thetroll@rata.vuw.ac.nz (The Troll) (03/16/90)
I have successfully managed to build Perl 3.0 Patchlevel 15 on a VAX 5400 (a rebadged MIPS box). However it fails the following tests if the compiler flag -O is used: op.eval; test 7, op.substr; test 40. Perl 3.0 patchlevel 8 did not have this problem (it had other problems instead). If anyone knows why this is happening could they let Larry know so that he can decide whether to fix it in a future release or not. If anyone else with a 5400 is having problems getting Perl to build, try switching off optimisation. print "I am a bear with a very small brain and I do not understand"; -- Alex Heatley Computing Services Centre Domain: alex@rata.vuw.ac.nz Victoria University of Wellington Path: I don't support bang paths P.O Box 600, New Zealand. Never mind the Titanic - is there any news of the iceberg?
ktl@wag240.caltech.edu (Kian-Tat Lim) (03/16/90)
In article <1990Mar16.044319.14338@kaukau.comp.vuw.ac.nz>, thetroll@rata (The Troll) writes: >I have successfully managed to build Perl 3.0 Patchlevel 15 on a VAX 5400 >(a rebadged MIPS box). However it fails the following tests if the >compiler flag -O is used: op.eval; test 7, op.[subst]; test 40. This appears to be a generic MIPS problem, as it also occurs on our SGI Iris 4D/240. I was hoping patch 15 would fix it, but it's still there. On our box, the string returned by the eval in test 7 of op.eval is getting stepped on somehow; compiling doio.c, eval.c, and str.c without optimization (so dbx does something useful) doesn't help the problem. op.subst (not op.substr) test 40 reports "Out of memory!", and our SYSLOG claims that perl wants 65536 4K pages of VM. I haven't tried to track that one down yet. Patchlevel 12 works beautifully on this machine. Fixed in patch 16? -- Kian-Tat Lim (ktl@wagvax.caltech.edu, KTL @ CITCHEM.BITNET, GEnie: K.LIM1) Perl is the Swiss Army chainsaw [of Unix programming]. -- Dave Platt's friend
chk%alias@csri.toronto.edu (C. Harald Koch) (03/17/90)
In article <1990Mar16.044319.14338@kaukau.comp.vuw.ac.nz> thetroll@rata.vuw.ac.nz (The Troll) writes: >I have successfully managed to build Perl 3.0 Patchlevel 15 on a VAX 5400 >(a rebadged MIPS box). However it fails the following tests if the >compiler flag -O is used: op.eval; test 7, op.substr; test 40. > >Perl 3.0 patchlevel 8 did not have this problem (it had other problems >instead). It does the same thing on the SGI 4D series machines, which also use MIPS processors. I haven't had time to look into it at all; I just reverted to patchlevel 8 (which works here). -- C. Harald Koch Alias Research, Inc., Toronto ON Canada chk%alias@csri.utoronto.ca chk@gpu.utcs.toronto.edu chk@chk.mef.org If awk is the Swiss Army knife of UNIX, then Perl is the Swiss Army chainsaw! -Author Unknown; related by <dplatt@coherent.com> (Dave Platt)
ktl@wag240.caltech.edu (Kian-Tat Lim) (03/17/90)
In article <14317@cit-vax.Caltech.Edu>, ktl@wag240 (Kian-Tat Lim) writes: > [about problems with op.eval and op.subst in patch 15]. At a suggestion from beldar@mips.com (Gardner Cohen), I recompiled without volatile. Everything works. Looks like the compiler is generating bad code, even though it gives no errors. -- Kian-Tat Lim (ktl@wagvax.caltech.edu, KTL @ CITCHEM.BITNET, GEnie: K.LIM1) Perl is the Swiss Army chainsaw [of Unix programming]. -- Dave Platt's friend
mellon@fenris.pa.dec.com (Ted Lemon) (03/24/90)
>I have successfully managed to build Perl 3.0 Patchlevel 15 on a VAX 5400 >(a rebadged MIPS box). However it fails the following tests if the >compiler flag -O is used: op.eval; test 7, op.substr; test 40. Excuuse me? The DECstation 5400 (not VAX 5400) is not a rebadged mips box. Have you ever heard of a Mips box with a Q-bus in it? Have you ever heard of a little-endian Mips Box? It does have a Mips R3000 chip in it, of course, but it's a completely different product. Sorry to interject this here, but I have trouble ignoring misinformation. _MelloN_