This is the mail archive of the
frysk@sources.redhat.com
mailing list for the frysk project.
Re: Register window workflow / potential advanced features
Andrew Cagney wrote:
Adam Jocksch wrote:
Adam Jocksch wrote:
The register window has been updated, and I've updated the frysk
website to include a workflow for the register window:
http://sourceware.org/frysk/workflow/register/
Looking at this image:
http://sourceware.org/frysk/workflow/register/register_with_formats.png
for the default case, where the byte order matches what is natural for
the inferior, I don't think users will be interested in and may even
be puzzled by the explicit (LE) or (BE) suffix. One possible way
around this is to have three options: Hex; B.E. Hex; L.E. Hex. With
hover-help clarifying the difference? I'm not sure.
Another advanced feature is to consider how these selections interact
with different ISA's (Instruction Set Architectures). If there's a
"Hex" (Natural Byte Order) then that selection can be applied to all
architectures.
It's also good that you've avoided the mistake of having a BE vs LE
toggle switch. In the real world, there's a weird and wonderful
continuum between the two extreems(1). While we'll hopefully never
have to handle those variants, as you've done, we don't want to assume
that BE and LE are the only alternatives.
Andrew
(1) For instance, little byte, big word.
I like the idea of having a default which corresponds to whatever the
natural byte order is of the particular architechture. That will of
course lead to some redundancy (i.e. LE natural order and the LE formats
both displayed), but that seems to be more of a user concern.
The idea of there being more complex formats really is something that
needs to be looked at as this window progresses, as there are a lot of
common issues between vector formatted registers, the idea of
showing/interpreting status bits, and the examples you just mentioned.