Hello, I just compiled AST-OPEN 2014-12-24 and tried out the new KSH. Platform is "sol8.sun4".
For KSH 2014-12-24 (latest beta?) it core dumps readily and quickly on certain text input in interactive mode. Scenario: a> ksh $ #nn <- literal three characters "#nn" <instant core dump) After typing a '#' character, almost any combination of two characters will cause a core dump. A GDB backtrace on the core file yields: #0 ed_emacsread() #1 slowread() #2 sfrd() In a related matter: The same thing happens on the MacOS-X platform (Darwin 14.3.0 -- the latest stable version). There (on MacOS-X) it occurs on KSH version '93u 2011-02-08'. But no core dumps occur under this scenario on the MacOS-X version of KSH '93u+ 2012-08-01' -- the version shipped with the OS. So at least this problem appears to have been fixed already for the latest MacOX-X release (by either you or perhaps by the Apple people). In other KSH matters: There appears to be many run-time bugs in the latest KSH. Maybe some are new but many seem to be old bugs that are just occurring more frequently for some reason. The most severe (I think) of these is for KSH to think that an entered command is falsely a function and then complain that it got an "exec format error" when trying to execute the entered command. It also puts out something that seems to indicate that KSH tries to read the binary program as a function definition and then failing on that as well. I will post more on this and other run-time errors and core dumps as I have some time to investigate them (which sadly is not much). But -- oddly enough -- this version of KSH (in AST 2014-12-24 and KSH '93v- 2014-12-24') actually appears to have less important failures than some of the older KSHs over the last several years (from 2011). So thanks very much for fixing the many bugs over the last few years. I will try to compile on the latest MacOS-X perhaps when some of the remaining CLANG compilation problems are worked out. Maybe fixing some of the CLANG problems will fix some compilation problems across other platforms -- let us hope! :-) In general -- great job! Thanks again, ---- David A. Morano, PE, PhD mor...@computer.org _______________________________________________ ast-developers mailing list ast-developers@lists.research.att.com http://lists.research.att.com/mailman/listinfo/ast-developers
