darkf has quit [Read error: Connection reset by peer]
darkf has joined #ocaml
darkf has quit [Read error: Connection reset by peer]
darkf has joined #ocaml
ohama has quit [Ping timeout: 252 seconds]
ohama has joined #ocaml
yacks has joined #ocaml
yacks has quit [Quit: Leaving]
yacks has joined #ocaml
pango has quit [Ping timeout: 276 seconds]
ttamttam has joined #ocaml
ttamttam has left #ocaml []
pango has joined #ocaml
awm22 has quit [Quit: Leaving.]
walter has quit [Quit: This computer has gone to sleep]
yacks has quit [Ping timeout: 256 seconds]
Yoric has joined #ocaml
Tobu has joined #ocaml
yacks has joined #ocaml
anderse has joined #ocaml
fraggle_ has quit [Ping timeout: 264 seconds]
fraggle_ has joined #ocaml
gnuvince has quit [Ping timeout: 272 seconds]
<pippijn>
morning
zpe has joined #ocaml
chambart has joined #ocaml
Trollkastel has joined #ocaml
anderse has quit [Quit: anderse]
zpe has quit [Ping timeout: 256 seconds]
ttamttam has joined #ocaml
eikke has joined #ocaml
<adrien_oww>
o/
hkBst has joined #ocaml
hkBst has quit [Changing host]
hkBst has joined #ocaml
<orbitz>
morrn
zpe has joined #ocaml
awm22 has joined #ocaml
zpe has quit [Ping timeout: 272 seconds]
thomasga has joined #ocaml
<wmeyer>
morning
<Leonidas>
morning wmeyer
cago has joined #ocaml
mika1 has joined #ocaml
ontologiae has joined #ocaml
djcoin has joined #ocaml
<wmeyer>
Leonidas: hi :)
<wmeyer>
sorry going to work at the moment
Snark has joined #ocaml
eikke has quit [Ping timeout: 252 seconds]
ontologiae has quit [Ping timeout: 256 seconds]
mfp has quit [Ping timeout: 272 seconds]
jbrown has joined #ocaml
mfp has joined #ocaml
ttamttam1 has joined #ocaml
ttamttam has quit [Read error: Connection reset by peer]
ttamttam1 has quit [Read error: Connection reset by peer]
ttamttam has joined #ocaml
ontologiae has joined #ocaml
hkBst has quit [Read error: Connection reset by peer]
hkBst has joined #ocaml
hkBst has quit [Changing host]
hkBst has joined #ocaml
Kakadu has joined #ocaml
eikke has joined #ocaml
talzeus_ has quit [Remote host closed the connection]
talzeus_ has joined #ocaml
eikke has quit [Ping timeout: 255 seconds]
hkBst has quit [Read error: Connection reset by peer]
hkBst_ has joined #ocaml
hkBst_ has quit [Changing host]
hkBst_ has joined #ocaml
talzeus_ has quit [Ping timeout: 264 seconds]
talzeus_ has joined #ocaml
mcclurmc has joined #ocaml
mika1 has quit [Quit: Leaving.]
hkBst_ has quit [Remote host closed the connection]
hkBst_ has joined #ocaml
hkBst_ has quit [Changing host]
hkBst_ has joined #ocaml
<gasche>
orbitz, pippijn: it's not really fair to say that the "industry effort for thread-safe runtime" is "zero", as you answered to SuperNoeMan
<gasche>
both OcamlPro and OcamlLabs are working on that
<gasche>
neither of them are aiming at a *concurrent* runtime, granted, but their (rather similar but not identical) multi-runtime design qualifies as thread-safe, that is explicit support for concurrency in the runtime
<gasche>
now, as whether they have a higher chance of success than the ultimately doomed Ocaml4mc project, we can't tell yet
mcclurmc has quit [Read error: Operation timed out]
<gasche>
but I personally have good feelings about this
cago has quit [Ping timeout: 260 seconds]
<gasche>
(whether that is exactly what people mean when they ask about concurrent runtimes is another question; but I'm not sure what people mean technically is actually relevant in this precise case)
<gasche>
it's funny to see that discussing Big Interpreter Locks is such a popular topic around here
<gasche>
it's like the third discussion I see in three days
Kakadu has quit [Ping timeout: 245 seconds]
<gasche>
maybe that would be a good topic for a blog post :]
<pippijn>
definitely
<gasche>
I've been avoiding controversial subjects so far
<gasche>
but that could be a good strategy to stir up readership :]
Tobu has quit [Remote host closed the connection]
<gasche>
(1) Concurrency in OCaml (2) Lazyness sucks (3) Type-level hackery is overrated
hkBst_ has quit [Read error: Connection reset by peer]
<orbitz>
gasche: Ok my 'zero' may have been too strong, I did make it clear after though taht AFAIK the work being done on paralleism in ocaml is closer to Erlang than Java
hkBst_ has joined #ocaml
hkBst_ has quit [Changing host]
hkBst_ has joined #ocaml
<companion_cube>
gasche: would you have insights on the choice lazy expr /fun () -> expr ?
