Reply at top for emphasis: You are definitely NOT butting in. At this point I'll take help from anyone. On Tue, Jan 6, 2015 at 6:50 AM, Brian Drummond <brian@xxxxxxxxxxxxxxxxxx> wrote: > On Tue, 2015-01-06 at 06:18 -0600, Cyd Haselton wrote: >> On Tue, Jan 6, 2015 at 3:34 AM, Andrew Haley <aph@xxxxxxxxxx> wrote: > > >> Regarding the generated map file for the gcc build, is it possible >> that either a) i'm looking in the wrong place or b) I need to add the >> generate map file command to Makefiles in the gcc build subdirs and/or >> different places in the gcc Makefile? >> >> I'll look int0 b). Let me know about a) > > Hope I'm not butting in here, but I saw something about mapfile > generation elsewhere... > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64492 > comment 15. > > It suggests that you can re-run the relevant ld command (which you can > find in the build log) with the --create-map-file argument and it should > do what you want. > > I feel your pain ... building gcc can be frustratingly difficult, if > you're slightly off the beaten path. > > -- Brian > I'll look into this...thanks. Could you tell me how --create-map-file differs from -Wl,-Map=mapfile? To recap: The problem is that the libgcc link command for 4.9 doesn't generate a mapfile because (and this is speculation on my part) the link command generates the error thus preventing the mapfile creation. I was able to generate a mapfile for the 4.8 libgcc link command, but there is no reference to dlopen in it. I went back and generated a mapfile for the gcc build in both versions, and dlopen is referenced in both by the same file
