logs archiveIRC Archive / Freenode / #emacs / 2015 / July / 2 / 1
wasamasa
and someone on github has a much better package they've named wc-mode, too
so I've had no chance of getting that into MELPA
it seems that this issue was fixed though: https://github.com/bnbeckwith/wc-goal-mode
ijp
I used to be annoyed by it, but I've came around to the way racket's planet did it: package names were basically (account, name) tuples
wasamasa
that's how I'd list these, too
miskatonic
fsbot will be reprogrammed into a new bouncer at melpa
ijp
ywot
tali713
ijp: verbose, but workable for package names, won't fix conflicting libraries. but generally you only need one python-mode or one wc-mode, etc.
wasamasa
yeah
so I've settled for the approach of picking the one I want
ijp
right, but you remember the mess a short while ago where there were N conflicting python-mode's all named python-mode
wasamasa
there's just python.el and python-mode.el
         

tali713
wasamasa: would it be possible to alter melpa so that it could prefix with github user name in case of conflict?
wasamasa
tali713: oh, I'm sure they could do loads of things
ijp
wasamasa: now that the current python-mode is in emacs, the fact that it is not the old python-mode that came with emacs is moot
miskatonic
there seem to be infinitely many python programmers using emacs
tali713
miskatonic: I doubt the veracity of that claim.
ijp
countable but not countably infinite
tali713
also,,cthulhu
fsbot
It was the eldritch scurrying of those fiend-born rats, always questing for new horrors, and determined to lead me on even unto those grinning caverns of earth's centre where Nyarlathotep, the mad faceless god, howls blindly to the piping of two amorphous idiot flute-players.
miskatonic
of course it does not stop with modes that highlight and auto-index python code, but there are thpose more or less integrated python development environments, such as all those rope* things
tali713
miskatonic: I believe we were referring to modes /named/ `python-mode'. not arbitrary modes used with python.
miskatonic
this doesn't make it easier for those trying not to get lost on palnet Melpa when looking for topols for python development.
tali713
miskatonic: true, but that's a separate problem.
miskatonic
lovecraft would have programmed in python
ijp
nah, php
tali713
that problem, I'm afraid, is not solvable. there will always be one more library written for one persons personal convenience. it renders scanning the list of all packages to see if something appeals a pitfall filled approach.
ijp
but a custom build that segfaults 1/5 of the time
tali713
,ijp++
fsbot
Noted, tali713. One rms-point for ijp!
tali713
ijp: is that an improvement or a downgrade?
ijp
yes
         

rrix
rms-points
dang
tasslehoff
I have a formula under my org-table that sums a column. Now I want to multiply each with an 'amount' from the neighbouring cell. Help.
miskatonic
I never got the hang out of the org mode
tasslehoff
me neither, but the effort I've put into it is not great. I use it for simple tables, logging research and hour registration.
benaiah
lovecraft would definitely have gone with malbolge
ijp
,sh*t-#emacs-says
fsbot
I'll take "it's possible that it will still break" over "it's not possible that it will not break" any day
wasamasa
speaking of horrible languages
https://github.com/Fuco1/be-irc-bot
kovrik
wasamasa: that's crazy
wasamasa
kovrik: inorite
kovrik: how's the coffee?
kovrik
wasamasa: we have disgusting coffee at work, but that's better than no coffee at all
wasamasa
(Action) is way too curious over things he doesn't even deal with
I wonder what day I'll cave in and go for coffee over tea
kovrik
wasamasa: sometimes I wonder if the place I'm working at is the only place in NZ where bad coffee exists. Because everywhere else coffee is good here
wasamasa
(Action) wonders what business mousepads have in new zealand
kovrik
wasamasa: :)
wasamasa: anyway...wanted to ask if there is such a thing in emacs as 'on-buffer-created' callback? I guess the answer is NO (can't google anything)
wasamasa
kovrik: first of all, emacs has hooks
kovrik: second, you don't really want to hook that
the thing is that for whatever reason string manipulation is a second class citizen in elisp and you're encouraged to work in buffers instead
kovrik
wasamasa: but AFAIK there is no 'on-buffer-created' hook in there. Why it is bad to use hooks for that?
don't get it
wasamasa
this has led to a rather incredible amount of buffers being created and killed for temporary tasks
so, if you hook something like `buffer-list-update-hook' with something of O(n) complexity, that's a great way to slow down other packages using loads of temporary buffers
jlf
wasamasa: "kill 'em all, let the GC sort 'em out"
ijp
where N is what?
wasamasa
ijp: n as in number of buffers
ijp
and where is this parameter being used?
wasamasa
I've learned this the hard way while wondering why the newest feature in firestarter made ZNC replay in circe unbearably slow
kovrik
wasamasa: Got it. The thing is that I read your comments on github...and I find proposed syntax to be bloated. Current shackle rules a very simple. Don't want to break that simplicity
ijp
I swear, big O has probably caused as much confusion as simplification
wasamasa
it was because I felt like catching buffer renames by adding a function going through all buffers and optionally setting a buffer-local variable to this hook
once I let-bound the hook to be empty, performance was back to old levels
so there you go, an useless hook in emacs!
ijp
a*
wasamasa
I've ended advising the `rename-buffer' function instead
and eventually rewrote it a bit more to not need to track the renames at all
ijp: so, if my function loops through all buffers and I refer to their number as n, wouldn't that be a complexity of O(n)?
ijp
it might be, it might not
wasamasa
anyhow, I've made the experience that it's a bad idea adding anything taking up time to that hook
jlf
ijp: for example?
ijp
jlf: the underlying assumption is that the operation on a given buffer is O(1)
jlf: or do you mean my more general comment?
jlf
well, of course, or else he would have mentioned another independent variable
since he didn't we can safely assume he was talking about walking a list of length n
wasamasa
kovrik: I don't really have a problem with it and find the condition variant slightly more elegant
ijp
jlf: why independent?
wasamasa
kovrik: the action variant would work better to demonstrate a new possible keyword
kovrik: both could be combined
jlf
ijp: i assume you're trolling
ijp
jlf: consider checking for uniqueness in a list
jlf
?
ijp
you are looping over a list, but the cost at each stage is dependent on the size of the list
jlf
there's nothing to suggest that
wasamasa
kovrik: I would prefer to not reinvent cond in a bad way and instead make it simple to hook into shackle in a more elegant manner than by offering before/after hooks
kovrik
agree
ijp
jlf: all the statement says is "my function loops over a list and the size is n"
you can't go from a lower bound to an upper bound
wasamasa
ijp: if you're interested in the actual code, here you go: https://github.com/wasamasa/firestarter/commit/290a665fda83e9137643c67a1776c94cb39f781f
rudybot
http://teensy.info/BmevhkVMQx
jlf
ijp: no, it was "going through all buffers _and optionally setting a buffer-local variable to this hook_" [my emphasis]. the natural assumption is (mapc (lambda (buf) (when (some-predicate buf) ...)) (buffer-list))
ijp
as usual, we're talking about two completely different statements
jlf
(Action) shrugs
wasamasa
kovrik: I've written code for the action variant already and will write code for the condition variant tomorrow
kovrik: shall I attach a patch to make erc not set the server buffer current?
« prev 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 next »