hannes changed the topic of #mirage to: https://mirage.io - bug cleaning day every first friday in month (14:00 UTC - late, next: April 6th) - next call April 4th 16:00 BST https://github.com/mirage/mirage-www/wiki/Call-Agenda - this channel is logged at http://irclog.whitequark.org/mirage/ - MirageOS 3 is released - happy hacking!
ousado has quit [Read error: Connection reset by peer]
ousado has joined #mirage
dmj` has joined #mirage
olle has joined #mirage
Haudegen has joined #mirage
Ulrar has joined #mirage
<Ulrar> Hey, quick question, do I need to do anything specific to get DNS resolution in mirageos ? When I try to do a Client.get (cohttp) I get a "name resolution failed: unknown endpoint type"
argent_smith has joined #mirage
argent_smith has quit [Quit: Leaving.]
<Ulrar> Seems like the answer is yes, you do have to do something. I figured it out reading the http-fetch example
<Ulrar> Now figuring out how to handle tls ..
mort___ has joined #mirage
mort___ has quit [Client Quit]
mort___ has joined #mirage
mort___ has left #mirage [#mirage]
argent_smith has joined #mirage
Haudegen has quit [Remote host closed the connection]
mort___ has joined #mirage
Haudegen has joined #mirage
mort___ has left #mirage [#mirage]
<mato> Is there a catch-up today at 17:00? There's been no announcement so far and nothing on the Wiki page.
<mato> 17:00 CEST that is.
<hannes> mato: yes, I'll post a mail now
<kit_ty_kate> hi everyone, is there a parser that would take ssh private keys and would give a Nocrypto compatible type (e.g. Nocrypto.Rsa.priv) somewhere already ?
<kit_ty_kate> reynir maybe ?
<mato> kit_ty_kate: @haesbaert might have something like that...
TImada has joined #mirage
ricarkol has joined #mirage
djwillia has joined #mirage
yomimono has joined #mirage
<hannes> hello!
<yomimono> hello! I was just reading your pinata post at https://hannes.nqsb.io/Posts/Pinata
<djwillia> hi all!
<hannes> dan! :) yomimono! :)
<lobo> hi
<TImada> Hi all!
* mato waves
<kit_ty_kate> \o yomimono
<kit_ty_kate> hi all
<Drup> hi
<ricarkol> hello everybody
<Drup> yomimono: so, what's your next library on your whacking list ? :p
<hannes> welcome to the mirageos call! :) the agenda is at https://github.com/mirage/mirage-www/wiki/Call-Agenda#18-april-2018
* Kensan waves
<hannes> atm there aren't many items on the agenda, we can discuss any other business at the end (feel free to add your topic there just now)
<yomimono> drup: suggestions? ;)
<hannes> first item is from myself, mirage-fs API discussion (https://github.com/mirage/mirage-fs/issues/20)
<hannes> background story is that I started to use mirage-fs in my newest project (caldav server), and I've some suggestions for improving the API.
<Drup> yomimono: you should open a letterbox, people that feel their libraries are up to the challenge!
<kit_ty_kate> hannes: thanks! Everything seem to be in x509.Encoding.Pem (it seems so anyway)
<hannes> feel free to raise your opinion in the mirage-fs issue
Lortex has joined #mirage
<hannes> and off to the next topic, Solo5 development mailing list pre-announcement (@mato)!
<yomimono> as with all big and sensible API changes, I think we should do it and regret that we didn't a year ago
<hannes> yomimono: thanks to the efforts to split up mirage-types into a zillion libraries, it's now easier to change one API, such as the FS, at a time :)
<hannes> mato: would you mind to talk about your pre-announcement?