Kakadu has joined #ocaml
<gasche>
My gut feeling is to favor (fun () -> ...) for simplicity as long as memoization is not a hard requirement
<pippijn>
companion_cube: the former only evaluates expr once
<companion_cube>
pippijn: yes, but I heard there is an overhead when you only evaluate once
<pippijn>
yes
<companion_cube>
(I used to use lazy for debug messages, where you evaluate at most once)
<pippijn>
you need to remember whether you evaluated it, yet
<gasche>
syntactically lazy is marginally nicer
<gasche>
(eg. you've got lazy patterns)
<gasche>
but in practice people don't know about lazyness in OCaml setting
<orbitz>
i have never really used it, i don't understand it enough
<gasche>
so e.g. they'll be surprised by your API if it exposes lazy usage, and this will require them a bit more time to process it
<gasche>
orbitz: it's fairly simple specification-wise: it's just like (unit -> 'a), with a side-effect that caches the value after the first computation
<companion_cube>
gasche: oh, lazy patterns
<companion_cube>
didn't know about that
<companion_cube>
may be very useful to implement Okasaki's stuff
<gasche>
indeed
<gasche>
they're a 3.12 addition iirc.
<companion_cube>
I think many people also don't know about pattern matching in records
<rks_>
companion_cube: why would you say that?
<pippijn>
as in | { value = 3 }?
<companion_cube>
yes
<darkf>
or record updates <3
<companion_cube>
because some people in my lab didn't know it :)
malo has joined #ocaml
<companion_cube>
I think pattern matching examples are very often only on algebraic types
<pippijn>
that's strange
<pippijn>
because that's about the first thing I learnt when I started with ocaml
<companion_cube>
wow
<gasche>
there are actually fascinating theoretical underpinnings to the question of destructuring product types by projection or pattern-matching
<darkf>
wat
hkBst_ has quit [Read error: Connection reset by peer]
<gasche>
(in polarized/focused type theories, one style is a characteristic of positive sums, the other of negative sums)
<gasche>
s/sums/products/, sorry
<darkf>
gasche: I understand what these words mean but not in that sequence
<darkf>
:-)
<gasche>
positive products (would rather use pattern-matching), negative products (would rather use field projections)
<gasche>
but in practice, well, it's just nice to have both
hkBst_ has joined #ocaml
hkBst_ has quit [Changing host]
hkBst_ has joined #ocaml
RagingDave has joined #ocaml
Yoric has quit [Ping timeout: 246 seconds]
<gasche>
orbitz: my understanding is that async/futures, fork-join and message-passing are slowly becoming the ways most people think about concurrency, with bare access to shared-memory becoming a specialist business
<gasche>
the same way we don't need to apologize about having a GC anymore
<gasche>
(while the OCaml or Java communities did spend an important effort justifying that choice to "the crowd" not that long ago)
<gasche>
maybe we should rather build up an argumentation about why OCaml cannot "just ship code to the GPU"
<gasche>
(as all fancy languages currently can, right?)
cago has joined #ocaml
<companion_cube>
it looks like no-copy message passing is also trendy
<companion_cube>
within a same process, that is
<gasche>
yes, but that's more of an implementation question
<gasche>
rks_: this Merlin bug you just closed, do you think it would have been appropriate of me to close it directly?
