If I create an empty character array by selecting ⍬ from an array that
contains a string and a number, the result becomes what seems to be an
empty array with non-zero shape.
The issue can best be explain by the following interaction. Note the
strange output from 8⎕CR:
a←('foo' 9)[⍬]
In the function Tokenizer::scan_real(), there is a call to exp10(). This
function is unique to glibc and does not exist on OSX (and likely not on
FreeBSD either, although I haven't checked).
Changing the call to exp10(x) to pow(10, x) makes it work.
Regards,
Elias
Gentle people,
Please find attached QUADTF.tc.gz. gunzip it and run apl (svn rev 110)
thus
apl --TM 2 -T QUADTF.tc
Examine QUADTF.tc.log and observe that the SCORES variable is not being
reconstructed in the workspace with either the migration or extended
form of the inve
On 2014-01-28 14:23:01, Juergen Sauermann wrote:
> that was on purpose because the default ⎕PW is 80 and on some
> 80 column terminals this causes an extra empty line to be printed.
If there are terminals like that, wouldn't it be easier to just initialize
⎕PW to 79 while keeping relation between
Hello Jrgen,
With regard to item 1:
a) I've rerun the valgrind test against SVN 111 and valgrind still
complains.
b) Before I suggested adding the '[]' to the 'delete', I made the change
and performed the valgrind test. Valgrind did not complain about double
deletions. It is possible t
Gentle people,
[I'm resending this email because I left the `[Bug-apl]' prefix off the
the subject line in the previous email. Also I've added some additional
valgrind detected memory errors.]
valgrind runs with options --leak-check=full
--show-leak-kinds=all --track-origins=yes
Dear Fred,
ad 1)
the delete is actually correct. the Cells to which the pointer points
have beed destructed before
and delete[] would lead to double destruction. I have inserted a
variable long_ravel in Value.cc which
hopefully confuses valgrind enough to not complain anymore.
ad 2)
I have ch
Hi,
thanks. Fixed in SVN 110.
/// Jürgen
On 01/28/2014 08:12 AM, Elias Mårtenson wrote:
Very easy to reproduce this one:
⍉⍬
This gives the following error:
==
Assertion failed: !shape.is_empty()
i
Hi,
that was on purpose because the default ⎕PW is 80 and on some
80 column terminals this causes an extra empty line to be printed.
/// Jürgen
On 01/28/2014 01:18 PM, Kacper Gutowski wrote:
Hi,
After lines are neatly wrapped in r109 in SVN, but I think there's some
off by one error; it beha
Hi,
After lines are neatly wrapped in r109 in SVN, but I think there's some
off by one error; it behaves as if ⎕PW was ⎕PW-1.
⎕PW⍴'X'
XXX
X
-k
10 matches
Mail list logo
