On 21 Apr 2016, at 09:36, Grisha Levit wrote: > A few issues inspired by the coproc proposal in this thread: > http://lists.gnu.org/archive/html/bug-bash/2016-04/msg00050.html > > 1. coproc unsets readonly NAME after the process completes [..] > 2. Segfault if NAME is an existing associative array [..]
these two can and in my opinion should be checked and ruled out before allocating coproc, at the begining of execute_coproc. > 3. If NAME is $,?,@, etc, bash will create a new array variable with that > name. This zeroth index of this array variable will then take precedence > over the real variable in some contexts these and others will be no longer an issue, when the patch in the thread you mention gets merged see https://lists.gnu.org/archive/html/bug-bash/2016-04/msg00053.html cheers, pg