than the popc.txt file.)
>
> Tips on rebuilding a current poplog are relevant.
>
> http://www.cs.bham.ac.uk/research/projects/poplog/sysdoc/rebuilding.txt
>
>
>
> It also occurs to me that finding out what Waldek Hebisch had to do when he
> ported
d as
>
> v15.6301-amd64-old.tar.bz2
>
> available here
> i.e.
> http://www.cs.bham.ac.uk/research/projects/poplog/amd64/v15.6301-amd64-old.tar.bz2
>
> also available on Waldek's web site:
> http://www.math.uni.wroc.pl/~hebisch/poplog/
> with his instruc
ng "port" is probably too much -- my work was essentially to
cross-compile existing sources.
BTW: I have just noticed that diffs on that page are a bit misleading,
they are "revesed diffs". Normal conventin is to have original
first and modified version second, but I did that in opposote
order.
--
Waldek Hebisch
e same way, but I did not have
time to work out detalis). I knew how to do this in assembly,
but clearly when efficiency is the same Poplog version is
better.
--
Waldek Hebisch
So you
need running Poplog to compile from sources. Current Poplog
distribution essentially assumes that you already have almost
full Poplog environment and want to compile new version (say
after small modification). Script at:
http://www.math.uni.wroc.pl/~hebisch/poplog/build_pop
assu
e moved
compared to what script expects or build can not find system files).
At
http://www.math.uni.wroc.pl/~hebisch/poplog/build_pop2
there is script that I used today to rebuild 'newpop11' on
Debian 8. I needed the following patch to avoid problems
in the new binary:
http://ww
OEXEC) = -1 EBADF (Bad file descriptor)
while on Debian 8 the second line is missing. 'sysopen' reads 'errno'
and different value of 'errno' causes mishap. ATM I do not know
why 'fcntl' is called.
--
Waldek Hebisch
src/syscomp/asmout.p' is doing this).
Anyway, after applying patch I was able to compile Poplog
on buster.
Patch in attachement.
--
Waldek Hebisch
diff -ru pop.bb/src/declare.ph pop/src/declare.ph
--- pop.bb/src/declare.ph 2016-12-31 02:09:30.237527016 +
+++ po
can be done in many ways,
the above works and is resonably efficient. Probably would
be nicer in hexadecimal notation. One could macro name
numbers. OTOH the need to sign-extend is ugly and ugliness
of code above may serve as warning to reader.
BTW: For permananent use definitely deserves a comment like:
;;; Sign extend from 32-bits to 64-bits
--
Waldek Hebisch
I have upload to github:
https://github.com/hebisch/poplog
cleaned up version of sources of 64-bit Poplog. ATM it is only
core part (what is needed to build corepop). I have removed
from sources ugly hack and replaced it by systematlically
marking external calls with SE, like:
_extern[SE
>
> Good! I'll try to login to https://github.com/hebisch/poplog
> and provide some user-test-feedback.
> == crg.
Note that ATM to compile it you need to add several files
from other Poplog sources and adjust build script.
In other words, currently this is mainly of interest
should do _less_ testing
when opening files. Namely, non-existing 'src.wlb' is normal during
build, but Poplog issues MISHAP because one of extra checks fails.
Similar things happens during popc compilations: popc tries to
autoload things and issues MISHAP when local autoload directory does
not exist. IMO having no local autoload directory is normal and
popc should just try to open file and treat failues the same way:
it file can not be opened then it is like non-existing file.
--
Waldek Hebisch
d symbolic
links. Copies too easily lead to diverging versions and confusion
when wrong copy is in use. Symbolic links have similar problems
and additinally may be easily broken.
Concering what I wrote about autoloading: I can manage autoload
by setting 'popautolib'. But 'poplocalauto' normally should be
unused. The trouble is that if it does not exist, then I get
mishaps. OTOH when it points to empty directory things work.
I am not sure in Poplog always behaved that way, or it is due
to same recent change (I remeber that popc acted in weird ways
on error so it may be old).
