15:04 <hannes> hello

15:05 <yomimono> hello!

15:05 <TImada> hello!

15:05 <hannes> it's 16:00 BST and an odd week, should we get started on a MirageOS call!?

15:05 <hannes> feel free to adjust the agenda at https://github.com/mirage/mirage-www/wiki/Call-Agenda#11-july-2018 <now> ;)

15:06 <yomimono> yeah, let's do it :D

15:06 <hannes> ok, is reynir around? :)

15:07 <hannes> looks like he isn't :/

15:08 <hannes> in the meantime, TImada I've a question -- you commented on https://github.com/mirage/mirage-entropy/pull/39 (which adds mirage-entropy support for aarch64) -- does the assembly instruction do what it is supposed to do (i.e. the same as arm32)? ((I saw your comment about the isb, but failed to find any comment about the read_virtual_count)

15:10 <TImada> hannes: me?

15:10 <hannes> TImada: ups sorry mixed that up :(

15:11 <TImada> hannes: No problem!

15:13 <yomimono> OCaml 4.07 is out :)

15:13 <hannes> \o/

15:14 <TImada> Good work!!

15:14 <yomimono> (I had nothing to do with it but I'm excited anyway!)

15:14 <TImada> What are major updates on 4.07?

15:16 <hannes> from a MirageOS perspective (build system): bigarray is now part of stdlib, there is no longer a need to compile and link an otherlibs.a

15:16 <lobo> ohai everyone

15:16 <yomimono> from a user perspective: several error message improvements :)

15:17 <mort___> \o/

15:17 <hannes> also, in respect to bigarrays, the GC has been tweaked https://github.com/ocaml/ocaml/pull/1476 which should result in less expensive bigarray allocations - something we may be able to observe in MirageOS unikernels

15:19 <hannes> i have a partial PR for ocaml-freestanding 4.07 support at https://github.com/mirage/ocaml-freestanding/pull/39 (but there's some shell-makefile escape issue, I'm not sure how to escape $ from configure.sh -> Makefile include -> pkg-config file)

15:19 <TImada> hannes: yomimono: thanks, sounds good for me. :-)

15:24 <hannes> is there anything else someone wants to talk about? looks like we've plenty of time ;)

15:25 <TImada> I will have a presentation talk about MirageOS with Netmap at ICCBDC2018 http://www.iccbdc.org/

15:26 <yomimono> cool, congratulations TImada!

15:26 <hannes> TImada: \o/ nice - would you mind to send abstract + link to paper on mirageos-devel?

15:27 <TImada> Thanks! I will do it when my paper become available on the Web.

15:28 <TImada> expected in several weeks

15:32 <hannes> i finally deployed my automated let's encrypt certification -- https://github.com/roburio/udns/blob/master/mirage/dns_mirage_certify.mli allows you do request a certificate, you need to run https://github.com/hannesm/ocaml-letsencrypt/tree/nsupdate/mirage as a secondary NS though

15:38 <yomimono> hannes: since you have ops, any chance you could update the banner? next bug cleaning day would be 3 August, and next call 25 July

15:39 <yomimono> (also congrats on the let's encrypt setup :) )

15:39 <Drup> Having a library to do the whole let's encrypt setup would be really neet

15:39 <Drup> neat*

15:41 <hannes> Drup: that's how I deploy tls.nqsb.io and hannes.nqsb.io these days, see https://github.com/mirleft/tls-demo-server/commit/565fdbe972e0c92c49294cf2120bbfbc9021bba4 for example changes

15:41 <Drup> grmbl, I need to spend some time to device-ify dns and tls

15:42 <hannes> i.e. the unikernel needs a secret to talk to the name server, and a hostname, and a seed for the private key

15:43 <Drup> (well, me or someone else :p)

15:44 <Drup> But the current situation is really bad, you need to write so much boilerplate and functor apps by yourself to use TLS

15:44 <Drup> that's what devices are for, we shouldn't have to do that

15:46 <Drup> Sorry, slight rambling, please continue the mirage call :p

15:47 <hannes> Drup: I'm all in on making this easier to use, but tbh I wouldn't know how to do that..

15:49 <Drup> I'm sure yomimono is an expert at writing devices by now :p

15:49 <yomimono> yomimono wishes to make more experts rather than making more devices - pls suggest tactics :)

15:50 <Drup> I haven't been very succesful at that either, so I'm not a good person to ask

15:50 <yomimono> (also, once current messes are complete, probably it will be time to look at block devices which are in more need of significant overhaul imo)

15:50 <yomimono> (by "current messes" I mean "stuff I already promised to do")

15:51 <yomimono> but it's a good point -- it would be nice to have a better story for TLS, and I don't think we have an issue tracking that

15:51 <yomimono> there is one for defining the signatures for resolvers IIRC

15:51 <yomimono> which would go quite a ways toward making dns less painful

15:52 <hannes> i investigated this a bit, and there's an API for host name resolution in conduit, which is implemented by ocaml-dns... I wanted to extract this from conduit... and maybe rethink what it is supposed to do

15:52 <Drup> yomimono: in general, making a bullet lists of "things to do to make devices out of a library" would be nice

15:52 <Drup> that includes standardizing the signatures

15:53 <hannes> esp. now that i have some unikernels which use both udns and ocaml-dns.... because i failed to use udns with cohttp for host name resolution...

15:53 <yomimono> yeah, I was kind of trying to do that when writing up the docs about making mirage-qubes but that article is about too many things to be clearly useful

15:53 <yomimono> I agree that that would be a good document to have

15:54 <Drup> Well, that's section 3 of the wip functoria paper :>

15:54 <yomimono> ha, everything fell out of my head since I last read it :D relatedly let me know when you want a proofreading pass

15:54 <Drup> yeaaah, we postponed it again

15:55 <Drup> I'll yell at everyone in due time, in a few weeks

15:55 <yomimono> sgtm

15:57 <yomimono> (brb, relocating)

16:00 <Drup> hannes: small completely off-topic question: do you know about CVE caused by lack of santisation of extracted groups after parsing by a regular expression ?

16:00 <hannes> Drup: not off my head

16:01 <Drup> where should I look if I wanted to see if there are some instances of that ?

16:02 <hannes> Drup: I usually search bugtraq mailing list or the cve directly https://cve.mitre.org/

16:03 <Drup> "The modifier.regex_replace.php plugin in Smarty before 2.6.19, as used by Serendipity (S9Y) and other products, allows attackers to call arbitrary PHP functions via templates, related to a '\0' character in a search string. " *snort*

16:03 <reynir> sorry, got a surprise visit from my brother

16:17 <hannes> reynir: no worries, I pushed it to the next meeting in two weeks :)

16:48 <reynir> Ok, thanks :)

16:49 <reynir> (got a surprise visit by my parents just before haha)