This is the mail archive of the binutils@sources.redhat.com mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: patch for binutils include/ansidecl.h.


On Sun, Sep 22, 2002 at 09:34:07PM -0700, Doug Evans wrote:
>  > That is a real limitation and annoying rule.  It is often impossible to
>  > find out *WHY* a change was done.  Which is a Bad Thing.  I can read the
>  > "cvs diff -r X -r X+1" to see the what.  I really wish GNU ChangeLog
>  > entries would grow the *WHY*, or the culture be changed to tell the *WHY*
>  > rather than just repleat the ChangeLog entry in the CVS commit log.
> 
> fwiw, from http://www.gnu.org/software/guile/changelogs/guile-changelogs_3.html
> 
> One should never need the ChangeLog to understand the current
> code. If you find yourself writing a significant explanation in the
> ChangeLog, you should consider carefully whether your text doesn't
> actually belong in a comment, alongside the code it explains. Here's an
> example of doing it right: 

The example you show isn't a good one -- it obviously fixes a bug.  Often
GCC changes seem to be purely for change sake.  An explanation of why
such changes are a Good Thing would be nice.  Examples:

    2002-09-17  Dale Johannesen  <dalej@apple.com>
	* cfgcleanup.c (try_forward_edges):  Do not forward a
	branch to just after a loop exit before loop optimization;
	this interfered with doloop detection.

Actually this is actually a good example of a useful change -- it does
include *why* the change was needed.

    2002-09-17  Steve Ellcey  <sje@cup.hp.com>
    * config/ia64/elf.h: Remove CPP_PREDEFINES.

Why??  It certainly had a value ("-Dia64 -Amachine=ia64"), where are
these getting defined now?  Grepping for ``-Dia64'' doesn't show
anything, greping for ``ia64'' gives too many hits to be useful.  So I
can't eyeball where this is now being defined.  Now maybe it isn't as it
is in the application name space and that is bad and being a new
platform, that define wasn't ingraned yet.  But the ``-Amachine=ia64''
should still apply.  I now have to scan the mailing lists to see where
this was discussed, or build the compiler myself to ensure no
functionality has been lost for the FreeBSD/ia64 compiler.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]