2026-06-10 07:33:31 <iv4nshm4k0v > 
ACTION listens to http://he3.magnatune.com/all/03-Open%20the%20Source-Paolo%20Pavan.mp3 

2026-06-10 07:38:48 <DKordic     > KipIngram: Re: Self-Mddifying Code; I completely agree.  Maybe Linux is a stupid joke.
2026-06-10 07:38:49 <DKordic     > I can't find the ChangeLog.  In recent Linux versions all Processes are ""mmap""ed with their ELFs.  Changes to the File will be reflected (""sync"") in the relevant Processes!!
2026-06-10 08:15:44 <iv4nshm4k0v > So far as I can recall, it's been the case for as long as I've used Linux.  E. g., RedHat 5.2, or perhaps it was some other distributions using Linux 2.0.36 and libc6, already had such behavior.
2026-06-10 08:20:39 <iv4nshm4k0v > So when, say, upgrading shared libraries (especially the "core" ones, such as aforementioned libc), the package manager has to be careful to unlink (or rename) the destination files before writing new ones.  Or else.
2026-06-10 08:29:33 <xentrac     > yeah, I think it was already the case in libc5
2026-06-10 08:30:13 <xentrac     > though I'm not sure if I ever upgraded a Linux using libc5.  I was using Slackware until well after that
2026-06-10 09:50:28 <Stalevar    > http://put.nu/files/yJcR_Yx.png
2026-06-10 10:13:50 <veltas      > KipIngram: Guess it depends which ABI / OS the competition is for
2026-06-10 10:21:01 <veltas      > This is such an awesome project https://c64os.com/
2026-06-10 10:22:08 <veltas      > In fact I think Linux added ELFs originally so they could use mmap, which a.out didn't allow in general (?)
2026-06-10 10:22:29 <veltas      > Or COFF or whatever they used before
2026-06-10 14:00:46 <MrMobius    > Stalevar: ?
2026-06-10 14:01:01 <Stalevar    > MrMobius, ?
2026-06-10 14:02:17 <MrMobius    > Stalevar: what's the png you posted?
2026-06-10 14:02:56 <Stalevar    > MrMobius, it's Pascal code which renders txt files into picture run through itself
2026-06-10 14:04:04 <Stalevar    > I have implemented substitute characters for what doesn't exist in the custom font by cloning some font8x8.c from linux kernel source and using sed to replace 0x with $ and  /* */ with { }
2026-06-10 14:04:46 <Stalevar    > They are rendered in square tofu marks similar to how non-existing UTF-8 show up as tofu with hex codes
2026-06-10 19:22:43 <oak         > Good morning! I was trying to figure out if there's a built-in word for printing a number without trailing space or newline. Anyone know of one?
2026-06-10 19:24:01 <lofty       > I don't entirely understand forth myself, but, I thought this could be done by "#s"?
2026-06-10 19:25:07 <lofty       > (as the saying goes, "post a wrong answer and somebody will come along and correct you" and then I'll learn something)
2026-06-10 19:25:21 <oak         > hehe
2026-06-10 19:25:39 <oak         > lofty: I looked at that. It seems part of "pictured numeric output", which I had a hard time understanding.
2026-06-10 19:25:50 <oak         > #s by itself gives me a stack underflow.
2026-06-10 19:26:10 <lofty       > "#s" needs to be inside "<#"/"#>"
2026-06-10 19:26:29 <lofty       > and the result is a "c-addr u" string you can "type"
2026-06-10 19:27:37 <oak         > Do you know what lookahead word I use with TYPE?
2026-06-10 19:28:08 <oak         > Like, https://gforth.org/manual/Formatted-numeric-output.html#index-_0023s--ud-_002d_002d-0-0--core uses "TYPE SPACES"
2026-06-10 19:28:42 <xentrac     > oak: maybe use 0 .r
2026-06-10 19:28:46 <oak         > oh that's not lookahead is it? >_<
2026-06-10 19:28:54 <xentrac     > .r is to print a number right-justified in a fixed-width field
2026-06-10 19:29:07 <xentrac     > if the number doesn't fit in the field it prints all the digits and horks the field width
2026-06-10 19:29:12 <xentrac     > so you can just print in a 0-width field
2026-06-10 19:29:35 <xentrac     > like, 37 0 .r 53 0 .r says "3753"
2026-06-10 19:30:15 <oak         > Aah. Thank you xentrac.
