@Sergey: yes, debug errors, and all crashes, are worth sending.

I can't make sense of this trace.  foldx processes the fold verb. unquote resolves a name into a verb/adv/conj for execution; dbunquote executes for debug.  That makes sense.  But the next thing should have been a call to xdefn to  run getJsonEl, which would call the parser again, hit the error, and suspend; but none of that is there.

I will need help from someone who can build from the source on Linux, and can spend the time to shoot this.  That person might be from Jsoftware or from outside.  Volunteers requested!

hhr

On 1/21/2020 11:56 AM, Raul Miller wrote:
Your suspicion is correct: I get an index error with debug turned off.

Here's a partial stack trace

0   libj.dylib                    0x0000000115ca1cde jtdebug + 110
1   libj.dylib                    0x0000000115ca228d jtdbunquote + 253
2   libj.dylib                    0x0000000115cd0951 jtunquote + 1153
3   libj.dylib                    0x0000000115c47e4c jtfoldx + 188
4   libj.dylib                    0x0000000115cc3417 jtparsea + 2807
5   libj.dylib                    0x0000000115ca2172 jtparsex + 162
6   libj.dylib                    0x0000000115c9cd93 jtxdefn + 6707
7   libj.dylib                    0x0000000115c615b7 jtdfs2 + 39
8   libj.dylib                    0x0000000115ca220f jtdbunquote + 127
9   libj.dylib                    0x0000000115cd0951 jtunquote + 1153
...

Thanks,


----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to