<gasche>
I considered doing that
hkBst_ has quit [Remote host closed the connection]
hkBst_ has joined #ocaml
<rks_>
gasche: yes you could have :)
eikke has joined #ocaml
<rks_>
I was troubled by your reponse at first, because I had a parse error in my toplevel, but then I saw your remark on camlp4, which explained it :)
cago has quit [Read error: Connection reset by peer]
cago1 has joined #ocaml
cago1 has quit [Quit: Leaving.]
cago has joined #ocaml
cago has quit [Client Quit]
cago1 has joined #ocaml
hkBst_ has quit [Ping timeout: 272 seconds]
hkBst__ has joined #ocaml
<gasche>
have you done a pull request on ocaml.org yet?
<gasche>
everyone should consider contributing something, so that we get a great website :]
talzeus_ has quit [Remote host closed the connection]
* avsm
seconds gasche
hkBst__ has quit [Ping timeout: 256 seconds]
* adrien_oww
did :D
<adrien_oww>
can I get a cookie?
talzeus_ has joined #ocaml
<rks_>
gasche: about merlin?
<rks_>
we've talked about it yes, no one took the time to do it yet though
<gasche>
that was a chan-wide suggestion, not to you specifically
<gasche>
but yes, there may be a tooling session where Merlin would definitely have its place
<gasche>
like a page about "The development environment(s) of the up-to-date OCaml programmer"
<gasche>
if it isn't there, create it!
<gasche>
(cf. the not-so-old Geany discussion on the mailing-list)
<rks_>
so I haven't had the courage to modify it yet :'
frederict has joined #ocaml
frederict has left #ocaml []
UncleVasya has joined #ocaml
hkBst has joined #ocaml
hkBst has quit [Ping timeout: 272 seconds]
cago1 has quit [Quit: Leaving.]
hkBst has joined #ocaml
talzeus_ has quit [Remote host closed the connection]
mcclurmc has joined #ocaml
talzeus_ has joined #ocaml
hkBst has quit [Ping timeout: 256 seconds]
hkBst_ has joined #ocaml
hkBst_ has quit [Changing host]
hkBst_ has joined #ocaml
ontologiae has quit [Ping timeout: 255 seconds]
hkBst_ has quit [Ping timeout: 248 seconds]
hkBst_ has joined #ocaml
Trollkastel has quit [Read error: Connection reset by peer]
hkBst_ has quit [Changing host]
hkBst_ has joined #ocaml
hkBst_ has quit [Ping timeout: 246 seconds]
mcclurmc has quit [Ping timeout: 252 seconds]
mcclurmc has joined #ocaml
mcclurmc has quit [Ping timeout: 255 seconds]
chambart has quit [Ping timeout: 246 seconds]
_andre has joined #ocaml
weie has quit [Quit: Leaving...]
talzeus_ has quit [Quit: where is the love...]
talzeus_ has joined #ocaml
hkBst_ has joined #ocaml
ontologiae has joined #ocaml
hkBst_ has quit [Ping timeout: 272 seconds]
walter has joined #ocaml
hkBst_ has joined #ocaml
hkBst_ has quit [Changing host]
hkBst_ has joined #ocaml
zpe has joined #ocaml
zpe has quit [Remote host closed the connection]
cago has joined #ocaml
hkBst__ has joined #ocaml
hkBst_ has quit [Ping timeout: 264 seconds]
thomasga has quit [Ping timeout: 256 seconds]
hkBst__ has quit [Ping timeout: 255 seconds]
hkBst__ has joined #ocaml
hkBst__ has quit [Excess Flood]
hkBst__ has joined #ocaml
talzeus_ has quit [Remote host closed the connection]
talzeus_ has joined #ocaml
chambart has joined #ocaml
talzeus_ has quit [Remote host closed the connection]
Neros has joined #ocaml
zpe has joined #ocaml
hkBst__ has quit [Remote host closed the connection]
hkBst__ has joined #ocaml
Nahra has quit [Ping timeout: 264 seconds]
chambart has quit [Ping timeout: 264 seconds]
ggole has joined #ocaml
zpe has quit [Ping timeout: 252 seconds]
chambart has joined #ocaml
chambart has quit [Ping timeout: 272 seconds]
ollehar has joined #ocaml
hkBst__ has quit [Ping timeout: 264 seconds]
hkBst has joined #ocaml
Nahra has joined #ocaml
Nahra has quit [Remote host closed the connection]
cdidd has quit [Read error: Connection reset by peer]
eikke has quit [Ping timeout: 255 seconds]
beckerb has joined #ocaml
zpe has joined #ocaml
groovy2shoes has joined #ocaml
jamii has joined #ocaml
eikke has joined #ocaml
weie has joined #ocaml
zpe has quit [Ping timeout: 272 seconds]
groovy2shoes has quit [Quit: groovy2shoes]
Neros has quit [Ping timeout: 272 seconds]
groovy2shoes has joined #ocaml
Neros has joined #ocaml
yacks has quit [Read error: Operation timed out]
hkBst has quit [Remote host closed the connection]
hkBst has joined #ocaml
awm22 has quit [Ping timeout: 248 seconds]
cago has quit [Quit: Leaving.]
