[bug #51237] Deadlock in Ctrl-C handler on Windows

[Eli Zaretskii]

Follow-up Comment #1, bug #51237 (project make):

Thanks for the analysis and the patch.
Your patch is too large to accept without legal paperwork.  Would you be
willing to assign copyright for your changes to the FSF, so we could accept
your contribution?  If so, I can send you the legal form and instructions to
fill and submit it.

Also, I wonder if your implementation is the best one (or maybe I'm missing
something).  You've inserted a call to check_need_sleep into the main loop of
reap_children.  But if the 1st arg of reap_children is non-zero, Make will
block inside process_wait_for_any, and AFAIU signaling the event you added
will not interrupt that wait.  Wouldn't it be better to instead add the event
handle to the handles on which process_wait_for_any waits?  Then the reaction
to an interrupt should be much faster and more reliable, I think.  Or am I
missing something?


    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?51237>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/