I don't have those compilers.  If you can find the errors I'd be obliged.  g7x5 has code to predict the length of an allocated block, and that testcase may be wrong for 32-bit now (I tightened up allocation a bit).

The others seem like compiler issues.

Henry Rich

On 5/1/2020 10:26 AM, Akakima wrote:
Looks like it is fixed. Thanks ...

By the way, i used clang for Cygwin. (clang version 8.0.1)
Only 1 test fails:

g7x5

Then i tried with gcc. (gcc version 9.3)
The folowing tests fails:

g6x
g7x5
g8x
gdll_df

In both cases, gdll does not fails anymore, after a simple modification of gdll.ijs.
----------------------------------------------------------------------
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

Reply via email to