q66 has joined #ocaml
smondet has joined #ocaml
awm22 has joined #ocaml
groovy2shoes has quit [Quit: groovy2shoes]
Tobu has joined #ocaml
zpe has joined #ocaml
<fasta>
Is there some nice library which already does replacing the first occurence of a regular expression in a file by a replacement pattern?
<orbitz>
define 'nice'
<orbitz>
I wrote my own wrapper around ocaml-re to do very trivial replacements, but there is Str and pcre for more complicated things. I'm guessing those don't fit your criteria for nice
cago has joined #ocaml
mcclurmc has joined #ocaml
<fasta>
orbitz: nice would mean "already does it", like sed does.
<fasta>
I am not sure, but doesn't sed -i use nearly constant memory usage?
<fasta>
-usage
<gasche>
ah, in *a file*
<gasche>
that would be more of an IO question
<fasta>
If one first needs to build a big string in memory, then it would be wasting memory.
<gasche>
(but thank you for suggesting the reddit/ocaml link of the day!)
<fasta>
Sed is not a Stream EDitor for nothing.
<gasche>
I'm afraid ocaml-re at least (and probably the other quoted above) work on strings
mk270 has left #ocaml []
<gasche>
I'm not sure what would be the right abstraction layer to operate on files
<fasta>
gasche: a file already is an abstraction. Why need more?
<fasta>
You can have a ByteLikeThingie, but that would just be equivalent to a file.
<gasche>
I'm wondering about how to do that in the OCaml world
<gasche>
you could operate on Bigarray, and then have you mmap the file
<gasche>
but maybe you care more about arbitrary streams than dead files (trading performance for flexibility)
<gasche>
in fact
<gasche>
ocaml-re interface would be quite easy to adapt to incremental matching
<gasche>
it looks like it's not completely exposed right now
<gasche>
but you should be able to read your big file (or stream or whatever) by batches of a fixed buffer size, run the regexp, and get information about whether you found your stuff (and can just cat the rest), found nothing, or found a partial match and need to feed more input to the matcher
<fasta>
gasche: and how would the writing part go then?
<fasta>
gasche: unless you would use redirection, of course.
<fasta>
gasche: I am talking about a pure OCaml solution.
<ggole>
What if the matched part is larger than the buffer, rather than just straddling a boundary?
<ggole>
You throw away the first bit before finding whether it is a match or not.
<gasche>
well you probably have to use as much memory as the size of the longest partial match in the stream
<gasche>
hm
<gasche>
no
<gasche>
that's the case for a streaming interface
<ggole>
If the file supports seeking you could just rewind
<thelema_>
gasche: not at all; DFA can do incremental matching, and then reverse match to find start of match, then rewind
<gasche>
but if you know you have a file, yes
<gasche>
all three of us agree :]
<thelema_>
fasta: to answer your question, no; there's no ocaml library that does what you want in a streaming fashion
<fasta>
Then we go for Perl.
<ggole>
You really do need to handle the non-seek case.