--
Waldek Hebisch
if it hits untested routine,
examine its arguments etc.
At:
http://www.math.uni.wroc.pl/~hebisch/poplog/poplog_arm.tar.gz
you can find tarball of my developement tree on Banana Pi.
It can rebuild itself. Languages (Pop11, Lisp, ML, Prolog)
and textual ved seem to work. ATM X and extra packages ar
ge. And IMO similar
change should be done for other architectures.
--
Waldek Hebisch
es, so
there are few conditionals for this.
The point is: interfacing code has to follow rules set by
operating system. Those rules are essentially arbitary
and unpedictible, Poplog simply has to adapt.
--
Waldek Hebisch
ing factor is time,
I have various commitments and plans... It is tempting to
delegate some work for students, and small support money could
help with this, but ATM I have no suitable student (bright ones
are interested in different things).
Another thing is that once port is stable it makes sense to do
some advertising. But IMO it is too early for this...
--
Waldek Hebisch
I have put new version at
http://www.math.uni.wroc.pl/~hebisch/poplog/poplog_arm.bz2
This version can do external calls and xved is able to show
edit window. There is still few unimplemented things, but
in my testing I was not able to trigger them.
The tarball contains 'pop' dire
Aaron Sloman wrote:
>
> If windows linux does not support 32 libraries only 64 bit poplog can be
> used. I am somewhat out of touch but I believe the latest 64 bit poplog is
> available only from Waldek's site:
>
> http://www.math.uni.wroc.pl/~hebisch/poplog/
My curre
Aaron Sloman wrote:
>
> Waldek,
>
> Thanks for your rapid and helpful response. Your contribution sounds
> extremely useful.
>
> > My current version is at github:
> >
> > https://github.com/hebisch/poplog
>
> I am not a github user but I would definit
t 'pop/src/mem_rapp'.
It contains raport of all system calls made by corepop, and
probably there is some difference in calls or results related
to the crash.
--
Waldek Hebisch
nable commit messages git gives
you a more information about changes than typical change log.
--
Waldek Hebisch
0x7a2000
...
brk(0x7a5000) = 0x7a5000
mprotect(0x7a4000, 4096, PROT_NONE) = 0
mprotect(0x7a4000, 4096, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
Google says that EACCES is due to SELinux (SELinux forbids writing
and executing the same memory and Poplog wants to do this).
--
Waldek Hebisch
t EACCES is due to SELinux (SELinux forbids writing
> and executing the same memory and Poplog wants to do this).
>
Somewhat longish story about the same problem (and some
solutions/workarounds):
https://nullprogram.com/blog/2018/11/15/
--
Waldek Hebisch
the
> output.p window.
>
> I presume it's not finding the 64 bit X11 libraries for some reason.
> I don't know how to ensure that poplog gets the right libraries.
My guess ATM is that Poplog is getting right libraries due to
action of system dynamic linker. But later Poplog
is trying to load libraries again and this time gets it wrong.
On Debian it finds no library and probably ignores the wrong
load. On Fedora it seems that finding 32-bit library unloads
or at least messes with 64-bit one.
--
Waldek Hebisch
>
> MACHINE 2: This got much further but there's a library problem for X
> (probably easy to fix?)
>
> a shared server running
>
> 2.6.32-754.14.2.el6.x86_64Scientific Linux release 6.10 (Carbon)
>
> This produced a partly usable basepop11 a
und a Linux kernel bug which silently changes the
> > PARENB/CREAD/CSIZE bits in c_cflag on pty.
I do not think there is any relation. Various distributions use
the same kernel and IME kernel is pretty reliable and compatible
across versions.
> It's a bit _too_ coincidental f
why Poplog wants to specify directories where
the libraries live. The attached patch just uses empty
string as a defalut path for libraries. Hopefully this
will be portable.
To apply do:
patch -p 0 < lib2.diff
in 'poplog_base' directory.