<mato> Right, so some time ago I conducted an informal vote among the top-5 Solo5 contributors about what to do for an additional communication channel for open-ended Solo5 development discussions
<mato> And the overwhelming preference was to set up a new mailing list
<mato> This is now (thanks mainly to hannes) ready for use, and I'm pre-announcing it here, feel free to subscribe at solo5-subscribe@lists.h3q.com
<mato> I will announce this on the mirageos list and also add it to Solo5 README.md once a majority of the core folks have subscribed to the list
<hannes> (there's btw as well https://www.mail-archive.com/mirageos-devel@lists.xenproject.org/ in case you're wondering where to browse archives)
<hannes> thanks mato for getting this rolled out!
<mato> In other random Solo5 news, ukvm on OpenBSD + Mirage/Solo5 is getting closer to merging, coming to a repo near you soon! :-)
<yomimono> :D
<hannes> TImada: btw you may want to open a pull request for your solo5-netmap branch (by clicking on "create pull request" at https://github.com/Solo5/solo5/compare/master...solo5-netmap:netmap -- so that one day this will become upstream :)
<djwillia> very cool on the OpenBSD!
<Kensan> +1! (Lot of persistence on by Adam on OpenBSD!)
<hannes> the OpenBSD stuff was mainly done by adam steen / theotheradam (who's based in australia and an OpenBSD contributor!)
<TImada> hannes: I want to do ..., I'm now checking the new APIs on Solo5. Please wait a while!
<djwillia> mato: ricardo and I are getting closer to the point where we will want to talk about the steps to move forward with the "seccomp backend" so hopefully we can start that discussion in another week or two
<hannes> TImada: sure :)
<hannes> djwillia: \o/
<mato> That sounds great. My current priorities are a) mailing list setup b) test and merge openbsd support c) arm64 ci
<mato> then I'd basically like to do the long-proposed renaming, write a bunch of docs and do a release ASAP
<mato> So IMO things like netmap, seccomp, nikhil's shmstream-based net device will be best tackled after a release.
<djwillia> sounds good!
<hannes> mato: yes, I agree. otherwise there will never be a solo5 release (and the last one is from more than a year ago)
<mato> djwillia: ricarkol: the sooner you guys subscribe to the list, the sooner i'll send out the announcement about the list :)
<djwillia> just sent my confirm to the bot :)
<Kensan> mato: Looking forward to a Solo5 release since that will mean it will come with Muen support :)
<mato> hannes: indeed. so we stick to the things that are in-flight and almost ready, and do the rest after
<djwillia> oh wow it has been a long time since a release if the Muen support isn't in there!
<hannes> (https://github.com/Solo5/solo5/releases/tag/v0.2.2 is the most recent release)
<mato> yes, forever.
<hannes> djwillia: neither the FreeBSD support
<mato> OT: Anyone looked into the recent Oracle announcement of "GraalVM"? The "polyglot" language interop looks interesting, possibly also relevant to multi-language libOS scenarios.
<Drup> mato: graal isn't really new
<Drup> (iirc, they sort of force a common representation, since everything boils down to a common VM)
<yomimono> slightly more on-topic: kit_ty_kate made a nice tool for checking whether packages are ready for ocaml 4.07, which is now in rc2
<mato> Is it JVM-based? Was not clear to me from an initial skim of the announcements.
<Drup> mato: yes
<yomimono> you can see conversation about it at https://discuss.ocaml.org/t/ann-opam-check-all-are-your-packages-broken-for-the-upcoming-ocaml-releases/1875/2 , and we should try to make sure our wodge of packages are as ready as possible (given that the xen stuff is still lagging)
<mato> Drup: ah, ok. Then mainlu the language interop is the new bit I guess. I've not seen any serious attempts at that before.
<mato> yomimono: I also noticed that Lwt 4.0.0 has been release, IIRC in marrakesh we found a few things in MirageOS that will break with that...?
<hannes> mainly Lwt_sequence is now deprecated (or already removed?)
<yomimono> mato: yes, there's a tracking issue in mirage/mirage for that.