<ggole>
Pipes are pretty common.
<fasta>
I kind of hate it, but well, if they have solved it perfectly, who cares.
<gasche>
well
<ggole>
I guess you can avoid the large allocation though, in other cases.
<thelema_>
ggole: the easiest practical way to handle the non-seek case is by using a backtracking algorithm to do matching
<gasche>
fasta: you should send an email to Jerome Vouillon, to see if he would be willing to implement that in ocaml-re
<thelema_>
ggole: the downside is that you have exponential worst case time.
<gasche>
I'm not saying you'd have to use it then, but that may help the next person having the same need
<thelema_>
gasche: it would be difficult to implement in ocaml-re because iirc, it uses proper automata theory
<gasche>
also, could you be precise about whether you need streaming or just a file-ish thing with a seeking capability?
<gasche>
thelema_: how does that make it more difficult?
<thelema_>
gasche: because DFA can't save location of beginning of match
<gasche>
hm, I need to think about that
<gasche>
but have to leave now
<thelema_>
gasche: cheers
<gasche>
thanks
<fasta>
gasche: I use both, but now only a file-ish thing.
barbapapa has joined #ocaml
<thizanne>
aaa
<thizanne>
sorry
tianon has quit [Quit: "I'm a very important man. I've got a tower!"]
eikke has quit [Ping timeout: 256 seconds]
Kakadu has quit [Quit: Page closed]
thomasga has joined #ocaml
awm22 has quit [Quit: Leaving.]
tianon has joined #ocaml
maufred_ has quit [Ping timeout: 245 seconds]
maufred has joined #ocaml
smondet has quit [Remote host closed the connection]
weie has quit [Quit: Leaving...]
talzeus has joined #ocaml
Neros has quit [Ping timeout: 245 seconds]
yacks has joined #ocaml
smondet has joined #ocaml
thomasga has quit [Quit: Leaving.]
cago has left #ocaml []
hkBst has quit [Quit: Konversation terminated!]
eikke has joined #ocaml
mika1 has quit [Quit: Leaving.]
ttamttam has quit [Quit: ttamttam]
Neros has joined #ocaml
malo has quit [Remote host closed the connection]
darkf has quit [Quit: Leaving]
djcoin has quit [Quit: WeeChat 0.3.9.2]
Kakadu has joined #ocaml
mcclurmc has joined #ocaml
eikke has quit [Ping timeout: 246 seconds]
Trollkastel has joined #ocaml
<fasta>
gasche: the preceding kind of shows that OCaml doesn't need more marketing, but more libraries.
<thelema_>
fasta: totally agreed; until recently, the OCaml ecosystem hasn't been conducive to using external libraries
<fasta>
thelema_: yeah, which is kind of surprising considering it is so old. Today it is a no brainer that if you want your language to succeed that you need Facebook widgets ;)
<fasta>
I think it's tools, tools, tools.
<fasta>
I also think it's better to *not* publish broken tools like Merlin.
<fasta>
If a certain community produces to much things which don't work, they will get ignored after some time.
<fasta>
too*
<thelema_>
I agree that it's important that tools be a certain level of working.
<thelema_>
I would guess that merlin's emacs problems come from the fact that its main developer is a vi user.
<thelema_>
and knows little of emacs
Trollkastel has quit [Read error: Connection reset by peer]
<asmanur>
what are merlin's emacs problem thelema_ ?
<asmanur>
fasta: flaming on IRC is easy; reporting bugs isn't, please save your time or to the latter.
<asmanur>
do*
<fasta>
asmanur: the problem is that if someone makes such basic mistakes, it means they aren't worth the time to report the bug in the first place.
tianon has quit [Ping timeout: 276 seconds]
<thelema_>
asmanur: I think mostly trouble installing it. I've gotten over that and am just annoyed by how it wants to auto-complete "in" with "ins".
<fasta>
asmanur: they should just take the hint and put on their site "There is a non-functional Emacs backend".
tianon has joined #ocaml
<fasta>
People use github like it is Twitter; just because you can write something, doesn't mean that you should publish it.
<asmanur>
thelema_: yes, the installation process is a bit hard -- but I believe that is common to any mode distributed with ocaml
<fasta>
Installation is not the problem.