Note: default patch should be last o
c/newpop -link -x=-xt -norsv
at the endo of 'build_pop2' to
$usepop/pop/src/newpop -link -x=-xm -norsv
I have neither Motif nor Lesstif so I can not test if this works,
but in principle all lower-level machinery for this to work is
in place.
--
Waldek Hebisch
xved window reliably
> causes a segfault. ISTR there was an emacs package :)
That is because poplog_base did not contain 'pop/x/ui' subdirectory.
I had to setup new machine (Debian 9), so in the process I installed
Motif (apt-get install libmotif-dev). After adding 'pop/x/u
Aaron Sloman wrote:
>
> Waldek wrote:
>
> > I have put package with ui added and patch to shared library path
> > at:
> >
> > http://www.math.uni.wroc.pl/~hebisch/poplog_base2.tar.bz2
>
> Thanks. I am about to try it. However: it should be
>
&g
gards,
I do not use Windows, but I have been told that now sbcl works
quite well on WSL. Poplog is simpler, so in principle there
should be less problems.
--
Waldek Hebisch
results
of half of expected size. Apply patch below:
diff -ru poplog_base/pop/src/x86_64/aarith.s trunk/pop/src/x86_64/aarith.s
--- poplog_base/pop/src/x86_64/aarith.s 2019-06-30 02:11:26.602946205 +0200
+++ trunk/pop/src/x86_64/aarith.s 2019-07-06 21:57:40.184992189 +0200
@@ -421,6 +421,7 @@
DEF_C_LAB (_posword_mul_high)
movl(%USP), %eax
movl8(%USP), %edx
+shll$1, %eax
addq$8, %USP
imull %edx
movq%rdx, (%USP)
Or fetch:
http://www.math.uni.wroc.pl/~hebisch/poplog/poplog_base4.tar.bz2
--
Waldek Hebisch
- 1;
--- packages/popvision/lib/newbytearray.p 2003-07-15 08:36:48.0
+
+++ packages/popvision/lib/newbytearray.p 2019-07-06 23:22:05.831409167
+
@@ -12,7 +12,7 @@
uses popvision
uses oldarray
-defclass bytevec :byte;
+;;; defclass bytevec :byte;
define newbytearray = newanyarray(% bytevec_key %); enddefine;
--
Waldek Hebisch
not previously
> tried resizing an xved window.
>
> This on Fedora 28, kernel 5.0.16-100.fc28.x86_64
On Debian (8 and 9) with LXDE as window manager, when I point
at window boindary, press left button and drag the XVed window
changes size with no bad effects.
--
Waldek Hebisch
ssemble
say. If there is problem with alignment, it should be
clear given info above. For other problems we shoul at least
get some clues.
--
Waldek Hebisch
eis I'm using Windows 1803, just built FriCAS with
SBCL-1.4.5 (bundled by Ubuntu 18.04) and SBCL-1.4.7 without problems.
I think this SBCL problem is fixed.
--
Waldek Hebisch
ried resizing
> an xved window on the latest poplog are Waldek on a debian machine
> and myself on fedora 25 (works), fedora 27(doesn't work) fedora 28(doesn't)
> SL6.10 (works) CentOS 7.6 (works).
Graham reported crash.
--
Waldek Hebisch
db prompt type
info registers
and
disassemble
Then post the results.
--
Waldek Hebisch
Thanks to debug info from Aaron I have found problem
with stack alignment and now have a patch that should
fix resize problem. Apply patch below or fetch
http://www.math.uni.wroc.pl/~hebisch/poplog/poplog_base5.tar.bz2
(the only difference from version 4 is that it has patch
already applied
omewhat hacky, I wonder
if there is better solution.
