You didn't misunderstand; it did crash for me on the load & that's what
I reported, but the detail was unimportant.
Henry Rich
On 2/2/2021 8:31 AM, 'Michael Day' via Beta wrote:
Sorry, Henry, I must have misunderstood, yet again!
Fine, if that's what you expect. However, you had said:
"..... if you just typed
load 'misspelledname'
J would crash. "
So I tried it, expecting J to crash as I thought you'd said it
would. So when it
didn't crash, I was surprised. I didn't realise you meant deferred
gratification.
Also, apologies for JVERSION - I'd copied that from the other, J902,
session, and
meant to change it. The crash did indeed occur, eventually, happen
for J903.
Cheers,
Mike
On 02/02/2021 13:15, Henry Rich wrote:
I'm not quite following it all. If you are surprised that the 'load'
didn't crash but something later did, that's OK, the load corrupted
memory for the later code.
Henry Rich
On 2/2/2021 7:18 AM, 'Michael Day' via Beta wrote:
That's odd! (re Henry's msg below)
In JQt 903, I've just tried
load 'misspelledname'
It fails, of course, but not disgracefully:
Yer'tis-
JVERSION
Engine: j902/j64avx2/windows
Release-a: commercial/2020-12-05T13:36:01
Library: 9.02.08
Qt IDE: 1.9.1/5.12.10(5.12.10)
Platform: Win 64
Installer: J902 install
InstallPath: c:/d/j902
Contact: www.jsoftware.com
load 'misspelledname'
not found: c:/users/mike_/misspelledname
|file name error: script
| 0!:0 y[4!:55<'y'
BUT - I looked for the loop in load, found it loops for "fls"
which is defined earlier by
fls=. getscripts_j_ y
so I tried
getscripts_j_ 'misspelledname'
The session hangs for a couple of seconds, then disappears
altogether.
In J902, I get the rather more friendly
getscripts_j_ 'misspelledname'
┌─────────────────────────────┐
│c:/users/mike_/misspelledname│
└─────────────────────────────┘
The first time I tried this in J903, the session was not clean.
However,
I restarted JQt 903 in order to be able to populate this message
with copies
from the session, so I can't blame detritus lying around from
whatever I'd
been doing earlier, except possibly for startup.ijs .
Here are the relevant bits of startup.ijs. There's other stuff, but
only verbs which
I like to have available, not run as startup.
load'~user/affinity.ijs'
load'~user/olddefs.ijs'
NB. ". :: 1: '9!:5 (1)' NB.enable nameref caching! NB'd for now
setproc_affin_ 2?4 NB. choose 2 out of 4 core
whereDefined=: 3 : '(4!:4{.;:y) {:: (4!:3''''),<''Source of
definition not found for '',''.'',~y'
echo 'CHECKING OLD DEFINITIONS'
allolddefs 1
Sorry for a lengthy message,
Mike
On 01/02/2021 13:59, Henry Rich wrote:
This is funny... The bug afflicted for_xyz. loops over boxed
iterands where the loop exited because of an error. The (load)
verb has such a loop, so if you just typed
load 'misspelledname'
J would crash.
I guess that Mike and Bob did something like that & got a crash;
then, good citizens that they are, they tried to clean it up to a
simple failure using a previous version. Of course, that made the
problem go away.
Fixed for next beta.
hhr
On 1/31/2021 5:59 AM, 'Michael Day' via Programming wrote:
Henry asked me privately to provide a copy of the offending
script, however long-winded.
Well... I deleted the J903-beta installation, then copied up from
the zip file, and then ran
the jreg, updateje and updatejqt cmd files, all with admin
privileges. (Windows 10 as before)
Despite Henry saying he'd take a script however long (and dirty by
implication), I removed
excess baggage, and cleaned it up a bit, all in J902, creating a
new script, not replacing the
old. It still ran under J902. I copied it to the new J903-beta
user folder, to sit there alongside
the old script for which J903 had crashed yesterday.
Computing Lore/Law rules: the new script runs cleanly. The OLD
script runs cleanly!!!
This suggests that something had gone wrong with the update from
beta-b to beta-c.
I'll still pass the scripts (old and new) and data to Henry in
case he wants to plumb their depths.
Cheers,
Mike
On 30/01/2021 15:10, 'Michael Day' via Programming wrote:
Thanks.
(NB replying to Programming - would otherwise be Beta)
Unfortunately, for a reason I can't ascertain, JQt 903 beta-c
crashes when I load
my home-made UK Covid data checker. Nothing revolutionary, but I
like to download
the daily updates and see how we're doing.
So I'd like to revert to beta-a - (I don't see how to debug my
script if I can't see what's
happening before it disappears!)
In Windows 10.... I've got the zip file for beta-a, have tried
copying to my J903-beta
folder but don't then have JQT. I tried updateje.cmd which took
me back to beta-c !
Please remind me how to revert to beta version a. Otherwise I'll
have to stick to
J902 . (Might be a good thing and stop me pestering Henry with
misreadings!)
Thanks again,
Mike
On 29/01/2021 19:31, Eric Iverson wrote:
J903-beta-c is available for windows/linux/macos (fixes bug in
the beta-b).
If you already run J903-beta, then upgrade is easy:
load'pacman'
'upgrade'jpkg'' NB. ensure base library and addons are current
'upgrade'jpkg'jengine'
If you have not yet installed J903-beta, time to get started!
https://code.jsoftware.com/wiki/System/Installation/J903
Release notes:
https://code.jsoftware.com/wiki/System/ReleaseNotes/J903
----------------------------------------------------------------------
For information about J forums see
http://www.jsoftware.com/forums.htm
--
This email has been checked for viruses by AVG.
https://www.avg.com
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
