Ok, which branch/tag do I use for egcs 1.1.2?

craig@jcb-sc.com craig@jcb-sc.com
Fri Feb 26 04:56:00 GMT 1999


> GP> Regarding to more (detailed) explanations, I'm not sure about
> GP> this, as I cannot recall any questions on the list concerning
> GP> this issue lately. What did you have in mind?
>
>Nothing specific.  Anything Craig (or anyone else) doesn't understand
>about the process that's worth a line there, I guess.  It's a bit
>difficult for people who made the mistakes years ago to know :-/.

Well, my modus operandi lately has been to act like a naive user,
reading the directions, following them, with the usual occasional
forgetfulness that entails.  I've found this makes it easier for me
to design software that actually works the right way, instead of
the way that was convenient to its developers when it was first
being created, because I learn the little lessons that aren't so
easily learned using the brute-force, invent-and-memorize-every-
useful-incantation approach that seemed oh-so-attractive to me
some 20 years ago.

Anyway, if that "update -P -r branch..." recipe isn't shown as the One
Right Way to check out a branch in the docs, then the docs need
fixing.  (Yes, I realize I could read *all* the specifications and
derive the One Right Way.  But I'd rather understand what ordinary
people go through when trying to deal with computers.)  I don't know
which docs I mean, though, offhand.

And, yes, this means I don't particularly care for the way utilities
like diff, patch, and so on were designed for use as a means to
distribute changes to software -- the idea of having to specify special
options to get obviously correct behavior is a very bad idea.  (This
might not be the fault of diff and patch per se, though, in the same
sense that it isn't exactly the fault of the make utility that we
insist on using it the way we do.)

        tq vm, (burley)


More information about the Gcc-patches mailing list