Anyway, at
http://www.math.uni.wroc.pl/~hebisch/poplog/poplog_base6.tar.bz2
there is tarball containing new corepop. I also slightly
modified 'pglink' so that it does not build 'rsvpop11'
when '-core' is given. Th
le
> both here:
> http://www.cs.bham.ac.uk/research/projects/poplog/amd64/poplog16
>
> and on his web site:
> http://www.math.uni.wroc.pl/~hebisch/poplog/
>
> identified as 'tarball', i.e.
> http://www.math.uni.wroc.pl/~hebisch/poplog/poplog_
a new one.
There may be some other problem, but installing 32-bit suff for
64-bit Poplog is a WRONG thing, at best it masks real problem.
Sorry if what I wrote sounds harsh, but unforunately wrong
information spreads out fast and it is important to rebuke it
quickly.
--
Waldek Hebisch
k with files and directories
and create distribution tarball that way, that is fine too.
If you want me to put documentation on Github and forward changes
there I can do this. That way documentation would be automatically
included in any future "base" tarball. If documentation is not on
Github, it would be somewhat awkward for me to add it each time when
I create new "base" tarball.
--
Waldek Hebisch
#x27;nopie.diff' patch.
'nopie.diff' is included in 'poplog_base...' tarball that
I provide. Just in case I also attach it to this message.
OTOH if you have this line in 'pop/pop/poplink_cmnd', then
we have deeper problem. I such case we should ask how
to create n
imply uses 'exec' system call
and all business with calling interpreter (that is the shell)
is done by kernel.
And './bar.sh' is full relative pathname, without relying
on PATH (which normally does not contain current directory).
--
Waldek Hebisch
000.0]
>
> ;;; MISHAP - FLOATING-POINT OVERFLOW
> ;;; INVOLVING: 2.0 1024
> ;;; FILE : /home/hakank/poplog/me/test_sqrt.p LINE NUMBER: 19
> ;;; PRINT DOING
> ;;; DOING: ** trycompile
> """
AFAICS that is OK, already 2**1024 is too big
16 if readline if it is invoked AFTER the pop11
> prompt has been printed (i.e. after Setpop, not on the linux command line):
>
> % pop11
>
> Sussex Poplog (Version 16.0 Thu 23 Jan 20:46:14 GMT 2020)
> Copyright (c) 1982-1999 University of Sussex. All rights reserved.
>
> Setpop
> : readline()=>
> ? this should work
> ** [this should work]
Yes, now proglist comes from parsing user input from terminal.
> POPRACER EVIDENCE
I had no time to look at popracer, but problem must be a bit more
subtle.
--
Waldek Hebisch
There is a change: Poplog used to
roll its own emulation of termcap, which did not keep up
with changing environment. AFAIK now it uses system library.
Apparently others had less motivation for testing other bits
were not enabled.
--
Waldek Hebisch
g this will not have V15.65. If anone wants to
> get involved it can be obtained here:
>
> http://www.cs.bham.ac.uk/research/projects/poplog/latest-poplog
>
> using the pre-linked poplog (because linking V15 basepop11 no longer works
> in recent versions of linux).
There is an extra problem: V15.65 does _not_ work on old 32-bit
systems (the only 32-bit ones that I have). It needs post 2007
(or maybe 2008) ld.so, otherwise it crashes. It well-known
incompatibilty: any program linked without
--hash-style=both
option to linker will crash on older systems.
--
Waldek Hebisch
volving saved images).
--
Waldek Hebisch
hat change are you proposing. In V16 USE_TERMINFO is defined
to true in 'pop/src/unixdefs.ph' (the definition lives now in
this file). So what exactly you think should be changed?
--
Waldek Hebisch
say that basic design principle here is that
any thing which happens, happens as response to events and normally
user program, in this case Poplog can change even handling in
almost arbitrary ways. So, problem should be resolvable
entirely within Poplog. And solution is likely to be quite
simple, we just need to find it...
--
Waldek Hebisch
ourne shell scripts.
AFAICS current system is inefficient, inflexible and hard to
maintain. Concerning inflexible: there are many points
which can _theoretically_ be used to change Poplog behaviour.
