------- Additional Comments From row at telocate dot com 2010-04-14 22:56 ------- The point is, I should be able to define a constant-valued symbol, e.g. "mySym = 0x100;" *outside* (and before) the MEMORY command, then use the symbol (i.e. a meaningful name instead of a bare literal) inside the MEMORY command, *and* in logically related expressions in the rest of the script. As Alan Modra already pointed out, this bug is just the result of an oversight in the existing implementations, wherein the (informal) specification was not completely implemented. The application fails to do what its documentation says it should, and what common sense suggests and good programming practice requires; and this specific shortcoming raises a significant barrier to writing more-maintainable linker scripts.
Dave Korn, a long time ago, provided a detailed description of this issue and sketched a solution: http://www.cygwin.com/ml/binutils/2004-03/msg00540.html. My YACC skills are very rusty, or else I might try and fix this myself. I'm developing embedded systems where this (unimplemented) feature would be quite handy. I need to use the MEMORY command to build an accurate model of the target memory layout, that will drive the rest of linking process, not to mention defining symbols that will drive bootloading, memory initialization, etc. Yes I might be able to use some sort of kludged work-around; but why not just fix the bug? I'm dismayed that something as obvious and potentially useful as this feature has gone un-addressed for so long... Row Reith, r...@telocate.com [rr] -- What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | http://sourceware.org/bugzilla/show_bug.cgi?id=4643 ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is. _______________________________________________ bug-binutils mailing list bug-binutils@gnu.org http://lists.gnu.org/mailman/listinfo/bug-binutils