<yomimono> biggest issue is indeed lwt_sequence
<yomimono> dinosaure is working on a replacement for that :)
<mato> thanks, have subscribed to that
<hannes> AoB? (remember the next retreat will be in Marrakesh from October 3rd till 10th. book your journey now ;)
<mato> one other things I wanted opinions on is derpeter's writeup on ARM64, https://github.com/mirage/mirage-www/pull/604
<mato> specifically whether or not mirage-www is the right place (both topic-wise and ease-of-process-wise) for that
<mato> feel free to state your opinion in the PR
<mato> djwillia: ricarkol: It'd be awesome if you could make it to the next retreat
<djwillia> yes it was a great time back in december
<hannes> yes -- and for the next we should maybe get TImada, wei chen, adam steen, Kensan there as well and have some solo5 discussions :D
<yomimono> hannes: the CI failures are because fat-filesystem version 0.12.1 didn't include the `fat` binary called by the test scripts
Haudegen has quit [Remote host closed the connection]
<mato> yup. we should start making a list of who will invite who in the robur discourse, so we don't forget
<mato> hannes: also gianluca :)
<yomimono> versions 0.12.2 and 0.12.3 have it, so perhaps 0.12.1 should be marked unavailable in opam-repository
<hannes> mato: yes!
<hannes> yomimono: thanks for investigating, I'll make a PR in opam-repository
<Kensan> hannes: It's marked in my calendar but not booked yet.
<hannes> (https://github.com/ocaml/opam-repository/pull/11821 is the fat-0.12.1 being marked unavailable)
<yomimono> pretty crusty set of selections by the solver for that test run
mort___ has joined #mirage
<yomimono> thanks for running this meeting, hannes!
<hannes> oh, i forgot to say it is over now, but i suspect you people inferred this by the silence
<hannes> yomimono: you're welcome! :)
yomimono has quit [Quit: "time to play in the snow"]
<hannes> so thanks for participating, happy hacking, and see you again in two weeks! (on May 2nd)
<mato> thanks hannes for leading the meeting!
Lortex has quit [Quit: Page closed]
TImada has quit [Ping timeout: 260 seconds]
thomasga has joined #mirage
<thomasga> hi there, I guess I am late :-)
<mato> yeah, you missed all the fun :)
<hannes> luckily there's a recording at https://irclog.whitequark.org/mirage/2018-04-18
<thomasga> (yes I am re-reading that :p)
<hannes> thomasga: is there a plan to implement mirage-fs interface in terms of irmin/git? AFAICT, wodan does irmin->mirage-block..
<thomasga> no plan as mirage-fs is too crap :-) but adding something wouldn't be difficult (we are already implementing KV_RO and the full Irmin API: https://github.com/mirage/irmin/blob/master/src/irmin-mirage/irmin_mirage.mli
<thomasga> but that would be a good idea once we have a better mirage-fs signature
<hannes> well, on my/our part, we develpo against mirage-fs for persistent storage, and instead of using mirage-fs-mem, i'd appreciate one day to use a git repository :D
<hannes> on another topic: shows one of the 4.07.0 issues... how are we supposed to fix this? pass on -thread for everything that uses lwt?
<kit_ty_kate> I remember seeing something about -thread being deprecated or something but I'm not sure
<kit_ty_kate> but I have no clue here :/
<kit_ty_kate> oh wait
Haudegen has joined #mirage
<kit_ty_kate> oh no that's fine I thought it was using a bad version of ocamlfind
<kit_ty_kate> oh no it did
<hannes> oh and it seems that we need lwt.4.0.0 first, then we can ocaml 4.07.0...
<kit_ty_kate> hannes: ok I'm rerunning everything now
<kit_ty_kate> that sure fix the issue
<kit_ty_kate> it should be finished by tomorrow morning
<hannes> ok
<kit_ty_kate> s/sure/should/
<thomasga> hannes: do you want every write to commit with a non-descriptive commit message?
