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: fallout from decimal sh_info readelf patch


On Thu, 6 May 2004, Alan Modra wrote:
> On Wed, May 05, 2004 at 06:59:36PM -0400, Hans-Peter Nilsson wrote:
> > And while I'm here, please people, please avoid changing the
> > default output or output of common options for binutils
> > programs.  Make changes depend on some new non-default option.
>
> Sometimes, I think a change to default output is worth the pain of
> updating the testsuite.

Sometimes, but rare enough.  I'd like to avoid having it
perceived as a rule that it's worth *any* pain to change the
output format of whatever binutils program.  It's *not* the
testsuite changes that is the primary reason to avoid output
format changes; it's all other use.

Alternatively, maybe we should instead have an option called
something like "--canonical-format" that people can use in
scripts, and we never change that output format no matter what
(for most "what"'s)?

> Incidentally, if we ever implement Roland's suggestion,
> http://sources.redhat.com/ml/binutils/2003-04/msg00435.html
> to do away with dynamic section symbols, we'll be up for a lot more
> pain than HJ's fix..

That surely would be a "change in output", but not a "change in
output *format*" which is what I really meant.  When the "output
change" reflects actual file content change, I have no
objection.

brgds, H-P


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