nls-technical
[Top] [All Lists]

Re: [nls-technical] [Fwd: Re: NLS/AUGMENT question.]

To: Ken Harrenstien <klh@panix.com>
Cc: NLS Restoration Technical Discussion <nls-technical@chm.cim3.net>, Alex Bochannek <alex@p9.com>
From: Mark Crispin <mrc@CAC.Washington.EDU>
Date: Sat, 3 Mar 2007 13:58:14 -0800 (PST)
Message-id: <alpine.OSX.0.83.0703031341090.21994@pangtzu.panda.com>
On Sat, 3 Mar 2007, Ken Harrenstien wrote:
> Hmm.  The only way that "makes sense" makes sense is in the sense of
> "this is how the compiler code was written, so it makes sense that the
> code is doing the wrong thing".    (01)

Indeed.  This definitely seems to be a bug in gcc.  I've noticed a 
distressing tendency of recent versions of gcc to be of ever-lower quality 
compared to earlier versions.    (02)

> I think this is a subtle linguistic misinterpretation of the intent of
> "register"    (03)

I've seen this out of certain compiler writers (IBM's were also this way) 
who grasp at straws to justify their compiler's behavior than in doing the 
right thing.    (04)

This is akin to protocol implementors who believe that a "SHOULD" in an 
RFC means "MUST NOT".  I've met a few of these.  Or worst, the ones who 
claim that Postel's robustness principle means that syntax checks are 
banned.    (05)

> which was to serve as a HINT that this (otherwise "auto")
> variable should be given priority for optimizations -- the only thing
> that was implementation-dependent is whether or not the compiler
> actually put them in machine "registers"    (06)

That's what K&R says.  But what do K&R know, they just invented the 
language........    (07)

> I'd still have expected better from GCC people.  Maybe.    (08)

I don't.  Sheesh, these days most of the C library is non-reentrant, so 
you now are pretty much prevented from calling *any* C library function in 
a signal handler -- even if the signal is going to exit the process and 
has no intention of dismissing.  That's right, you can't even do a 
printf() in a signal handler with GNU's C library, much less a syslog(). 
Obviously RMS has forgotten about PCLSR.    (09)

Sadly, GNU/FSF is rapidly pushing itself into irrelevancy.  Too bad they 
also did such an excellent job of attacking any open source work that 
isn't part of their church.    (010)

> I agree with MRC.  Go ahead and patch it appropriately and get on with
> things.    (011)

And it seems to me that the very simplest patch is
        -Dregister=
in the CFLAGS in the makefile.    (012)

-- Mark --    (013)

http://panda.com/mrc
Democracy is two wolves and a sheep deciding what to eat for lunch.
Liberty is a well-armed sheep contesting the vote.    (014)

_________________________________________________________________
Message Archives: http://chm.cim3.net/forum/nls-technical/
Shared Files: http://chm.cim3.net/file/work/project/nls-restore/
Community Portal: http://www.computerhistory.org/  
To Post: mailto:nls-technical@chm.cim3.net
Community Wiki: http://chm.cim3.net/cgi-bin/wiki.pl?NLS_Restoration    (015)
<Prev in Thread] Current Thread [Next in Thread>