Question about dead code in build-infra/common/autoconf/basics.m4

[Magnus Ihse Bursie]

On 2012-09-17 09:28, Fredrik Öhrström wrote:
> 2012/9/17 Tim Bell <tim.bell at oracle.com>:
>> Based on my experiments this evening, lines 169...172 will never be
>> executed, because OPENJDK_BUILD_OS does not have a value at this point. The
>> if test at line 168 evaluates to false, even on Windows build systems.
> You are absolutely right. It just a minor oversight when Magnus
> restructured and
> modularized the autconf source files. There are in fact three more places where
> OPENJDK_BUILD_OS is used before its value is set. In BASIC_INIT and two
> places where MAKE_SPACE_SAFE is called. I am sure it will be relatively easy
> to fix.

You're right. And it's trivial to fix these places you identified; I 
recoginze two of them already but had forgotten about them. :-(

However, what's actually turned out not to be so trivial was to get 
*everything* right. There's a lot of the "chicken and egg" kind of 
problem in the configure script. I fixed several of them when I 
modularized the configure source; I introduced several more but fixed 
most of these later on. I'm still not confident that we're doing 
everything right, and what's worse, I'm not sure how to test that 
either. Typically the problems that are likely to be left are like the 
one you discovered -- code for handling an unusal error condition on a 
single platform.