freenode/#lisp - IRC Chatlog
Search
6:10:05
adlai
ACTION occasionally sees a comma instead of colon, this seems like a less formal convention... most IRC clients will highlight the message in either case
6:11:43
adlai
fbmnds: there is also an #ecl channel that might be useful since you are asking about an ECL-specific project
6:13:27
adlai
ACTION mumbles something about how maybe they'd get more excitement about ECL if it implemented the complete ANSI specification
6:13:47
beach
fbmnds: jackdaniel is the current maintainer of ECL, and he fulfills all the criteria. He has a family (new baby too), and he is very busy. But he is often around.
6:14:52
adlai
beach: no, I am only aware of one painful incompliance - the lack of :arguments in the long form of define-method-combination
6:16:13
adlai
I've considered taking this on as a project, since at least one other implementation (Clasp) has inherited that problem, by reusing ECL's code
6:17:18
beach
Indeed. And I was planning to use ECL to bootstrap SICL, so I might be interested as well. Plus, I have this long-term plan to try to create a Cleavir-based compiler for ECL.
6:19:49
adlai
it seems understandable to me that an attempt to write an efficient method dispatch system would skip that one feature, although the result is a nuisance, because it is a very useful feature
6:32:58
fbmnds
phoe: ref. [2020-03-28, 13:46 <phoe> nope, it's non-portable and SBCL-only] osicat runs on ECL too (https://pastebin.com/VSQDwAKu) - just saying
7:16:36
phoe
fbmnds: my answer was to "do other distributions have a sb-posix compatibility layer? ccl?" not to osicat itself
7:17:50
phoe
ECL follows some SBCL interfaces, but the last time I checked (and I might have failed) it doesn't have a layer that follows SB-POSIX
8:40:46
jackdaniel
adlai: lack of this argument is a bug, but saying that "ecl doesn't implement the complete ansi specification" based on one missing argument is quite a stretch
8:41:32
jackdaniel
even sbcl doesn't pass all tests from ansi-test suite (some of disrapencies were a deliberate choice at that!)
9:01:07
fbmnds
phoe: ah ok - I came across this when I looked into osicat as a dependency of cserial-port while porting from SBCL to ECL. I had problems with getting osicat to run until I realised that it runs well if ECL is compiled as default with gcc (had previously used g++)
9:06:13
jackdaniel
fbmnds: eql5 is actively developed and maintained by Paul Ruetz (the project author), I'm not sure what do you mean by "maintaining further" the project
9:07:20
fbmnds
btw, has anyone succeeded in building a reliable permanent connection to a microcontroller with cserial-port? I tend to believe that is practically not possible due to timing issues between send/response sequences. I am looking now to make my own FFI wrapper for a C library.
9:16:28
fbmnds
jackdaniel: I refer to upcoming maintenace challenges in a foreseeable future, given that Paul Ruetz mentioned to me in a private email conversation that he regards EQL5 as a hobby project on which he does not want to spend too much of his private time.
9:30:14
fbmnds
jackdaniel: I should add that I am also refering to the upcoming Qt6 release (https://www.qt.io/blog/2019/08/07/technical-vision-qt-6)
11:20:52
MrtnDk[m]
In Guile scheme, which command do you recommend for reading a line of text (from a file stored on a fixed disk, for instance)?
11:25:08
MrtnDk[m]
(Sorry, I guess I posted it in the wrong Lisp. I just reread the room discription of this one).
11:25:40
rgherdt
you can ask that on #guile or #scheme. But you probably want to take a look at read-line
11:50:36
MrtnDk[m]
<rgherdt "you can ask that on #guile or #s"> I tried readline, but it doesn't seem to work in newer guiles. (use-package (ice-9 readline)) or something.
11:55:10
rgherdt
MrtnDk[m]: don't confuse readline (GNU's tool) with scheme's read-line. The latter can be imported for instance from (scheme base) or rdelim
12:03:18
iissaacc
yo, im working with common lisp and i have a directory of 74000 files i need to process. (directory) is running okay but I'm only getting 50 filenames, anyone know why this could be?
13:08:15
MrtnDk[m]
rgherdt: Ahh! I thought "read-line" was a typo I made, since it generated an error and isn't mentioned in the texinfo documentation of the guile reference!
13:20:52
nabataeus
I'm asking because I'm attempting to emulate one though I don't know where to look.
13:26:42
nabataeus
phoe: Is open Genera an OSS version of the original Genera? Does it differ, or is just relicensed.
13:29:20
p_l
nabataeus: it's an unfinished (but usable in production) system that runs Genera 8.5 on top of Alpha CPU running Digital Unix (POSIX = "Open Systems" at the time)
13:34:42
p_l
nabataeus: yes, it's been pretty popular for a time to use hacked-up versions that run on amd64, the range of bugs can be subtle though
13:35:34
p_l
and not all software will work, as OpenGenera never actually hit planned 1.0, despite 1.0 and 2.0 being sold, and some software expects presence of old console interface (whereas OpenGenera only exposes a newer API that not all applications migrated to)
13:45:36
bitmapper
the x86_64 emulator was made by writing a program that translates alpha pseudo-assembly into C, and then just translating the alpha emulator
13:51:00
nabataeus
But I think their 'Genera Concepts' guide would better explain this. Either ways thanks for your time o/
13:51:38
p_l
nabataeus: VLM (Virtual Lisp Machine) essentially treated Alpha CPU as microcode engine
13:52:19
p_l
nabataeus: and using fine-tuned assembly implemented microcode engine for modified version of Ivory CPU (binaries between Ivory revs 0-4 and rev5 aka VLM are incompatible)
13:53:02
p_l
there's some C code which was taken from MacIvory/UX series code, except instead of talking to expansion board with real Ivory CPU it talks to the microcode engine
13:54:01
p_l
from hw other than CPU, the system presented is equivalent to UX-series symbolics lisp machines
14:52:38
adlai
nabataeus: there is still a teeny tiny market for the actual machines themselves, and their peripherals and replacement parts