copy` has quit [Quit: Connection closed for inactivity]
mort___ has joined #mirage
mort___ has quit [Quit: Leaving.]
mort___ has joined #mirage
mort___ has quit [Quit: Leaving.]
insitu has joined #mirage
insitu has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jermar has quit [Ping timeout: 240 seconds]
srenatus[m] has quit [Ping timeout: 272 seconds]
AltGr has joined #mirage
mort___ has joined #mirage
insitu has joined #mirage
insitu has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
srenatus[m] has joined #mirage
dexterph has joined #mirage
mort___ has quit [Quit: Leaving.]
jermar has joined #mirage
mort___ has joined #mirage
insitu has joined #mirage
insitu has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mort___ has quit [Quit: Leaving.]
yomimono has joined #mirage
mort___ has joined #mirage
<yomimono>
yay, new opam-depext with OPAMYES is in master of opam-repository!
* yomimono
runs around restarting travis builds
rgrinberg has joined #mirage
rgrinberg has quit [Client Quit]
rgrinberg has joined #mirage
yomimono has quit [Ping timeout: 264 seconds]
mort___ has quit [Quit: Leaving.]
dexterph` has joined #mirage
dexterph` has quit [Remote host closed the connection]
dexterph has quit [Ping timeout: 260 seconds]
dexterph has joined #mirage
insitu has joined #mirage
agarwal1975 has joined #mirage
mort___ has joined #mirage
yomimono has joined #mirage
<yomimono>
mirage configure -t
<hannes>
\o/
copy` has joined #mirage
<Drup>
yomimono: sorry for reacting a bit late
mort___ has quit [Quit: Leaving.]
<yomimono>
drup: my fault; I should've tagged you rather than assuming silence was agreement
<Drup>
well, tagging me by email wouldn't have helped, I was in holiday
<yomimono>
:D
<yomimono>
in that case I'm extra glad I didn't!
<yomimono>
(hobbies include not bothering people while they're on holiday)
insitu has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Drup>
(IRC and personal email is best for important things, github emails are just a big mass of stuff, I had 900 github notifications to catch up ..)
<yomimono>
yeah, I need an autofilter for them myself -- configure my mail client to sort them by repo or something
<yomimono>
right now I miss stuff all the time
<Drup>
Aliases are also a bit of an obscure feature
<Drup>
I was not very satisfied by the API, and I introduced them mostly to be able to add --unix/xen (and the hypothetic --release for www, which was never done), but I didn't advertized them, since the API is quite limited
<yomimono>
that does explain a lot
<yomimono>
--release also seems like a very weird thing to plumb in in that way
<yomimono>
I don't understand the semantics of that, the more I try to think about it
<Drup>
I don't know, I was never really convinced, by thomasga emited the idea, and the API allowed to implement both with a reasonable semantics
<yomimono>
it seems like a thing that it would make more sense to have an as an independent key anyway
<Drup>
well, it's not really an independent key
<yomimono>
...I definitely don't understand it properly, then
<Drup>
ok, maybe the name is bad then
<yomimono>
what was --release meant to do?
<Drup>
let's consider the case of www: when deploying, it's always deployed the same way, right ?
<yomimono>
in our current use case, sure, we always deploy a xen unikernel
<yomimono>
but I can imagine having "release" that might mean different backends
<Drup>
it's always "--net direct --dhcp false --ip 46.43.42.146
<yomimono>
aaaaaaah, okay, so it's more like a shortcut for a whole mess of keys
<Drup>
you could make an alias that would set all those options in one flag
<Drup>
yes
<Drup>
technically, you can also make aliases that are regular options, but there was an issue with that, so we removed it from the API, there are only alias flags now
<yomimono>
I suppose you could still implement this as an independent key and then have some logic in config.ml to set the other keys if that one is set?
<Drup>
yes, it's just more britle because you are reinventing the logic
* yomimono
nods
<Drup>
it works both ways too
<yomimono>
?
