malina has quit [Remote host closed the connection]
sparkyy has joined #ocaml
swalk has joined #ocaml
<Drup>
(that page is very very old, and hence partially wrong nowadays)
<rks`>
(were module types required to start with a capital letter in the past?)
andreas_ has joined #ocaml
<_y>
(is it not the case today?)
<zozozo>
it seems destructive substitution is quite restrtictive, i.e I can't seem to with something like: (A : S with type clause := formula list), though it can be done with a simple "=", is there a profound reason ?
<flux>
well, it seems it has been applied only to 4.05
sepp2k has quit [Read error: Connection reset by peer]
<flux>
if you are not able to update, patching your local ocaml 4.03 doesn't seem bad. just guessing the ticket might be moving slower during the summer re. releases for 4.03 and 4.04..
<theblatte>
so the goal is to fix ocaml while keeping a broken CPU?
sepp2k has joined #ocaml
<theblatte>
OCaml may not be the only thing affected by the bug
<flo2>
do i understand it correctly that just setting this fno-tree-vrp at configure time, fixes this issue?
<adrien>
flo2: how is the µcode update not possible?
<flo2>
adrien: you cant force everyone to do this code upgrade, unfourtnatly
jlam__ has joined #ocaml
zwild has quit [Ping timeout: 268 seconds]
<flux>
a library to reliably detect the situation would be nice :)
jlam_ has quit [Ping timeout: 240 seconds]
<flo2>
so what is the exact way to call configure?
dhil has joined #ocaml
<flo2>
flux: is there some kind of roadmap for the patch releases for 4.03?
<adrien>
flo2: depends on the setup
<adrien>
on linux distributions, you could
<adrien>
even on windows it could be OK
<flo2>
adrien: but on windows it does not affect because it does not use gcc?
<adrien>
you can use gcc on windows
<adrien>
but I see
<flo2>
adrien: we are using vc140
<adrien>
so, ok
samrat_ has quit [Ping timeout: 248 seconds]
<flo2>
adrien: but i am still uncertain what i should write as argument for the configure
butts_butts has quit [Remote host closed the connection]
aciniglio has joined #ocaml
zpe has joined #ocaml
jlam_ has joined #ocaml
jlam__ has quit [Ping timeout: 255 seconds]
zpe has quit [Ping timeout: 248 seconds]
<flo2>
or would it be better to just wait for a new 4.03 release with that patch?
freusque has quit [Quit: WeeChat 1.9]
freusque has joined #ocaml
<adrien>
is it planned to do that?
<freusque>
hello everyone. What are the advantages of a source-level package manager? I'm trying to understand what is the rationale of having opam versus using external tools such as nix.
<theblatte>
it would be great if opam shipped binaries too
govg has joined #ocaml
<octachron>
adrien, flo2, afaik, there is no planned bug fixes release of 4.03
<flo2>
octachron: thanks for that info, will that patch be applied to the next 4.04 release?
<octachron>
flo2, note that the patch only decreases the probability of the hardware bug being trigered by OCaml GC
<octachron>
without the corresponding μcode update, the bug can still be trigered by other sources
<flo2>
octachron: only decrease? i thought that would make sure that ocaml can not trigger this issue?
<flo2>
octachron: ok
<octachron>
flo2, at the core, this is an hardward bug ; it just happened that one of the OCaml gc tight loop triggered it with a quite high probability. The patch fix this probability to go back to normal background probability
jlam_ has quit [Ping timeout: 246 seconds]
jlam_ has joined #ocaml
<def`>
octachron: is that explained somewhere ?
maarhart has joined #ocaml
jlam__ has joined #ocaml
ryanartecona has joined #ocaml
Mercuria1Alchemi has joined #ocaml
maarhart has quit [Client Quit]
jlam_ has quit [Ping timeout: 240 seconds]
<gasche`>
flo2: we don't know if -fno-tree-vrp avoids the issue entirely
<gasche`>
we only know that it was not reproduced with the flag used, I think
jlam_ has joined #ocaml
<gasche`>
(I don't know if backports of 4.03 or 4.04 with the fix are planned, but this can be discussed; could you tell a bit more about what your use-case is, how you install OCaml, and why moving to more recent versions is not an option?)
