From: "John Merryweather Cooper" <jmcoopr@webmail.bmi.net>
Subject: Porting JGNAT public version 1.0p to OS/2
Date: 2000/04/03
Date: 2000-04-03T00:00:00+00:00 [thread overview]
Message-ID: <sehps9eueop30@corp.supernews.com> (raw)
Well, I've had to "hack" the makefile, but I've gotten a good part of the
package to compile. I believe some of the bugs in the EMX implementation of
make are showing through since the "clean" makefile wouldn't pass the .c
source files to the compiler ($< was returning " "). JVMLIST.EXE,
JVM2ADA.EXE, etc. fail on linking--it seems they cause a compile/link of
osint.ads/.adb which in turn wants a-adaint.c. But manually adding
a-adaint.o to the module list cause "duplicate references errors. For each
module that fails to link, I get three undefined symbol errors pointing to
functions whose source (potentially) is found in a-adaint.c:
_to_canonical_file_list_init
_to_canonical_file_list_next
_to_canonical_file_list_free
Since these utilities are supposed to only rely on the standard GNAT
library, are these symbols supposed to be defined in libgnat.a (but are not
in the EMX-OS/2 version--so that I need to port them to the standard
library) or is something else going haywire? :)
Back to the slave pits . . .
jmc
reply other threads:[~2000-04-03 0:00 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
replies disabled
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox