[PATCH] oepydevshell-internal.py: decode only when readdata is valid

[PATCH] oepydevshell-internal.py: decode only when readdata is valid

[changqing.li]

From: Changqing Li <changqing.li@windriver.com>

fix below problem:
pydevshell raises exception when maximize the python shell window.
when click maximize, rlist of select return ready object, but the
pty.read is None, so throw exception of 'NoneType' object has no
attribute 'decode', change to only decode when readdata is valid.

[YOCTO #11875]

Signed-off-by: Changqing Li <changqing.li@windriver.com>
---
 scripts/oepydevshell-internal.py | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/scripts/oepydevshell-internal.py b/scripts/oepydevshell-internal.py
index 04621ae..2d1bc77 100755
--- a/scripts/oepydevshell-internal.py
+++ b/scripts/oepydevshell-internal.py
@@ -63,7 +63,10 @@ try:
             (ready, _, _) = select.select([pty, sys.stdin], writers , [], 0)
             try:
                 if pty in ready:
-                    i = i + pty.read().decode('utf-8')
+                    readdata=pty.read()
+                    if readdata == None:
+                        continue
+                    i = i + readdata.decode('utf-8')
                 if i:
                     # Write a page at a time to avoid overflowing output 
                     # d.keys() is a good way to do that
-- 
2.7.4

Re: [PATCH] oepydevshell-internal.py: decode only when readdata is valid

[Richard Purdie]

On Tue, 2018-11-13 at 17:04 +0800, changqing.li@windriver.com wrote:
> From: Changqing Li <changqing.li@windriver.com>
> 
> fix below problem:
> pydevshell raises exception when maximize the python shell window.
> when click maximize, rlist of select return ready object, but the
> pty.read is None, so throw exception of 'NoneType' object has no
> attribute 'decode', change to only decode when readdata is valid.
> 
> [YOCTO #11875]
> 
> Signed-off-by: Changqing Li <changqing.li@windriver.com>
> ---
>  scripts/oepydevshell-internal.py | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/scripts/oepydevshell-internal.py b/scripts/oepydevshell-
> internal.py
> index 04621ae..2d1bc77 100755
> --- a/scripts/oepydevshell-internal.py
> +++ b/scripts/oepydevshell-internal.py
> @@ -63,7 +63,10 @@ try:
>              (ready, _, _) = select.select([pty, sys.stdin], writers
> , [], 0)
>              try:
>                  if pty in ready:
> -                    i = i + pty.read().decode('utf-8')
> +                    readdata=pty.read()
> +                    if readdata == None:
> +                        continue
> +                    i = i + readdata.decode('utf-8')

Usually you'd write this as:

if readdata:
    i = i + readdata.decode('utf-8')

Is there a reason not to do that here?

If you really want to test for None, you usually write:

if readdata is None:

which is a little more pythonic.

Cheers,

Richard

Re: [PATCH] oepydevshell-internal.py: decode only when readdata is valid

[Burton, Ross]

On Tue, 13 Nov 2018 at 10:03, Richard Purdie
<richard.purdie@linuxfoundation.org> wrote:

> Usually you'd write this as:
>
> if readdata:
>     i = i + readdata.decode('utf-8')
>
> Is there a reason not to do that here?

Whilst this code is being changed, is there a chance of getting a
partial read that has half of a utf-8 byte sequence?

Ross

Re: [PATCH] oepydevshell-internal.py: decode only when readdata is valid

[Changqing Li]

On 11/13/18 6:03 PM, Richard Purdie wrote:
> On Tue, 2018-11-13 at 17:04 +0800, changqing.li@windriver.com wrote:
>> From: Changqing Li <changqing.li@windriver.com>
>>
>> fix below problem:
>> pydevshell raises exception when maximize the python shell window.
>> when click maximize, rlist of select return ready object, but the
>> pty.read is None, so throw exception of 'NoneType' object has no
>> attribute 'decode', change to only decode when readdata is valid.
>>
>> [YOCTO #11875]
>>
>> Signed-off-by: Changqing Li <changqing.li@windriver.com>
>> ---
>>   scripts/oepydevshell-internal.py | 5 ++++-
>>   1 file changed, 4 insertions(+), 1 deletion(-)
>>
>> diff --git a/scripts/oepydevshell-internal.py b/scripts/oepydevshell-
>> internal.py
>> index 04621ae..2d1bc77 100755
>> --- a/scripts/oepydevshell-internal.py
>> +++ b/scripts/oepydevshell-internal.py
>> @@ -63,7 +63,10 @@ try:
>>               (ready, _, _) = select.select([pty, sys.stdin], writers
>> , [], 0)
>>               try:
>>                   if pty in ready:
>> -                    i = i + pty.read().decode('utf-8')
>> +                    readdata=pty.read()
>> +                    if readdata == None:
>> +                        continue
>> +                    i = i + readdata.decode('utf-8')
> Usually you'd write this as:
>
> if readdata:
>      i = i + readdata.decode('utf-8')
>
> Is there a reason not to do that here?
>
> If you really want to test for None, you usually write:
>
> if readdata is None:
>
> which is a little more pythonic.
>
> Cheers,
>
> Richard
>
Thanks. I will correct the patch as suggested.

//Changqing

-- 
BRs

Sandy(Li Changqing)

Re: [PATCH] oepydevshell-internal.py: decode only when readdata is valid

[Changqing Li]

On 11/13/18 7:10 PM, Burton, Ross wrote:
> On Tue, 13 Nov 2018 at 10:03, Richard Purdie
> <richard.purdie@linuxfoundation.org> wrote:
>
>> Usually you'd write this as:
>>
>> if readdata:
>>      i = i + readdata.decode('utf-8')
>>
>> Is there a reason not to do that here?
> Whilst this code is being changed, is there a chance of getting a
> partial read that has half of a utf-8 byte sequence?
>
> Ross

Seems "continue"  will delay  the deal of  "i"  to next loop, but will 
not lost data.

but I still change to let it continue to deal with "i" , in case there 
is still "i" not

outputed in last loop. it should be better.


changqing

-- 
BRs

Sandy(Li Changqing)