2026-06-10 19:30:28 <oak         > and lofty :)
2026-06-10 19:30:53 <lofty       > ^^;;
2026-06-10 19:31:52 <xentrac     > I had previously done something like : .n s>d <# #s #> type ; as lofty was suggesting and then I found .r
2026-06-10 19:32:58 <oak         > Got it. Earlier this morning I ended up doing it by hand
2026-06-10 19:33:00 <oak         > : .num ( n -- ) 10 /mod dup 0= if drop else recurse then [char] 0 + emit ;
2026-06-10 19:33:15 <oak         > 0 .r is great
2026-06-10 21:13:08 <xentrac     > I think maybe someone here on this channel asked me, "Why aren't you using .r?"
2026-06-10 21:13:18 <xentrac     > and that was how I found out
2026-06-10 21:47:00 <veltas      > I don't know xentrac but you said on 2024-08-19 "I just learned that I can do 0 .r instead of doing 0 <# #s #> type to output a number without trailing spaces"
2026-06-10 21:47:45 <veltas      > I quickly checked the logs to take credit for it, but looks like I can only falsely claim credit
2026-06-10 21:47:48 <veltas      > But I may do so
2026-06-10 21:49:40 <veltas      > It looks like you were checking out a few different Forths?  Maybe it was in some examples for one of them?
2026-06-10 21:56:00 <veltas      > What I don't get about crc's IRC logs is how they go back further than crc was in the channel
2026-06-10 22:10:38 <lofty       > um, another question: what's the difference between a "name token" and an "execution token"? I'm kinda struggling to find where these are defined in the standard.
2026-06-10 22:16:13 <veltas      > Name token will be in practice the address of a dictionary entry, starting probably at back link and near the name of the entry, followed by data/code
2026-06-10 22:16:38 <veltas      > Execution token will be address of the actual code, or the code field, depending on your threading model
2026-06-10 22:17:21 <lofty       > is a name token more flexible than an execution table because it assumes less about the physical layout of the dictionary?
2026-06-10 22:18:09 <veltas      > The point of the name token in the standard is to allow you to traverse the dictionary in a portable way, yeah
2026-06-10 22:18:54 <veltas      > Execution tokens are what you want most of the time, because those are used to efficiently execute a word
2026-06-10 22:19:35 <lofty       > I ask because e.g. https://forth-standard.org/proposals/forth-standards-backward-compatibility-and-modern-state-of-the-art-words-with-their-historical-traditional-counterparts-#contribution-424 talks about how FIND-NAME is better than FIND because it returns a name token.
2026-06-10 22:20:02 <veltas      > Depends what you're trying to do
2026-06-10 22:20:23 <lofty       > mostly "learn what is considered idiomatic forth" ^^;;
2026-06-10 22:20:49 <veltas      > The standard's not very idiomatic
2026-06-10 22:21:03 <veltas      > Believe it or not most forthers aren't using the latest standard drafts
2026-06-10 22:21:12 <veltas      > I think FIND-NAME especially is quite new
2026-06-10 22:21:31 <veltas      > I am pretty sure it didn't exist in the 2012 release of the Forth 200x standard(?)
2026-06-10 22:22:35 <veltas      > I don't think there's an idiomatic way to get 'name tokens' or traverse dictionary entries
2026-06-10 22:22:51 <lofty       > https://forth-standard.org/proposals/find-name#reply-174 <-- yeah, it's from 2018. fair enough.
2026-06-10 22:23:11 <veltas      > FIND is more reliable but even that has been implemented differently at times
2026-06-10 22:23:48 <veltas      > But at least since the ANS standard FIND has returned an execution token and whether the word's immediate, so it's pretty useful
2026-06-10 22:27:17 <crc         > veltas: my logs include some from my pre-IRC days, which I got from one of the guys on tunes.org who ran a logger back then. I maintained a mirror of his logs until he shut his bot down during the freenode/libera split.
2026-06-10 22:34:53 <veltas      > I was just trying to find a wayback machine of Mark's IsForth page, that directed you to the channel
2026-06-10 22:35:05 <veltas      > Did you find that on slashdot or something?
2026-06-10 22:35:27 <veltas      > Well maybe too long ago to remember, was 2003
2026-06-10 22:36:11 <veltas      > https://web.archive.org/web/20030528070111/http://isforth.clss.net:80/
2026-06-10 22:37:55 <veltas      > I was around 10 years old when you joined this channel, makes me feel young
2026-06-10 22:41:26 <veltas      > Well the freenode channel
2026-06-10 22:42:23 <veltas      > The previous version of that site on archive.org called the IRC network "linpeople.org", that's how long ago this was