) a resilient IoT basis design,
6) desperate desire to play waterpolo again,
7) several books to read and digest,
8) have tea with interesting people (not a conference).
Space for more...
@selenized If it takes the prospect of financial success to motivate people to read more, so be it. If someone wants to prove the earth is flat or vaccines cause autism and it gets them to read scientific articles instead of watching YouTube or commenting on Facebook, that's a win in my book.
@Aerdan It's just realism, the majority of scientists smoke weed.
At least in CS, you won't find many people who'd decline a spliff when there's no paper to be written.
