Re: PPP modules with 1.3.59

[turnbull@example.com (Stephen J. Turnbull)]

>>>>> "Craig" == Craig Oda <craig@example.com> writes:

    Craig> On Mon, 29 Jan 1996, Stephen J. Turnbull wrote:
    >> The latest rev seems to be 2.2.1d at SunSITE.UNC.edu, although
    >> 2.2.0e is also available.  Why, I don't know.

    Craig> This is probably ppp-2.1.2d.tar.gz which was uploaded on
    Craig> January 13.  I think it is for the 1.2.x kernel series.

Yes, that's right.  My bad.  I transposed the digits.

    Craig> I am now unable to compile the ppp network driver on two
    Craig> different machines.  Both machines are using Slackware 3.0
    Craig> with gcc 2.7.0 elf.  One of the machines has Slackware 3.0
    Craig> installed on a recently formatted partition which is
    Craig> unmodified.  Both compilations hang at the same place.

I feel for you.  I just built 1.3.59, and discovered that the Buslogic
SCSI driver *has* in fact been completely rewritten---hurray!---and
now it won't recognize the alleged BusLogic clone I'm using---boo hoo!
So I get a kernel panic---"no file systems found".  I'd panic, too....
Sigh.  I'll have to try the Adaptec driver while waiting for the
BusLogic driver's author to reply.

Where is the compile hang?  In the driver or in the kernel?  Which
file?  Do you have a line number?

    Craig> I've tried kernel versions 1.3.56 - 1.3.59 with the same
    Craig> results.  I've searched through USENET and haven't found

Did you use Alta Vista or a similar engine?

    Craig> too many references to this problem so I don't think it is
    Craig> something with the ppp driver itself.  Something might be
    Craig> wrong with my gcc installation, possibly I have messed up
    Craig> headers??  Perhaps I should install gcc 2.7.2??

Yes.  Messed up headers rarely results in a hang, that should be
detected by the compiler or result in a buggy application..  GCC's
x.y.0 releases are famous for this kind of thing.  They usually
contain new features like better optimizations and such, and often do
this.  GCC 2.7.1 is also rumored to have severe bugs in the C++ and
Objective-C compilers.  It's not clear to me that returning to 2.6.3
isn't the best idea.  I always keep two editions of GCC around---it's
worth the space when things don't compile.  Often one version will
catch a syntax error that blows up another.

Also, try 'gcc -v'; this often allows you to pinpoint which pass is
blowing up.  I think most system packages turn on -Wall and most of
the few -Wxxx that don't come with -Wall; give that a try if not.

-- 
                            Stephen J. Turnbull
Institute of Policy and Planning Sciences                    Yaseppochi-Gumi
University of Tsukuba                      http://turnbull.sk.tsukuba.ac.jp/
Tennodai 1-1-1, Tsukuba, 305 JAPAN                 turnbull@example.com