<asmanur>
(because you have to tweak the load-path to include opam's share path)
<fasta>
It simply doesn't function like it should.
<fasta>
asmanur: that's trivial.
<fasta>
Blaming the user is really the problem here. It's the developer who cannot write an Emacs mode.
<fasta>
That's fine; not everyone needs to be able to do everything.
<asmanur>
fasta: 15 messages spent telling how merlin's emacs mode suck could be better spent describing specifically your problems
<asmanur>
thelema_: concerning completion, it is being improved
<fasta>
asmanur: that would come down to a lecture on Emacs.
<asmanur>
(integration with completion-at-point is in experimental)
Anarchos has joined #ocaml
<thelema_>
asmanur: awesome.
<asmanur>
thelema_: however ac should only trigger after three characters so your bug should not happen
<thelema_>
asmanur: maybe I need to update?
<asmanur>
thelema_: perhaps -- it cannot hurt but I believe the three characters limit is kinda old
<thelema_>
I'm using the 1.0.0 release from opam
<fasta>
asmanur: are you the developer?
<asmanur>
thelema_: this should work then
<asmanur>
fasta: yes
<fasta>
The developer of Merlin should obtain 1000 random modules from opam and then do random movements in the buffer and execute random commands of his mode for 10 days straight. If it then still works, then I will be interested.
<fasta>
This is something which everyone can easily do and it saves you this kind of discussions.
<fasta>
If 10 days is too much do less, but you need something like that in place.
<asmanur>
fasta: you know, you are the only feedback that negative
<def-lkb>
Probably, but as the developer of merlin backend, I see really no point in getting you more interested
<fasta>
asmanur: sure, blame the messenger.
pango_ has joined #ocaml
<asmanur>
and some users are actually managing to use it
<fasta>
asmanur: fine, then continue to suck.
pango has quit [Ping timeout: 252 seconds]
<fasta>
asmanur: I told you how to improve your processes.
<fasta>
asmanur: if you then thing you know better (you don't), you are free to ignore it.
<fasta>
But don't be surprised when other people will agree.
<asmanur>
fasta: when did you tell me ?
barbapapa has quit [Remote host closed the connection]
<fasta>
asmanur: I just did.
<fasta>
asmanur: and you just said implicitly it was a bad idea.
<fasta>
asmanur: but it would have prevented my comments if you had done that before.
<fasta>
asmanur: you don't cover all your cases like a pr0.
<fasta>
asmanur: so, then I suggested a way to let the machine tell you how you are not covering all cases.
<asmanur>
or you could tell me some cases that are not covered
<adrien>
fasta: please no name-calling nor insults
<fasta>
asmanur: that would take forever.
<asmanur>
don't be exhaustive then
<fasta>
asmanur: I don't understand how you are not getting the message.
<adrien>
also, remember that merlin is a young project that you have a worfklow and environment that is fairly different from the authors'
<fasta>
I also presented a solution for that: write with big lettters that the Emacs backend is not functioning.
<adrien>
that it works for your environment which is not theirs is something that's very good
<fasta>
I have asked the same for other pieces of software and those authors agreed with me in the end.
<fasta>
This includes professors of universities.
<fasta>
(those usually write the worst software)
<rks_>
well fasta, that's because they were obviously smarter
<rks_>
being educated and all
<fasta>
I think I only know one professor who doesn't suck.
<fasta>
Postdocs generally do know their stuff.
<fasta>
adrien: it's also my opinion that any Emacs mode is supposed to make the existing state of affairs strictly better. If I can choose between adding 100 commands of which 50 don't work or 1 command of which 1 works, I pick the latter.
<adrien>
yes
<adrien>
perfectly fine
<adrien>
but don't be aggressive towards others
<fasta>
If I would report every bug one by one, it would mean that I support cowboy style development.
<fasta>
If OTOH, a bug is in some complex piece of code that was a mere oversight, then there is a different issue and I will report that.
<fasta>
I think everyone should follow that, such that only good project will float up and bad projects die.
<adrien>
that I don't care but I don't want to see aggressivity on the channel