Re: Confirming X02zlevi test failures

["Jun T." <takimoto-j@xxxxxxxxxxxxxxxxx>]

It seems fixing the zpty problem in BSD/Darwin is not easy.

If we must resort to a (hopefully temporary) workaround, 
I think 'bindkey -r "\e~"' is better than zpty_flush()
because the latter seems to be not 100% reliable.

In the patch below, instead of putting 'bindkey -r "\e~"'
in comptest, it is called just before the failing test in
X02zlevi, in order to minimize the interference with other
tests which are also using comptest.

Jun


diff --git a/Test/X02zlevi.ztst b/Test/X02zlevi.ztst
index 8f93902..0c346a4 100644
--- a/Test/X02zlevi.ztst
+++ b/Test/X02zlevi.ztst
@@ -37,6 +37,7 @@
 >3
 >CURSOR: 5
 
+  zpty_run 'bindkey -r "\e~"'
   zletest $'\e' $'~aI\e' $'~o\e' \~
 0:swap case on a blank line
 >BUFFER: i
diff --git a/Test/comptest b/Test/comptest
index 1f4dac6..9c92f96 100644
--- a/Test/comptest
+++ b/Test/comptest
@@ -162,7 +162,7 @@ comptest () {
 zletest () {
   local first=0
   for input; do
-    zpty_flush Before zletest
+    # zpty_flush Before zletest
     # sleep for $KEYTIMEOUT
     (( first++ )) && read -t 0.011 -k 1 < /dev/null
     zpty -n -w zsh "$input"