cvs commit: src/usr.sbin/sysinstall main.c

[Roman Kurakin]

Peter Jeremy wrote:
> On 2007-May-01 10:48:28 +0400, Andrey Chernov <ache at freebsd.org> wrote:
>   
>> On Mon, Apr 30, 2007 at 06:39:57PM -0700, Alfred Perlstein wrote:
>>     
>>> Using the strategy "commit to -current then suffer the fallout"
>>> is pretty bogus.
>>>       
>> The only possible. Nobody can run all ports at once. Kris already promise 
>> all ports build results with those changes in, lets see.
>>     
>
> Note that just building the ports with these changes will not demonstrate
> much.  This change alters the functionality of putenv() rather than the
> API/ABI so testing the change requires exercising the ports.  This is
> a much more difficult task.
>   
I suggest to install all ports sources and grep them at first. I am sure 
some of ports
could be marked as bug-less and other should be marked for exec-check or 
probably
for more accurate review not just grep.

rik
>   
>> Apples and oranges. select is BSD own function unlike putenv which is not 
>> BSD own and initially adopted without too much care.
>>     
>
> Then someone needs to fix the "HISTORY" section of getenv(3).  I've
> looked thru the TUHS and CSRG files and it didn't exist in V7 or any
> of the early BSDs.  I'm not sure when it really did appear.  getenv(3)
> in 4.3reno doesn't claim compliance to any standard for putenv().
>
>