Linking and regression problems on *

[Jan Weerts]

Harald Wellmann wrote:
>First of all, as a new subscriber to this list, let me say that I'm
>quite impressed with kaffe.

Dito.

I tried to use the jitterbug web interface to submmit this but got my
own submission bounced with a user unknown, so here I am using the
mailing list...

>However, on HP-UX, I didn't get nearly as far. The latest messages
>regarding HP-UX in the mailing archives are a few months old, so I'd
>like to know if anyone has managed to build and run kaffe 1.0.5 on
>HP-UX.

Same problems here on SunOS4.1.3_U1

>The first problem I've had is to do with shared libraries (as usual
>on HP-UX, I'm inclined to say...).

dunno about the 'as usual' part, but same problems here

>so I really wonder what's going wrong here. By the way, I'm using egcs-1.1.2.
>I also tried HP's "cc -Aa", but that didn't work at all, i.e. there were
>some make errors.

I tried some combinations of configuration parameters
   - OS: 4.1.3_U1, 4.1.4
   - gcc: 2.8.1, 2.7.2
   - vm: intrp, jit
   - libs: static, shared

to no avail. Always the same problems.

>results of "make check" don't look too promising:

Dito. Results are similar over here. I even get some more 'funny'
stuff, like *un*deterministic pass/fail/hang.

Some always fail, e.g DoubleNeg, DoublePrint.
Some sometimes fail, e.g. TestIntLong.
Some tests hang, e.g. ThreadState, UDPTest.
Sometimes compiling fails 
  with: exception.c:308: failed assertion `!INTS_DISABLED()'
  or: ld.so: call to undefined procedure _strlen from 0x.....

Setting VERBOSE=on or compiling by hand and asking for
extra jit information with -verbosejit barfs even more.
This usually gives far more INTS_DISABLED failed asserts.

Especially the undeterministic behaviour troubles me. I thought about
machine utilization as changing factor (shouldn't be), but could not
deduct anything by going to a near-single-user-mode.

Gee, I'd like to use kaffe
  Jan
-- 
Jan Weerts         for my PGP key: send email with subject "get pgp key"
                                             #include <spammers-warning>
I know, you believe, you understood, what you think I said; but I am not
sure, you understand, that what you heard, isn't what I meant.   [Nixon]