logs archiveIRC Archive / Freenode / #emacs / 2015 / August / 4 / 1
ryouma
it's an exciting distro
csed
That'd be awesome, yeah.
It is.
Gonna get a Libreboot X200 Pretty Soon TM. Plan on dropping Guix on it.
taylanub
ryouma: that'll take some time, but help always appreciated :D I see Arch as a good first goalpost.
hodapp
guix looks neat... maybe I'll set up a VM and forget about it.
ryouma
taylanub: not sure what you mean when you say nor found a chance...
what has a high barrier? guixsd?
taylanub
I mean, I'm not writing much Guile code, nor found a chance to contribute to Guile itself
miskatonic
forcer is a guile-guy
ryouma
ah, so guixsd might stil have a low barrier to entry, good
lancelot_of
csed, M-x package-refresh-contents takes several minutes, is it normal?
         

taylanub
yeah, contributing package recipes is relatively easy, though some packages can be tricky
(and some outright nightmareish)
csed
lancelot_of: Try doing an M-x package-install. M-x package-refresh-contents leaves a message in the minibuffer, so it looks like it's running forever.
lancelot_of
csed, is what I did, M-x package-install then solarized-theme and it does nothing saying [No match]
csed
lancelot_of: Huh, lemme see what you added under package-archives (the bit you got from the ELPA page on the wiki)
wgreenhouse
rudybot: seen aidalgol
rudybot
wgreenhouse: aidalgol was seen in #emacs twenty hours ago, saying ""You think you're funny, don't you?" "I think I'm adorable."", and then aidalgol was seen quitting fifteen hours ago, saying "Remote host closed the connection"
taylanub
since last week I've been tutoring a young 2-week intern at work, and from some of that experience, I can tell COBOL actually looks pretty neat.
wgreenhouse
taylanub: it has nil!
taylanub
it also made me re-re-re-realize how utterly bad Objective-C is. all the layers of dirt I need to explain to her, or handwave away...
benaiah
ugh, evil-mode ">>" and "<<" don't respect the current indentation...
lukas
that's just evil
benaiah
,rimshot
fsbot
*badum TISH*
taylanub
like I'm trying to teach the concept of a formatting string and its arguments, and the compiler starts whining because I need to explicitly cast an NSInteger to (long) then use %ld, otherwise it's ambiguous between 32b/64b. great fu*king distraction when you're trying to teach high-level concepts.
whenever you have (U)Integer variables in ObjectiveC and want to format a string, your code gets littered with (long)foo and (unsigned long)bar crap
dalias
sounds like NSInteger is your problem :)
tromey
printf is just an API disaster
csed
taylanub: Why not use something more high-level for teaching high-level concepts? I mean, you'd get bogged down in compiler warning if you just used C, don't want to imagine what Objective-C looks like.
tromey
there are so many issues like this
benaiah
isn't frequent casting normally considered a code smell?
dalias
yes
casts are a code smell
         

taylanub
csed: because it's iOS internship
tromey
like they added various char types to C but didn't add printf formats for them
dalias
tromey, ?
csed
taylanub; Ah, alright. That's a thing.
tromey
there is no way to printf a char16_t
dalias
you mean the char16/char32 stuff?
benaiah
taylanub: what is your opinion on Swift?
dalias
char16/char32 are pretty much useless
tromey
opinions vary
anxt
uint16_t ?
taylanub
benaiah: looks like a properly designed language opposed to Objective-C, but still carries abundant amounts of historic baggage from Apple for compatibility with existing C and Objective-C code...
miskatonic
Jonathan Swift of Lemuel Gulliver fame?
anxt
the whole unicode utf-8 thing baffles me
confounds
dalias
care to elaborate?
taylanub
if I want a non-garbage-collected language that still supports some modern constructs, I'd just run to Rust
dalias
i was asking anxt
anxt
oh, i just mean isn't that a char16?
benaiah
Rust seems exceptionally difficult to learn as a first low-level language
dalias
anxt, char16_t is defined as uint_least16_t which is uint16_t if the latter exists
but i think tromey's concern was that there's no way to print the character it represents directly via printf
but that makes sense, because a char16_t doesn't necessarily represent a whole character
anxt
yeah, printf annoys me
dalias
a format to print char16_t strings would make sense
but single char16_t's are not meaningful
they're code units, not codepoints
anxt
I've done a 6 or 8 months of C, and glib, and GLIB format macros in the middle of loggers, I wish there was a better way
dalias
(much like %c is useless for printing utf-8 characters, except in the special case where you know you have a member of the portable character set)
tromey
there's also no way to print a char16_t*
anxt
I should take a few hours and read about utf-8 and unicade
taylanub
is ASCII called "the portable character set" now? :P
dalias
tromey, *nod*
taylanub, they have slightly different meanings :)
tromey
and casts are mandatory in some situations, iirc posix has opaque typedefs that you have to widen by hand
it's just a mess
taylanub
anxt: UTF-8 is basically a mini protocol of metadata and payload
dalias
tromey, yes. casts are sometimes necessary, but they are code smells still
taylanub, that's one (imo bad) way to look at it
anxt
taylanub, yeah, I understood it to be a way to code unicode in variable length strings
tromey
does anyone here have a good understanding of regexps and NFAs?
I have a question
mrmyers
tromey: fire away?
taylanub
anxt: the metadata in each byte tells you whether that byte is the beginning of a multi-byte sequence (and if so, how long the sequence is), or whether it's a non-beginning member of such a sequence (so if a file is broken and cut in half, you can discard any leading bytes that belong to a sequence that was cut in half). the payload is just .. numbers. Unicode maps numbers to characters, so...
tromey
mrmyers: I was reading Russ Cox's pages about regexp implementation. and he discusses an extension to the Thompson approach to NFA simulation which can record capture group indices
anxt
taylanub, sounds pretty straightforward
tromey
mrmyers: and in this he says that the Thompson optimization still applies -- a given state can just appear once
mrmyers: but he doesn't explain why and I don't understand this; since it seems like you could enter the same state with different capture groups
mrmyers: he claims that this doesn't affect future parsing; which is true -- but it seems like it would affect the capture group result and so be important for correctness
I am really puzzled by this
taylanub
anxt: like many things, it's straightforward once you get it :P the Wikipedia article explains it pretty well IIRC.
anxt
there are just so many things I want to learn, and not enough hours
wilfredh
"This variable may be risky if used as a file-local variable."
anxt
that sounds like I could take 15 minutes and cherrypick a reasonable understanding
wilfredh
what sort of risk does this mean? Is file-local the same as buffer-local?
mrmyers
tromey: Not familiar with his extensions, this sounds pretty recent?
tromey
mrmyers: not really, he attributes the idea to Rob Pike's "sam" editor. I feel I am just missing some essential insight as to why this is ok
mrmyers: I can dig up the paper if you want
his writings on this topic are quite nice
miskatonic
sam is for mouse-lovers only
wilfredh
hmm, http://stackoverflow.com/questions/12180122/what-does-this-a-file-local-variable-mean suggests that file-local variable safety is just about automatically setting emacs variables when you visit a file
rudybot
http://teensy.info/Fd8o1cxe8n
taylanub
WTF, Unicode calls the ASCII underscore "LOW LINE"
tromey
PILE OF POO
taylanub
aww, don't have enough fonts to display that :P
tromey
me neither :(
« prev 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 next »