<Drup>
for example, if you have --foo an alias for -a 1 -b 2
<Drup>
and then you configure with -a 1 -b 2, then --foo is set to true directly
<Drup>
(including in the Key module in the unikernel)
<yomimono>
ah, I see
<yomimono>
that's quite nice
<Drup>
yeah, the design is okay, but there are issues when the value of alias are not booleans
<Drup>
(and CSP is really a mess for this semantics, so it's only configure-stage)
<Drup>
(if you look inside the implementation, Alias.opt is commented :D)
<yomimono>
It sounds like everyone who has an opinion is at least slightly in favor of removing it
<Drup>
Well, nobody really used it
<yomimono>
as is clear, I didn't even understand it :P
<Drup>
(which means you can remove it without anyone complaining :D)
<yomimono>
net contribution to functoria's gonna be so negative :D
<Drup>
that's fine, I had to introduce a lot of stuff inside functoria only so that it didn't break anything, we/you are now removing those things, I'm fine with that :D
<Drup>
(the implementation of aliases, in particular, was really difficult)
<yomimono>
:(
<Drup>
You're going to remove a bit more than simply the Alias module :D
<yomimono>
In that light I especially appreciate your good spirit about it
<yomimono>
oh dear
<yomimono>
...should I not step in this, maybe?
<Drup>
too late, you volunteered :D
<yomimono>
...story of my life
<Drup>
more seriously, though, we can simply remove --unix/xen, then maybe explain what aliases actually are and propose to deprecate them
<Drup>
(or just remove them instead of deprecating, since I doubt it's used anywhere)
<yomimono>
are there non-Mirage users of functoria?
<Drup>
not at the moment, no
<yomimono>
I couldn't speak for them but I haven't seen it used anywhere other than the mirage front-end tool
<Drup>
(I would have liked it, but I don't have time to work on it)
<yomimono>
it seems, though, that maybe a design that you're happier with could result from working under different constraints in the future
<yomimono>
(i.e. without people demanding that the same legacy stuff keep working)
<Drup>
I'm not sure I'm parsing this sentence correctly
<yomimono>
sorry, let me try again:
<yomimono>
if someone wants something like alias in the future, but there's no need to support stuff the specific legacy stuff you were designing around, one might get a better thing like Alias
<yomimono>
removing it now isn't necessarily removing it forever; it can come back better-stronger-faster
<Drup>
ah
<Drup>
yes, I agree
insitu has joined #mirage
<Drup>
also, if that was the case, it would come with an actual use case.
<yomimono>
:)
dexterph has quit [Ping timeout: 240 seconds]
<Drup>
grmbl
<Drup>
I'm midly unconvinced by the topkg conversion
<hannes>
Drup: could you elaborate which parts you're unhappy with?
mort___ has joined #mirage
rgrinberg has quit [Ping timeout: 240 seconds]
brson has joined #mirage
insitu has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
rgrinberg has joined #mirage
jermar has quit [Ping timeout: 276 seconds]
yomimono has quit [Ping timeout: 244 seconds]
rgrinberg has quit [Ping timeout: 252 seconds]
rgrinberg has joined #mirage
jermar has joined #mirage
AltGr has left #mirage [#mirage]
<Drup>
hannes: I already said quite a few things on slack, but basically, I consider oasis's project description superior in every way. Now all sort of information are either lost in the minutia of ocamlbuild's instructions (which are terrible to organize sublibs) or duplicated across many files.
<Drup>
Additionally, it adds a build dependency that I consider shouldn't be there.
<mort___>
drup: on slack? which channel…
<Drup>
topkg does only one things: distribution. It does it correctly, but it forgets all the other stuff, that oasis did quite well. That works well for bunzli libs, which are mostly one-file libs with no significant internal structure.
<Drup>
mort___: probably random or general
<mort___>
ah ok must've missed it
<Drup>
mort___: tbh, I tried to avoid a trollfest with him, so I didn't say much
rgrinberg has quit [Ping timeout: 250 seconds]
mort___ has quit [Quit: Leaving.]
rgrinberg has joined #mirage
insitu has joined #mirage
agarwal1975 has quit [Ping timeout: 258 seconds]
insitu_ has joined #mirage
insitu has quit [Ping timeout: 252 seconds]
insitu_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
insitu has joined #mirage
insitu has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]