<def`>
octachron: ahh I see, you were talking about ocaml patch, not microcode patch :P
<gasche`>
(yeah I also just noticed that octachron already said all of the above, apologies)
<flo2>
gasche: ocaml 4.05 will be released soon?
<octachron>
`def: yes, and that was my tl;dr version of various dicussions on the subject, I am not sure if there is a synthetic source on the subject somewhere
<def`>
octachron: yes that makes sense
jlam__ has quit [Ping timeout: 260 seconds]
<octachron>
flo2, is one week ago soon enough?
enterprisey has joined #ocaml
ziyourenxiang has joined #ocaml
FreeBirdLjj has joined #ocaml
FreeBirdLjj has quit [Read error: Connection reset by peer]
FreeBirdLjj has joined #ocaml
MercurialAlchemi has quit [Ping timeout: 260 seconds]
<flo2>
octachron: yes thats very good news :)
Guest11402 is now known as ousado
ousado has quit [Changing host]
ousado has joined #ocaml
zpe has joined #ocaml
<ousado>
Hi - it seems that a repeated external function call (get_time) is being optimized away and the result is being replaced with the result of the first call, is there an annotation to prevent that?
MercurialAlchemi has quit [Ping timeout: 268 seconds]
MercurialAlchemi has joined #ocaml
tane has quit [Quit: Leaving]
TarVanimelde has joined #ocaml
jlam__ has joined #ocaml
jlam_ has quit [Ping timeout: 240 seconds]
ryanartecona has joined #ocaml
mengu has quit [Quit: Leaving...]
sam____ has joined #ocaml
aciniglio has quit [Ping timeout: 276 seconds]
sam____ has quit [Ping timeout: 268 seconds]
TarVanimelde has quit [Quit: TarVanimelde]
maxton has joined #ocaml
<maxton>
Why is it that (exception) constructors don't support partial application?
infinity0_ has joined #ocaml
infinity0_ has quit [Changing host]
infinity0 has joined #ocaml
infinity0 has quit [Killed (verne.freenode.net (Nickname regained by services))]
<maxton>
They seem to work like functions, but if I have `` exception Ex of string `` I can't pass Ex to a function, but instead have to use (fun x -> Ex x)
<zozozo>
maxton: because constructors are *not* functions
<zozozo>
they are treated differently and also have a different syntax
jlam_ has joined #ocaml
jlam__ has quit [Ping timeout: 248 seconds]
pierpa has joined #ocaml
rpcope has quit [Ping timeout: 246 seconds]
rpcope has joined #ocaml
sam____ has joined #ocaml
enterprisey has quit [Read error: Connection reset by peer]
sam____ has quit [Ping timeout: 276 seconds]
TheLemonMan has quit [Quit: "It's now safe to turn off your computer."]
argent_smith1 has joined #ocaml
argent_smith has quit [Ping timeout: 240 seconds]
ryanartecona has quit [Quit: ryanartecona]
argent_smith has joined #ocaml
<Leonidas>
at least in OCaml. In SML and Caml Light they were like functions.
peterpp has joined #ocaml
argent_smith1 has quit [Ping timeout: 276 seconds]
<Leonidas>
zozozo: weird, I believe I compiled CL some time ago and constructors were like functions. Maybe I misremembered
rpcope has joined #ocaml
<zozozo>
^^
<Leonidas>
at least Caml had this at one point in its history
<lapinot>
imho this is consistant with the fact that constructors are not curryfied (although haskell takes the other path and has them curryfied and function like
<lapinot>
i do prs
<lapinot>
sry, on my phone... I do prefer the haskell way: there is one obvious way to assign a type to a constructor, and it's the type of a function, no matter how different it is from the compilers point of view
sh0t has joined #ocaml
cranmax has joined #ocaml
dakk has quit [Remote host closed the connection]
<_y>
Leonidas, it was, syntax-wise
kakadu_ has quit [Remote host closed the connection]