Hmm, I could swear that I have it in my .def file. I'll have to check it later, but it definately linked ok (made a 62k prc which is abour right), but then died with a bus error. Yes, floating point works with the fixed libgcc, yay! Now I'm fighting through multiple segments. Another app I've tried for multisegment "sort-of" works. I can move some functions to an alternate section, but not others (not launch related ones either). If I move some of them I get Chunk Underlocked error messages of all things. A friend told me there's a lot of talk about this on massena, but I can't get massena to behave half the time (last night it would display threads, but timed out displaying messages). > Subject: Re: Prc-tools v2 - Spurious code section ignored? > From: John Marshall <[EMAIL PROTECTED]> > Date: Sun, 12 Mar 2000 22:19:55 -0800 (PST) > X-Message-Number: 37 > > Chris DiPierro <[EMAIL PROTECTED]> wrote: > > testapp:calcs: warning: spurious code section ignored > > > > I don't see why it's calling this spurios as I use it and it seems to link > > it correctly > > Actually I'm surprised it links correctly. If you're actually calling > any of the functions in the calcs section I would expect undefined > references to __text__calcs at link time (if you're using multigen to > generate the assembly language stub from the same definition file that > build-prc is reading). > > > Is there a way to get gcc ot > > elaborate a little? > > Well, I guess I could document the error message a little... :-) > > This error message is produced when there is a code section in your > executable which is not listed in your definition file. (This is an > error because build-prc can't know which resource number to map the > section to.) > > The collection of names which appear in section attributes anywhere > in your project and the list of names in multiple code clauses in your > definition file need to be identical. Multigen and build-prc require > this so that they can generate the cross-section dispatch code correctly, > and you'll get errors if they are not identical: "spurious code section > ignored" for a section in an attribute somewhere but not in the .def > file, and "unknown code section" for a name in the .def file but not > in any attributes. > > (So the floating point's working now? :-)) > > John > -- For information on using the Palm Developer Forums, or to unsubscribe, please see http://www.palm.com/devzone/mailinglists.html
