Re: Exceptions in IPC::Open2
[p5sagit/p5-mst-13.2.git] / README.beos
CommitLineData
c4f23d77 1$Id: README.beos,v 1.2 1998/05/02 01:55:04 dogcow Exp dogcow $
2
3Notes on building perl under BeOS:
4
5GENERAL ISSUES
6--------------
7perl will almost compile straight out of the box with ./Configure -d, but
8there are a few gotchas:
9
10Currently, you have to edit config.sh and remove SDBM_File from the
11dynamic_ext= and extensions= lines. SDBM_File does not build properly
12at this time. You need to run ./Configure -S after editing config.sh.
13
14In addition, with mwcc, after doing `make depend`, you need to edit
15makefile and x2p/makefile and remove the lines that mention 'Bletch:'.
16This is not necessary if you're using gnu cpp.
17
18in short:
19./Configure -d
20remove SDBM_File from config.sh
21./Configure -S
22make depend
23remove Bletch: from makefile and x2p/makefile
24make
25
26Other than that, perl should build without problems. There are some
27technical comments in hints/beos.sh.
28
29OS RELEASE-SPECIFIC NOTES
30-------------------------
31
32PR1/PPC:
33See R3/X86. Same bug, different form.
34
35PR2/PPC:
36Signals are somewhat unreliable, but they can work. Use caution.
37The POSIX module is still somewhat buggy.
38
39R3/X86:
40Under R3 x86, there are some serious problems with the math routines
41such that numbers are incorrectly printed. This causes problems with
42modules that encode their version numbers - in particular, IO.pm will
43probably not work properly. This should be fixed under R3.1.
44
45The problem has manifested itself if you see something similar to the
46following during the compile:
47
48cc -c -I/usr/local/include -O -DVERSION=\"1.1504\" -DXS_VERSION=\"1.1499999999\" -fpic -I../.. IO.c
49(lots of 9's are the indication of the problem.)
50
51In the meantime, you can use the following workaround:
52
53make perl
54cd ext/IO
55cc -c -I/usr/local/include -O -DVERSION=\"1.1504\" -DXS_VERSION=\"1.15\" -fpic -I../.. IO.c
56cd ..
57make
58
59(Substitute the correct numbers if IO has been updated.)
60
61R3/PPC-
62There's math problems, but of a different kind. In particular,
63perl -e 'print (240000 - (3e4<<3))' gives a non-zero answer.
64I'm looking into this. There is no workaround as yet. Hopefully,
65this will be fixed in R3.1.
66
67CONTACT INFORMATION
68-------------------
69If you have comments, problem reports, or even patches or bugfixes (gasp!)
70please email me.
71
721 May 1998
73Tom Spindler
74dogcow@merit.edu
75