Re: what's with the loop with "in in" in u-boot.inc?

[Gary Thomas <gary@mlbassoc.com>]

On 2015-03-07 10:48, Bernhard Reutner-Fischer wrote:
> On March 7, 2015 3:04:52 PM GMT+01:00, "Robert P. J. Day" <rpjday@crashcourse.ca> wrote:
>>
>>   based on a recent post to this(?) list, i was curious about the
>> following loop construct in u-boot.inc:
>>
>>         for config in ${UBOOT_MACHINE}; do
>>             for type in in ${UBOOT_CONFIG}; do
>>                 if [ "${type}"x = "in"x ]
>>                 then
>>                     continue
>>                 fi
>>
>>   the words "in in" above don't appear to be a typo, as the next
>> condition explicitly checks for the value "in" and skips it. what's
>> the rationale for that? i used "git blame" to examine the commit that
>
> for may barf on empty iteration input.
> One usually uses for i in ${foo} '';
> and skips i with zero length. Any other is obviously fine, too though so nothing wrong or odd here.

But this code is only executed if ${UBOOT_CONFIG} is non-empty
so this test is nonsense.

My guess is that the code originally tested for ${UBOOT_CONFIG}
being empty this way, then the if/then/else was added (look at
the indentation) and the original test was never removed (which
does no harm)

>
> HTH,
>
>> introduced that, but it says nothing about that curiosity.
>>
>> rday
>
>

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------