[gnu.g++] Protocol Violation

pcg@aber-cs.UUCP (Piercarlo Grandi) (02/02/90)

In article <9001312044.AA18233@arkesden.sun.com> tiemann@sun.com writes:

    I made a mistake in pointing people to MIT machines other than
    prep.ai.mit.edu.  Please do not ftp to any machine but this one to get
    GNU C++ or related software.  If you have not yet gotten the latest
    release, please refrain from doing so until GCC 1.37 is available from
    prep.ai.mit.edu.  GNU C++ 1.37.0 will be available shortly thereafter.

Ok, ok. By the way, here is one of my peeves: It would be
helpful if you kept your sources (and if RMS did the samer for
GNU C) under RCS, as this would also make it easy to do what
Larry Wall does, and others do as well, providing upgrade
patches between releases; e.g. to have release 1.36 patchlevel
1-,1,2,3,4-,4 and so on. It could also be nice to have some
major patch sets to upgrade to a major release, e.g. 1.36 to
1.37 (straight through, that is a consolidation of all
patchlevels between the two).

Now that RCS is also GNU RCS, it is also politically correct to
use it :->, and it makes generating patchlevels a snap (and also
could do away with the ChangeLog, or else make maintaining it a
snap as well).

I don't see that there are that many changes between sub-minor
releases, and FTP'ing just patches, and applying just patches,
saves bandwidth, disk space, time; encouraging theuse of RCS,
e.g. for branching local mods, etc..., would be nice.

	I have another pet peeve, that the GNU project switch to
	use cake (which is free sw as well), especially for
	multi target entities like GCC, G++, etc... GNU make
	has much the same functionality, but with a syntax that
	is even more opaque than make's; by contrast, cake has
	a very clean syntax.
-- 
Piercarlo "Peter" Grandi           | ARPA: pcg%cs.aber.ac.uk@nsfnet-relay.ac.uk
Dept of CS, UCW Aberystwyth        | UUCP: ...!mcvax!ukc!aber-cs!pcg
Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk