IRC channel logs

2015-09-06.log

back to list of logs

<daviid>ACTION starts to have some good results wrt his grip-clutter toolbox and examples he's building while working on our Guile-Clutter bindings: not optimal yet, but some good progress recently made
<paroneayea>minor blogpost http://dustycloud.org/blog/guile-errors/
<nalaginrut>morning guilers~
<taylanub>ACTION 's mind is blown as ey reads http://www.r6rs.org/r6rs-editors/2006-April/001196.html which is a concrete specification of the confusing psyntax behavior ey had experienced
<amz3>héllo :)
<amz3>thx paroneayea nice trick
<amz3>I tried to guess, the solution... I was like "no `gensym` is overkill"
***cluck` is now known as cluck
<paroneayea>amz3: yeah I figure gensym is fine enough
<taylanub>paroneayea: hmm, the delimited continuations system offers `make-prompt-tag' for the very same reason AFAIUI. BTW your examples made me laugh :D
<taylanub>"Make a new prompt tag. Currently prompt tags are generated symbols. This may change in some future Guile version." looks like they also thought gensym is fine enough
<taylanub>paroneayea: oh and gensym takes an optional argument for a prefix string so your code could be (very) slightly simplified
<Tsutsukakushi>hai
<os_>him
<os_>him
<os_>i mean hi
<os_>anyway!, guile-gnome 2.16.3 link is broken here: https://www.gnu.org/software/guile-gnome/download.html
<rlb>civodul: ok, fwiw, looks like we're OK in Debian on all (primary) architectures now: https://buildd.debian.org/status/package.php?p=guile-2.0&suite=sid
<rlb>(i.e. armel is fine now)
<civodul>rlb: cool!
<civodul>so the fix was just to use -O2?
<rlb>Well, the fix was to remove some old packaging cruft, and allow dpkg-buildpackage to handle the flags, which I believe does mean -O2. The original problem likely was the absence of any -O option on armel.
<rlb>It was causing some kind of compiler offset overflow.
<civodul>ok
<mark_weaver>rlb: it would be good to verify that guile is being compiled with optimization on all architectures.
<mark_weaver>even if it doesn't result in breakage, it might result in a slow guile.
<rlb>mark_weaver: it should be now (or doing whatever debian intends via dpkg-buildflags, which except in very special circumstances should be -O3).
<rlb>s/O3/O2/
<mark_weaver>okay, sounds good :)
<rlb>mark_weaver: it also adds various fortify and stack protection flags, which I'd imagined might cause trouble, but they didn't...
<rlb>civodul: just curious - is there any plan to change the way the auto compile logging works? i.e. in the long run, I wondered if that was the kind of thing most programs do (i.e. dump output whenever they're doing "normal" caching).
<rlb>It hasn't really been a problem for me yet, though it is noisy, but the Debian bug report made me wonder about it.
<civodul>rlb: i admit i've become tired of those messages :-)
<civodul>we should probably remove it in 2.2
<civodul>at least disable it by default
<rlb>I suppose I could imagine being silent unless GUILE_LOG_AUTO_COMPILE was set (or whatever).
<civodul>yeah
<civodul>something like that
<rlb>Since of course there might be times when you'd really like to see what's going on in more detail.
<rlb>(without strace ;>)
<civodul>yep
<govt>somebody posted some logs on reddit, and mark_weaver acts like a giant twat throughout, is this typical behavior for him?
<mark_weaver>govt: can you be more specific in your criticism, or show me the logs?
<govt>I'm not criticizing you, I'm laughing at you
<daviid>govt: we are no used to such low level, unfounded, personnal and therefore completely useless comments here: are you an adult? I suspect you're a child, how old are you?
<govt>lol
<govt>mark called somebody asking for help and idiot, and more than once at that
<govt>somebody called him out on reddit, so i read the logs and jump in here to confirm
<mark_weaver>that doesn't sound like something I would do.
<mark_weaver>how do you know the logs are genuine?
<govt>the logs dont lie, theyre the ones hosted on gnunet
<mark_weaver>can you show them to us please?
<mark_weaver>link please
<taylanub>mark_weaver: I think it's best to just ignore trolls...
<govt>literally JUST SAID they are posted on reddit
<taylanub>I put govt on /ignore already
<govt>lol
<govt>thats like twitch-reflex tier mad
<govt>all because mark cant answer some noob's questions without shitting himself in chat
<taylanub>are our weak hash tables ephemeral? e.g. if I map an element of a vector to the vector itself in a weak-key hash table, will the vector and its element be reclaimed when no references to the vector nor its element remain outside of the hash table, or will the key be kept alive by the vector, and thus cyclically the vector too?
<mark_weaver>govt: provide the link to the gnunet logs in question, or go away
<mark_weaver>taylanub: I believe they will be kept alive in that case, unfortunately.
<taylanub>ok
<govt>mark_weaver as if you have any kind of influence over me
<taylanub>MIT/GNU Scheme supports weak and "ephemeral" tables separately, because apparently ephemeral ones are somewhat size and time inefficient...
<mark_weaver>taylanub: as I recall, libgc doesn't have good support for weak pointers.
<mark_weaver>actually, most garbage collectors don't have very good support for them, and the last time I tried to sit down and decide on a good semantics for them, I found it quite difficult.
<mark_weaver>but we could certainly do better than we do today, if we were able to play more with our GC.
***govt is now known as mflatt
<daviid>ot: is there an easy and quick way to ask a browser [epiphany in my case] to reset a page [or all pages, doesn't matter] alinks so it does consider all alinks unvisited?
<taylanub>daviid: I would guess clearing the history should do, but I'm not sure. (I use my browser without any shared state between sessions.)
<daviid>taylanub: I tried but that didn't do it. I want to avoid to quit/restart [which appears to have the default you mention here as well
<taylanub>if clearing the history and reloading the page doesn't do it, then no clue :\\