death was premature: fixinclude program working

death was premature: fixinclude program working

[Bruce Korb]

It turns out the streams bug on the Sun is that
they do *not* send the M_PROTO message, command S_HANGUP
until the writer actually exits and not merely closes
the pipe.  This is contrary to the documentation and
contrary to SVR4.2.  :-(  Invoking "ioctl( fd, M_PROTO, &cmd );"
did not help, either.

Anyway, on my Sun:

fixincludes    26 min.
inclhack.sh    23 min.
fixincl        13 min.

"inclhack.sh" and "fixincl" produce identical results.
They differ from "fixincludes" in that '#undef NULL'
starts in column 1 instead of 16, c++-style comments
are deleted instead of converted to c-style (this change
should be applied to "fixincludes" as well -- sys/audiovar.h
winds up with nested comments whithout the change), and
lastly two fixes to stdlib.h are applied in an order
that produces an innocuous difference.

I need a few people to volunteer to run these things
on their boxes and verify consistent results.
Then it can be more generally released.
Please contact me via email.  Thanks.
[[ The tarball is too big to email to the world. ]]

Regards,
	Bruce

Re: fixinclude program - does anyone care?

[Bruce Korb]

Bruce Korb wrote:
...
> Anyway, on my Sun:
> 
> fixincludes    26 min.
> inclhack.sh    23 min.
> fixincl        13 min.

Make that 8 min 30 sec - I changed the machine name selection
expression to somthing that cannot be used in the shell code
and saved another five minutes.

> "inclhack.sh" and "fixincl" produce identical results.
...

-- 
Bruce Korb                      | Data Design Systems, Inc.
Korbb at DataDesign dot com     | 45 Cabot Dr., Suite 110
Voice:  408-260-0280            | Santa Clara,  CA   95051
Fax:    408-260-0281            | USA