<thomasga> (it's certainly possible to do it, it was just not totally clear to me why it would be useful instead of using directly the Irmin API)
<thomasga> but it makes sense, if we want to make that a parametrizable device
<hannes> thomasga: for portability? ;)
<hannes> thomasga: does Irmin.XX extend mirage-fs? should it?
<thomasga> it would be great if it wasn't too far away. But `Irmin.write` will have to take a commit/author message
<thomasga> so perfect inclusion seems a bit difficult
<hannes> but yes, i see your point. in webdav country we have so-called properties, and by using http we have (a) an author (b) a user-agent...
<hannes> (but now food first, i can see your point and am not worried too much about rebasing our work on irmin API) :)
<thomasga> but maybe yes, have a look at the functions defined there https://mirage.github.io/irmin/irmin/Irmin/module-type-S/index.html#val-kind
<thomasga> maye the transaction type (https://mirage.github.io/irmin/irmin/Irmin/module-type-S/index.html#type-transaction) could be made abstract and be defined differently between the various backends
<thomasga> Also I think it's a mistake to not have partial reads (with offset/lenght) in the Irmin API, that will probably appear at some point.
mort___ has quit [Ping timeout: 264 seconds]
djwillia has left #mirage ["ERC (IRC client for Emacs 25.3.1)"]
ricarkol has left #mirage [#mirage]
mort___ has joined #mirage
pagurus` has joined #mirage
pagurus has quit [Ping timeout: 240 seconds]
pagurus` has quit [Ping timeout: 240 seconds]
pagurus has joined #mirage
thomasga has quit [Quit: Leaving.]
mort___ has quit [Quit: Leaving.]
mort___ has joined #mirage
Haudegen has quit [Remote host closed the connection]
<reynir> kit_ty_kate: there's a few different formats for ssh private keys
<kit_ty_kate> reynir: I had RSA without password working using the x509 library
<reynir> ok great :-)
Haudegen has joined #mirage
<kit_ty_kate> with password seems a lot harder though but I think I will just avoid using passwords for now ^^"
argent_smith has quit [Quit: Leaving.]
<reynir> Yea :/
<hannes> kit_ty_kate, reynir: if you find a spec what crypto / standard is used by ssh keys with password encryption, please let me know... I've seen various in the X.509 setting (and still have the long-term plan to implement them, but providing 3DES isn't anything I consider to be useful), I'm esp. missing ASN.1 OIDs for aes ciphers...
<hannes> thx thomasga for the links, I'll think more about it over the next days/weeks :)
<reynir> I think there's even more than one format with password encryption
<reynir> There's a newer format with more rounds or something in the key derivation
<reynir> I looked into the formats a bit a while ago and decided it was too complex
<hannes> reynir: i looked into pkcs8 and could only find 3des there :|
<hannes> "too complex" does not apply to me ;)
<reynir> heh
<hannes> i mean i'm not afraid of complexity here, more about implement weak crypto... i feel better in the honest way: no encryption, than weak encryption (3DES, red13)
<reynir> There's this, but I'm not sure what it applies to. I don't think it's the new format(?) https://github.com/openssh/openssh-portable/blob/master/PROTOCOL.key
hannes changed the topic of #mirage to: https://mirage.io - bug cleaning day every first friday in month (14:00 UTC - late, next: May 4th) - next call May 2nd 16:00 BST https://github.com/mirage/mirage-www/wiki/Call-Agenda - this channel is logged at http://irclog.whitequark.org/mirage/ - MirageOS 3 is released - happy hacking!
<reynir> hannes: but wrt. weak crypto I think then only the new format is relevant
<reynir> oh, my key is in the old format :|
<reynir> hannes: so, the new format seems to be documented in PROTOCOL.key and the implementation is here https://github.com/openssh/openssh-portable/blob/dc31e79454e9b9140b33ad380565fdb59b9c4f33/sshkey.c#L3239
<reynir> good night o/
Haudegen has quit [Remote host closed the connection]