RFR: 8039411 : Add environment variable support to fixpath

[Mike Duigou]

On Apr 8 2014, at 23:26 , Erik Joelsson <erik.joelsson at oracle.com> wrote:

> Indentation look weird around line 460 and 470 (tab vs space?), otherwise it's ok to me.

Corrected before push. Some text editor I was using was allowing tabs to sneak in.

Mike
> 
> /Erik
> 
> On 2014-04-08 21:31, Mike Duigou wrote:
>> I have made the changes Tim suggested along with other cleanups, additional diagnostics and refinements.
>> 
>> http://cr.openjdk.java.net/~mduigou/JDK-8039411/1/webrev/
>> 
>> I have not tested this change on mingw/msys and probably won't attempt to unless it is required.
>> 
>> Mike
>> 
>> 
>> On Apr 8 2014, at 08:10 , Tim Bell <tim.bell at oracle.com> wrote:
>> 
>>> Hi Mike
>>> 
>>> Looks good - one thing to pick on is line 402 - I'd like to see what was in var if the setting fails.
>>> 
>>> Tim
>>> 
>>> On 04/08/14 08:35, Erik Joelsson wrote:
>>>> Hello Mike,
>>>> 
>>>> My C is a bit rusty, but I think it looks good in general. If you are able to test it on mingw/msys I think that would be good since it's a pretty big change.
>>>> 
>>>> /Erik
>>>> 
>>>> On 2014-04-07 23:53, Mike Duigou wrote:
>>>>> Hello all;
>>>>> 
>>>>> Currently the fixpath utility used in windows builds expects that the first parameter it is passed will be the path of the executable. In some cases it's desirable to define environment variables which will apply during the execution of that executable. This change adds support for defining environment variables. The variables appear before the executable. Currently the command line parsing assumes that all arguments containing "=" before the command path are environment variables. (This precludes the executable having '=' in it's path, which is unlikely anyway).
>>>>> 
>>>>> The remainder of the changes were lint warnings suggested by Visual C. (mostly const)
>>>>> 
>>>>> https://bugs.openjdk.java.net/browse/JDK-8039411
>>>>> http://cr.openjdk.java.net/~mduigou/JDK-8039411/0/webrev/
>>>>> 
>>>>> Mike
>