From mboxrd@z Thu Jan 1 00:00:00 1970 From: Kevin Buettner To: Andrew Cagney , Kevin Buettner Cc: gdb-patches@sources.redhat.com Subject: Re: [PATCH] Conditionally include nlist.h in solib-legacy.c Date: Fri, 23 Mar 2001 08:40:00 -0000 Message-id: <1010323164002.ZM20219@ocotillo.lan> References: <1010320030437.ZM8757@ocotillo.lan> <3ABB762D.9867D7DC@cygnus.com> X-SW-Source: 2001-03/msg00435.html On Mar 23, 11:13am, Andrew Cagney wrote: > Looking back at this. I don't know that solib-legacy should ever be > appearing in the list of TDEPFILES. Given it can only work in native > builds, I think it should only be included in the NATDEPFILES. You're right. I put it in the TDEPFILES list for two reasons: 1) Laziness. (It was easier to just search for solib-svr4.o and add it to each line containing this file.) 2) To avoid link errors. My concern regarding #2 was as follows: If you put solib-legacy.o in host A's NATDEPFILES and then you configure an A cross B debugger where B doesn't have solib-svr4.o in its TDEPFILES, you'll end up with a link error when you're done. I've just recently learned, however, that this won't happen since NATDEPFILES is commented out when building cross debuggers. Since problem #2 is disposed of, all I need to do is overcome #1. (I'll try to get around to looking at it over the weekend.) > I suspect this was the real reason for the compile problems. Probably so. However, I also think that the solib-legacy code is better off using the autoconf supplied feature tests.