JI-9008287 - configure should provide better error reports about bootjvm binaries

[Dave Pointon]

Hi David ,

On Tue, 2013-11-19 at 06:28 +1000, David Holmes wrote:
> Hi Dave,
> 
> On 19/11/2013 1:34 AM, Dave Pointon wrote:
> > 
<snip>
> 
> Minor style nit but note that the pattern used everywhere is x$FOO = x. 
> So changing to use "j" or "r" is undesirable.
> 

OK, TFT. It was merely an obviously misguided attempt to imbue the
constant with unwanted meaning.

> MOre generally this seems solvable simply by changing the existing error 
> message to "an executable binary file for XXXX could not be found" - and 
> the user could see if it was missing or just not marked executable. :)
> 

Possibly, I was trying to implement a solution that covered the bases
that I encountered when attempting to track the problem down in as near
to good SW Eng practice e.g. SPoD etc., as I could manage.

> Re-use of the BOOTJDK_CHECK_TOOL_IN_BOOTJDK macro seems reasonable but I 
> don't think you really need to include the _WITH_OPT_ABORT in the name.
> 

The macro name is intended to reflect the fact that a failed call may,
or may not, cause the immediate cessation of the script according to the
value of the 3rd arg - there are a couple of instances where the script
doesn't immediately abort, it attempts to recover from the failure - the
fact that this attempt invariably fails is neither here nor there :-)

> I'm sure Magnus will have more comments.
> 
> Cheers,
> David H.
> 
> >
> > TIA & rgds ,
> >
> 

Rgds ,
-- 
Dave Pointon FIAP MBCS

Now I saw, tho' too late, the folly of beginning a work
before we count the cost and before we we judge rightly
of our strength to go thro' with it - Robinson Crusoe