On Tue, Jul 07, 2009 at 01:10:36AM +0200, Jiří Paleček wrote:
> Hi
>
> On Mon, 06 Jul 2009 23:51:12 +0200, Matt Helsley <[email protected]>
> wrote:
>
>> On Thu, Jul 02, 2009 at 06:09:16PM -0400, Mike Frysinger wrote:
>>> On Thursday 02 July 2009 16:32:53 Matt Helsley wrote:
>>> > On Thu, Jul 02, 2009 at 02:35:46PM -0400, Mike Frysinger wrote:
>>> > > in general, converting sh to bash is a bad idea as it can cause  
>>> issues
>>> > > for us embedded peeps.
>>> >
>>> > I suppose that's because of the memory bash consumes. Is there  
>>> anything
>>> > else it does which is bad for embedded?
>>>
>>> there is memory and disk usage concerns, but there are also systems  
>>> that cant
>>> run bash at all (e.g. systems without fork() -- all no-mmu Linux ports)
>>> -mike
>>
>> Here's that updated patch which hopefully should be able to replace 3
>> and 4 of my recent series. Unfortunately I haven't had a chance to test 
>> it so
>> I'm posting it in case someone else does. As the original author of the
>> patch Jiri is welcome to add his signoff as well.
>
> Thanks for doing this. I have some comments to the patch, see below.
>
> I have made my "version 2" of the patch too, containing also changes to  
> other files (like runpwtest.sh). If you're interreseted, I'll send it.

Sure. I'd have happily used that. My main concern is getting these fixed so
I can cleanly run the containers testcases. I think it's actually worse than
just removing "bashisms" because the interprettern is inconsistent with the
scripts and thus nobody can run them.

>
>> From: Jiri Palecek <[email protected]>
>>
>> This patch removes a bunch of bashisms.
>>
>> It was originally written by Jiri Palecek but lacked changes
>> requested by Mike Frysinger.
>>
>> Signed-off-by: Matt Helsley <[email protected]>
>> ---
>>
>> v2:
>>      Fix pushd/popd removal to save PWD and restore it using a regular
>>              variable.
>
> I wonder if this couldn't be done by a simple "cd -", or even left out  
> completely.

Hmm. I didn't think it could be left out completely because of the way
$PWD was inconsistently used. Hence some trash could be left if we
didn't cd to LTPTMP before and remove it after the test body.

>
>>      Remove empty echoes in favor of printf.
>
> This hasn't been done thoroughly (some empty echoes remain in the patch). 
> Also, the patch misses one thing, that if you use printf, you have to add 
> an extra linefeed at the end, ie:
>
>  echo -e "foo\n"
>
> has to be replaced by
>
>  printf "foo\n\n"
>
> for it to work equivalently.

Ack, you're right.

>
>>      Remove non-bashism "kill -s KILL ..." and keep "kill -9 ..."
>>              If this is still a problem then it could be placed in
>>              a separate patch.
>>      Keep the exact formatting in cron0X tests -- use printf instead of echo.
>>      Use non-greedy expansion of tty device in ioctl test.
>>      Change the increment syntax to use : $(( X += 1 )) instead of:
>>              X=$((X + 1))
>
> I changed this to (IMHO cooler syntax, it's for the sake of aesthetics  
> after all) ": $((X++))" in my patch.

As Mike has noted, dash didn't work with that.

Cheers,
        -Matt

------------------------------------------------------------------------------
Enter the BlackBerry Developer Challenge  
This is your chance to win up to $100,000 in prizes! For a limited time, 
vendors submitting new applications to BlackBerry App World(TM) will have 
the opportunity to enter the BlackBerry Developer Challenge. See full prize 
details at: http://p.sf.net/sfu/blackberry
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list

Reply via email to