[nicl] RFR Adding -L option to jextract tool

[Henry Jen]

Thanks, looks good for a start. We might tweak what to trigger checking later, either -l alone or always.

Cheers,
Henry

> On Feb 21, 2018, at 7:54 PM, Sundararajan Athijegannathan <sundararajan.athijegannathan at oracle.com> wrote:
> 
> Yes, it seems fine.
> 
> Our emails crossed :) Responded to your earlier email.
> 
> -Sundar
> 
> On 22/02/18, 9:22 AM, Henry Jen wrote:
>> It’s fine as those type are not directly handled from cursor, the parameter declaration is pointer type.
>> 
>> Cheers,
>> Henry
>> 
>>> On Feb 21, 2018, at 12:26 PM, Henry Jen<henry.jen at oracle.com>  wrote:
>>> 
>>> It print out an warning on symbols doesn’t exist in library, but still generate code for it. It’s compatible to what happens now.
>>> 
>>> In the future we may want to skip unresolvable symbols, so a return value from symbol checker is probably good unless we think using Exception is preferred.
>>> 
>>> I will test out what will happen with a function pointer parameter, like the useFunction in the attached header file. It’s a parameter declaration and eventually will have a type to be defined later. Not quite remember what exactly will happen there, will let you know when I have a chance to test it.
>>> 
>>> Cheers,
>>> Henry
>>> 
>>> <FnType.h>
>>> 
>>> 
>>>> On Feb 21, 2018, at 5:00 AM, Sundararajan Athijegannathan<sundararajan.athijegannathan at oracle.com>  wrote:
>>>> 
>>>> -L option can be used to specifiy one or more shared object search paths. This path is used to search for symbols during jextract run. Any missing symbol results in a waring being issued during a jextract session. Note that the -rpath option is for runtime search of shared objects. -L is similar search path for jextract "tool time".
>>>> 
>>>> Webrev: http://cr.openjdk.java.net/~sundar/jextract_L/webrev.00/index.html
>>>> 
>>>> Thanks,
>>>> -Sundar