Re: Fishing for advice...

Ove Kaaven (ovek@arcticnet.no)
Wed, 23 Jun 1999 17:36:32 +0200 (CEST)


On Wed, 23 Jun 1999, Shaw Terwilliger wrote:

> > OK. I did a compile and it wrecks here... Anyone have a clue?
> > .............................................................................
> > make[3]: Entering directory `/home/tw/abisource/abi/src/wp/ap/unix'
> > Building with [LicensedTrademarks:Off Debug:Off].
> > ap_UnixApp.cpp:
> > ap_UnixApp.cpp: In method `UT_Bool AP_UnixApp::initialize()':
> > ap_UnixApp.cpp:99: warning: unused variable `UT_Bool bVerified'
> > ap_UnixApp.cpp: In method `UT_Bool AP_UnixApp::parseCommandLine()':
> > ap_UnixApp.cpp:510: no matching function for call to `AP_BuiltinStringSet::dumpBuiltinSet (char[13])'
> > make[3]: *** [/home/tw/abisource/abi/src/../src/Linux_2.2.5-15_i386_OBJ/obj/ap_UnixApp.o] Error 1
> > make[3]: Leaving directory `/home/tw/abisource/abi/src/wp/ap/unix'
> > make[2]: *** [build] Error 2
> > make[2]: Leaving directory `/home/tw/abisource/abi/src/wp/ap'
> > make[1]: *** [build] Error 2
> > make[1]: Leaving directory `/home/tw/abisource/abi/src/wp'
> > make: *** [build] Error 2
> > .............................................................................
>
> I can't really tell here... it looks like you might have a half baked build.
> You said you did try a recent CVS checkout; you might try it again if
> you were using "cvs update" to grab it. You may also want to do "make clean"
> or "make realclean" to see if it cleans out old objects. I just checked
> a tree out (during writing this message) on one of our faster machines and
> it built all the way through.

Strange thing. I just did a few makes in a row a few days ago (maybe with
a few different options, like some builds with ABI_OPT_DEBUG=1 and some
without, and with two different kernels on the same source tree etc,
without make clean in between of course), and suddenly that error came up.
Then I looked for dumpBuiltinSet, but I didn't see it anywhere in the
relevant .h files. So I just commented the line out, since it appeared to
be a debugging feature anyway... but now it's in the .h file (again?).
Ghost in the machine? Of course I could have missed it, but could the
compiler too?



This archive was generated by hypermail 1.03b2.