In practice choosing wrong point leads to more or less
wrong behaviour. Scattered and hidden dependencies make
it hard to find correct change point.
--
Waldek Hebisch
s into repository.
Up to now I did not find time to do serious work on items above,
so if you folks think that it is more important to ship something
now than to clean up things I will not object. Still, I think
that is better to skip a package in case when nobody can tell
if this package is working.
--
Waldek Hebisch
ether.
>
> I've had a quick glance at what the bham install script produces, and it's
> a little different to what's in your repo, so now I don't know which source
> to cohere around. The download script references your work which is great,
> but the readme at ht
g on 'delete' button to killing Poplog _all_ happens
in Motif code. For me, most likely problem place is code that
created window. Namely, normaly various possible behaviours
are chosen at window creation time. Unfortunaly, window creation
seem to be spread out over rather large area of code...
--
Waldek Hebisch
tup image, then they should be added
to 'pop/lib/lib/startup.p'. This requires a bit of work,
as things depending on X should go into proper conditional.
BTW: 64-bit V15 had a bug in memory management and putting
stuff in 'mkstartup' probably was a workaround. But no
such workaround should be needed in V16.
--
Waldek Hebisch
.
But for some reason most of the time handler is not
called...
--
Waldek Hebisch
asked about '$usepop/pop/teach/teachfiles'). You wrote
that file is where it should be. In such case it remains
a mystery why ved can not open it via 'teach'. Does it
work if you try to open copy of 'usepop/pop/teach/teach'
in ved?
--
Waldek Hebisch
n other
folks. BTW: '-no-pie' is also problem on ARM (and presumably
on 32-bit Intel processors). I think that simple solution
is possible: main build script (or 'configure' before build)
should try compiling a test program giving '-no-pie' to
compiler. If that works
port usepop
to the above.
Try in pop11
systranslate('$usepop') =>
If usepop is set correctly you will see value that you have
in your .bashrc
--
Waldek Hebisch
ght (c) 1982-1999 University of Sussex. All rights reserved.
> >
> > Setpop
> > : help rclib
> > Unknown terminal type: xterm-256color
>
> That suggests that your value of $TERM at the time when you start poplog is
> xterm-256-color. That's not recognized by p
AFAICS the sh version was missing:
usepop=`pwd`
export usepop
at the start. To allow real sh (and not only bash), instead of
source $usepop/pop/com/poplog.sh
it needs
. $usepop/pop/com/poplog.sh
Command for unpacking packages-V16.tar.bz2 needed fix, correct
version is
tar xfj ../../package
On Thu, Aug 20, 2020 at 03:05:21AM +0100, Aaron Sloman wrote:
> Thanks *very* much Waldek.
>
> I have fetched this
>
> http://www.math.uni.wroc.pl/~hebisch/poplog/build_all.sh
>
> and installed it here:
>
> http://www.cs.bham.ac.uk/research/projects/poplog/
ck of what is tested (however shallow the
test) and what got no testing.
OK, tell me what you think about issues above.
--
Waldek Hebisch
gt; experience of designing editors with autosave mechanisms could come up with
> a clean new specification.
AFAICS the change _directly_ should not help at all. So most
likely this changes code flow in such a way that avoids
executing failing code. Without way to reproduce problem
there is no way to find out what is the real reason.
But is is reasonable to expect that real problem is much
more general than saving files.
--
Waldek Hebisch
/news.v14_1 pop/help/xnews.v13_9
Those are all old news files for obsolete Poplog versions.
--
Waldek Hebisch
I have put at:
http://www.math.uni.wroc.pl/~hebisch/poplog/poplog_base7.tar.bz2
new poplog_base7.tar.bz2. It has all recent fixes applied to
github repo. AFAICS it should work with Aaron's recent
'getpoplog.sh'. Changes:
- fixed arrayfile and passing and returning single
ents could give simple
commands generating line drawings. By delaying and
rewriting they could do simple animations. Technicaly
the code was quite crude, but it was fit for the
purpose.
