On 2019/02/08 19:41, Jouni Malinen wrote:
That looks excessive pretty ugly. Wouldn't it be simpler to just add
support for None return without separating processing based on version?
I.e., something like this (and well, I'd merge in the applicable
.decode() changes from patch 1/4 to this patch taking into account they
alone do not really fix these same read() lines):
diff --git a/tests/hwsim/vm/parallel-vm.py b/tests/hwsim/vm/parallel-vm.py
index df6c19b..d2b0def 100755
--- a/tests/hwsim/vm/parallel-vm.py
+++ b/tests/hwsim/vm/parallel-vm.py
@@ -90,7 +90,10 @@ def vm_read_stdout(vm, i):
ready = False
try:
- out = vm['proc'].stdout.read().decode()
+ out = vm['proc'].stdout.read()
+ if out == None:
+ return False
+ out = out.decode()
except:
return False
logger.debug("VM[%d] stdout.read[%s]" % (i, out))
@@ -191,9 +194,11 @@ def show_progress(scr):
running = True
first_running = True
try:
- err = vm[i]['proc'].stderr.read().decode()
- vm[i]['err'] += err
- logger.debug("VM[%d] stderr.read[%s]" % (i, err))
+ err = vm[i]['proc'].stderr.read()
+ if err != None:
+ err = err.decode()
+ vm[i]['err'] += err
+ logger.debug("VM[%d] stderr.read[%s]" % (i, err))
except:
pass
@@ -247,8 +252,10 @@ def show_progress(scr):
running = True
try:
err = vm[i]['proc'].stderr.read()
- vm[i]['err'] += err
- logger.debug("VM[%d] stderr.read[%s]" % (i, err))
+ if err != None:
+ err = err.decode()
+ vm[i]['err'] += err
+ logger.debug("VM[%d] stderr.read[%s]" % (i, err))
except:
pass
> That looks excessive pretty ugly. Wouldn't it be simpler to just add
> support for None return without separating processing based on version?
I can understand because my previous modification is like it.
My purpose of this patch is not to catch exception if not needed.
Because catching exception makes debugging more difficult.
Indeed, sometimes I needed to comment out try statements temporarily
while doing python2 to 3 project.
Which do you like ?
_______________________________________________
Hostap mailing list
Hostap@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/hostap
