Regarding my two patches against config.sub

[Yann Droneaud]

Hi,

I've just sent two patches to improve config.sub.

Yann Droneaud (2):
      config.sub: use $name instead of $1
      config.sub: be more liberal on input case: accept upper case name

 ChangeLog  | 10 ++++++++++
 config.sub | 72
+++++++++++++++++++++++++++++++++++++-----------------------------------
 2 files changed, 47 insertions(+), 35 deletions(-)

The rationale behind this is: 

when using ./configure --host=<name>, --build=<name>,--target=<name>,
<name> is not the output of config.guess and can be written in
upper-case. For example ./configure --host=X86-LINUX

With the current config.sub, I'm getting:
    $ ./config.sub X86-LINUX
    Invalid configuration `X86-LINUX': machine `X86' not recognized

With the two patches, I'm getting a better result:

    $ ./config.sub X86-LINUX
    x86-unknown-linux-gnu

Some concerns:

- tr tool was not previously used in config.sub, but it is in
config.guess. If it cannot be used, perhaps sed -e
'y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/' can be used
instead. But the 'y/' command is not used in config.guess: not portable
enough ?

- MiNT: the way MiNT was recognized seems a bit difficult to understand.
  In uppercase it has a any prefix, but in lowercase it has a dash, but 
  not every time. Might be interesting to check for real.

Regards.

-- 
Yann Droneaud
OPTEYA