--
Waldek Hebisch
up (with similar
> comments for the other languages) that should suffice.
>
> At present this is handled in the directory
>
> $usepop/pop/pop
>
> which includes, besides basepop11 (hard) links to it with different names,
> e.g.
> pop11, clisp, pml, prolog, ved, xved
>
> which are all hard links to basepop11.
All what changed is that there are less links. More precisly
newpop no longer creates links to doc, help, im, ref and teach.
OTOH 'teach teach' in pop11 or ved works as before.
I changed 'pop/src/newpop_options', changing it back to old
version would give back the links.
--
Waldek Hebisch
In fact, searching
code for string 'ku' or 'kd' gives me nothing...
So it seems that we need to write appropriate 'install_escape'
(it needs to handle multicharacter sequences, so may be a bit
more complicated than simple assignment) and do assignments
for known keys sequences. ATM I see no code like this...
--
Waldek Hebisch
On Tue, Aug 25, 2020 at 08:21:30AM -0400, os9106...@sneakemail.com wrote:
> On Tue, 25 Aug 2020, hebisch-at-math.uni.wroc.pl |pop-forum| wrote:
>
> ...
> > > I haven't yet understood how vedescapetable gets its values. The file
> > > vedtermcap.p looks rel
On Tue, Aug 25, 2020 at 09:56:19AM -0400, os9106...@sneakemail.com wrote:
> On Tue, 25 Aug 2020, hebisch-at-math.uni.wroc.pl |pop-forum| wrote:
> ...
> > > Isn't the code for vedsetkey in $usepop/pop/ved/src/vdsetkey.p doing the
> > > job
> > > of your in
rary functions. Poplog only uses library calls, which
is OK.
--
Waldek Hebisch
fferent ad-hockery which gives 'pc'
(that is current content of machine type).
Maybe we should introduce new parameter 'sys_machine_architecture'
and give it appropriate value, and remove uses of
'sys_machine_type'?
--
Waldek Hebisch
and
disassemble 0x077B12890-40,0x077B12890+40
to see failing code. Note: 0x077B12890 is program counter
of failing instruction. This may change from run to run,
gdb will tell you program counter of failing instruction.
Program counter is also shown in output of 'info registers',
as 'rip'.
--
Waldek Hebisch
is above code in corepop, so either this is
some wild jump (unlikely) or incrementaly compiled code
- 0x7ab1e0 is round number and PC coincides with address
so there is good chance that this is first instruction of
generated code and error is because memory is marked as
non-executable (info from disassemble and register content
could verify if this is really first instruction)
Not much, but smells like SELINUX...
--
Waldek Hebisch
local functions, it used to generate code on the stack but
apparently now is implemented in different way. What remains
is dynamic linker, Java, various Lisp-s, ML-s etc. Most
probably use use several 'mprotect' calls and copy to
add new code.
--
Waldek Hebisch
(say buch of configuration
changes with some dependent of kernel features).
To put it differently, most guesses about software problems are
wrong. Here there are several factors involved, so chance
of resolving problem without digging deeper is virtually 0.
--
Waldek Hebisch
op11 code,
only to C file). Of course full rebuild will do.
--
Waldek Hebisch
--- ../trunk/pop/extern/lib/c_core.c 2020-02-07 18:23:46.0 +
+++ pop/extern/lib/c_core.c 2021-05-06 23:31:15.365927914 +
@@ -2154,9 +2154,10 @@
|
= 45
execve("/mnt/lv3/kompi0/poplog/pp6.pp3y/pop/pop/pop11", ["pop11"], [/* 71 vars
*/]) = 0
and later
personality(0x /* PER_??? */) = 4456448
The calls with 0x as argument ask for current personality, we
need last value which in hex is 0x44. On my system default is
0x40 (decimal 4194304).
--
Waldek Hebisch
ent process. At least on my machine the opposite seem
to "work": when I clear READ_IMPLIES_EXEC, then I get errors.
