<dexen>
does a bare coma has any meaning between list elements?
<dexen>
i've placed some here and there seemingly out of habit, and it seems to not affect the program at all...?
<Regenaxer>
it is a read macro, which places the next item into the *Uni global
<Regenaxer>
This is normally used for strings to be localized
<Regenaxer>
Look up "read macros" in the ref, and the *Uni global
<Regenaxer>
So yes, there is no visible effect :)
<Regenaxer>
: *Uni
<Regenaxer>
will show something though
<dexen>
great
<dexen>
starting with PicoLisp is a surreal experience in that some of expected functionality isn't really there, or is handled in unexpected ways. like binary file read/write
<dexen>
tho that's not a bad thing, as per the epigram #19 "A language that doesn't affect the way you think about programming, is not worth knowing."
<Regenaxer>
I see :)
<aw->
hi dexen, welcome
<dexen>
thank you, thank you
rob_w_ has quit [Remote host closed the connection]
<aw->
i send a reply through GitHub as with the previous pull request
<Regenaxer>
Hi dexen, tankf33der, aw-!
<Regenaxer>
In fact, I don't know rlook well, I did not write it
<Regenaxer>
I think it was Jon Kleiser
<Regenaxer>
I can gladly take the patch
<Regenaxer>
What does it fix?
<aw->
i'll try it now, i think it fixes the search issue in new browsers
<Regenaxer>
tankf33der, I tried to find out why pil21 is so much faster on divisio
<Regenaxer>
n
<Regenaxer>
but cannot see a simple reason
<tankf33der>
Regenaxer: ok
<Regenaxer>
Mysterious ;)
<aw->
nope doesnt fix
<aw->
Regenaxer: im not sure what that patch fixes, dexen would need to explain
<Regenaxer>
ok
<Regenaxer>
The focus at the end is commented
<dexen>
for UX, i want the docs to give focus to the search <input>. sadly the "autofocus" attribute doesn't seem to work with frames
<Regenaxer>
Besides that, only the order of the attributes in the text field are changed?
<dexen>
so instead i added a short JS that focuses the search <input>. the comments are to handle older browsers that don't like JS/CSS. it's similar to how previous block of JS and CSS in the same file is commented out
<Regenaxer>
but isn't "IIwindow.document.getElementById('search').focus..." commented out?
<dexen>
it's not. there are 3 lines inside the new <script> tag. first line is a HTML comment (for old browsers), second line is the actual JS code without comment marks; 3rd line is again HTML comment for old browsers.
<dexen>
tbh i wouldn't do the HTML comments on my own, but that's how it is done in other documentation code in those files
<Regenaxer>
I see
<Regenaxer>
So for me anything is fine
<Regenaxer>
(I never read docs in a JS-enabled browser anyway)
<Regenaxer>
If nobody objects, I take your version, ok?
<dexen>
later on i'll probably also skip the HTML/JS documentation and plug something directly into my IDE (Acme) but for starting, the HTML/JS docs are pretty good
<Regenaxer>
What can I write into doc/ChangeLog?
<Regenaxer>
A short note
<Regenaxer>
and credits to?
<Regenaxer>
"... by dexen"?
<aw->
it changes nothing
<aw->
still doesn't auto-focus for me
<dexen>
aw, what is your environment? i've only tested it on the (latest) Firefox and Chromium
<aw->
linux suckless 'surf' browser
<dexen>
alright let me install and check it out
<dexen>
version surf-2.0 hopefully? or an older one?
<dexen>
can't get to it now, as i'm having trouble compiling dependencies for the browser. sorry
<Regenaxer>
No problem!
<Regenaxer>
Shall I postpone the change for now?
<dexen>
recommend "Reference lookup search input autofocus", and "by dexen" is fine
<Regenaxer>
All right! Good
<Regenaxer>
Will be in the next release. Thanks!
<dexen>
great thank you
rob_w has quit [Quit: Leaving]
<dexen>
the (port) function seems to work only with IPv6...? or am i reading it wrong.
<dexen>
asking because it bombs for me with "IP socket error: Address family not supported by protocol"
dexen has quit [Read error: Connection reset by peer]
dexen_ has joined #picolisp
dexen_ is now known as dexen
<Regenaxer>
(port) does work fine with IPv6, but it needs a dual stack at runtime
<Regenaxer>
oop
<Regenaxer>
s
<Regenaxer>
... works fine with IPv4 I mean
<Regenaxer>
so, the OS must support it
<dexen>
alright, makes sense
<dexen>
is there a way to modify the stdin/stdout file descriptors - like the unix dup2() function - so they are attached to different socket for the subsequent (exec ...)?
<Regenaxer>
Not without 'native' calls I think
<Regenaxer>
'pipe' does something like that internally
<Regenaxer>
You mean an (exec ..) after a (fork)?
<dexen>
yes. trying for a very basic functionality; a TCP listener that forks off worker processes and assigns them the connected socket as the stdin/stdout
<Regenaxer>
I think this can be done with pipe plus exec
<Regenaxer>
I find here stuff like: (pipe (exec 'java "-cp"
<dexen>
i'm getting there, but still having trouble properly propagating the eof state up & down the chain
<Regenaxer>
Just close the file descriptor/socket/pipe etc. The Lisp side receives NIL then
<dexen>
i'm probably doing something very silly; i wanted to indicate to the child process that there's no more content to *read*, and to switch to *writing* the response instead
<dexen>
but using (out Pipe (eof T)) seemingly made the pipe closed in *both* directions
<Regenaxer>
hmm (eof) is a pure input function
<Regenaxer>
Closing the file descriptor here is not a good idea
<Regenaxer>
Better send some dedicated message
<Regenaxer>
Anyway (eof T) does not close anything
<Regenaxer>
it sets the *input* status to indicate EOF upon ext read. Very seldom needed
<Regenaxer>
What kind of child process is that? Is it under your control?
<dexen>
yes tho i would like not to change it too much. the idea is to learn picolisp better, so i'd rather understand what i'm doing wrong :D
<dexen>
(rd) without count reads "one item", what does that mean? and is there a helper to convert the large integer into smaller integers suitable for (wr) ?