IRC channel logs
2016-01-13.log
back to list of logs
<davexunit>your blog posts usually make it to the front page, but no one comments on them. <davexunit>probably because they are sufficiently technical and don't have click-bait titles <wingo>tx for kind words. fear not, the bloviation will continue <peterbrett_work>wingo: Unfortunately I can't make it to your FOSDEM talk due to other talks that are more relevant to the people who are paying for me to be there <davexunit>you'll be able to watch the recording later, at least. <wingo>somehow i got roped into three talks! <wingo>or maybe just one and a quarter <wingo>perhaps i can free myself from a snabby thing; was going to copresent with a colleague but the slot is short <wingo>then there is a panel, who knows what happens there, maybe we just say nice things and make a group hug <peterbrett_work>wingo: Re your blog post :) Guile's … lengthy … build is causing me problems with CI for Windows builds of gEDA <wingo>yeah :/ do you bundle guile? <peterbrett_work>Because there's no way to sensibly compile Guile as a .dll at the moment (thank you, GC) <wingo>understood. perhaps if you stick to one version though you can do the GUILE_FOR_BUILD trick <wingo>but yeah, i wish it were otherwise. i use luajit in another project and compared to guile it's a delight to embed :/ <peterbrett_work>Luckily I'm doing cross builds so I can use the host's distro-packaged Guile as GUILE_FOR_BUILD <peterbrett_work>(I think I might have to fork gEDA to get the people who are all, "But I can't install Guile 2.0 on my Gentoo / NetBSD / RHEL 4 workstation !!!!!" <ArneBab_>madsy: I tried getting mingw working last weekend, but ran into gentoo bugs, so it will be till after this week before I can try again (and find out why mingw doesn’t build). <peterbrett_work>ArneBab_: Also try out minipack please (just so I have some idea whether it only works on my machine) <ArneBab_>peterbrett_work: can you remind me again after this weekend? I have my PhD defense on Friday. <ArneBab_>Estimating the uncertainties of CO₂ surface fluxes from atmospheric measurements <ArneBab_>I implemented that part (for ground based remote sensing) <peterbrett_work>You know — I actually really like the suggestion of including the compiler's .go files in the release tarball <wingo>peterbrett_work: they are currently architecture-dependent. or at least, dependent on the word size and endianness of the target <peterbrett_work>wingo: Did you see my mailing list question about libguile on LLP machines? <wingo>peterbrett_work: no i didn't <peterbrett_work>So, basically, numbers.c refuses to compile on Windows x64 because it's LLP64 rather than LP64 <ArneBab_>LLP64 for backwards compatibility — that actually makes sense if you have to stay binary compatible. <ArneBab_>which Windows has to do, because it’s built as base for proprietary software <paron_remote>ACTION feels restless and doesn't want to work on the rest of his TODO list... works on the Guile tutorial he's writing instead. <karhunguixi>let me know if you want a newb to run through it at some point, and give some feedback from this perspective <paron_remote>(I promised myself I wouldn't explain car and cdr in this tutorial and I just did so anyway) <paron_remote>oh well, it's hard to get by in scheme without seeing a lot of cars and cdrs around anyway <rekado>why is explaining car and cdr really needed? Pattern matching is all the rage in programming languages, so couldn't we just advertise "match" instead? <rekado>every other book on scheme introduces car and cdr. <davexunit>rekado: I think pattern matching is a slightly more advanced notion <davexunit>introducing pairs and how to manipulate them is good <nalaginrut>rekado: maybe it's too easy to make exercises with match for a textbook ;-P <amz3>I tried to avoid introducing car & cdr, but for that you need to introduce literals for list and pair which is... more complex that explaining a procedure...