--
Waldek Hebisch
--- ../trunk/pop/extern/lib/c_core.c2020-02-07 18:23:46.0 +
+++ pop/extern/lib/c_core.c 2021-05-12 18:04:
build corepop
on oldest supported system: linking on new system tends to
introduce dependence on newer library.
--
Waldek Hebisch
nt Birmingham
tarballs). But '-Wl,--hash-style=both' is likely to cause
trouble when link is done on old system.
Corepop with patch, _not linked to ncurses_ and linked as
above should work on wide range of machine.
--
Waldek Hebisch
sion
files should be fixed by adding apropriate 'uses' (I have some
fixes of this sort in my priviate popvision tree, but it is
probaly far from complete).
--
Waldek Hebisch
/hebisch/poplog .
2. Make directory for corepop:
mkdir -p target/pop
3. Fetch corepop:
wget https://www.math.uni.wroc.pl/~hebisch/poplog/corepop.amd64
4. Turn on executable bits on fetched corepop:
chmod 755 corepop.amd64
5. Move corepop to proper place:
mv corepop.amd64 target/pop
do not want to depend on envirinment.
Namely startup matters for small utility programs that
are run from other programs (which may be unaware of
fact that utility is written in Poplog).
--
Waldek Hebisch
t; system. I think it's pretty common, but maybe not universal.
> The syntax ${0##*/} does the same thing in bash, but I don't know about
> other shells. Maybe test for the presence of basename in configure and only
> insert the lines if it is found.
I would not worry about "basename", it is pretty standard command.
--
Waldek Hebisch
use different corepop:
0. Go to directory where you will build poplog
1. Fetch sources:
git clone --depth=1 https://github.com/hebisch/poplog .
2. Make directory for corepop:
mkdir -p target/pop
3. Fetch corepop:
wget https://www.math.uni.wroc.pl/~hebisch/poplog/corepop.i386
4. Turn on
ed from other
programs or command line. And certainly we do not want a little
utility to suffer long startup time.
--
Waldek Hebisch
ld AIX source tarball:
https://www.cs.bham.ac.uk/research/projects/poplog/new/powaix-15.52.tar.gz
--
Waldek Hebisch
/packages/lib'
looks like? I see there 'vedhelplist', 'vedreflist' and 'vedteachlist'.
--
Waldek Hebisch
name);
compile(file);
valof(name)();
enddefine;
define fun();
loader('ffun.p', "fun");
enddefine;
---
I wonder, are there any traps. Or maybe there is
better way to do this?
--
Waldek Hebisch
s like things randomly fail.
ATM I do not know what the problem are, but there is good
chance that problem with popracer is the same as problem with
rc_ants_demp.p.
--
Waldek Hebisch
ockfile, master and newc_dec from the distribution.
--
Waldek Hebisch
cluded in 'c_dec.p' and
'neural' should use 'oldc' instead of 'c' as language
tag. Both changes are very simple and after that
'teachneural' works OK.
--
Waldek Hebisch
rces build fine on it. With new
sources rc_ants_demo.p has the same problem as on 64-bit machine.
With old 15.61 rc_ants_demo.p works fine. So, it seems that I will
have to do some bisections to find out which change introduced the
problem...
--
Waldek Hebisch
ously correct", but C compiler (using freedom given
to it by C rules) produced different effect than intended.
Now rc_ants_demo.p runs form me. Popracer goes a bit more forward,
now popracer controls actually do something, but simulation crashed
complaining that it needs a number. I got rather long mishap message
involving a number and a list...
--
Waldek Hebisch
ler (using freedom given
> > to it by C rules) produced different effect than intended.
>
> Was that this fix?
> https://github.com/hebisch/poplog/commit/f592a3154b4011bd9a025580fd7a4fc96b87e0de
>
> <https://github.com/hebisch/poplog/commit/f592a3154b4011bd9a
1 - 100 of 149 matches
Mail list logo