> Hello,
>
> I've just done a nightly build on b86 tree (x86 platform). The mail_msg
> file generated at the end of the build reported no errors in any of the
> build sections. However, there were a bunch of errors listed under
> "Build noise differences". Should I care about those? What are these
> build noise differences?
In general, "yes," you should care about this. By "care," I mean that you
should pay attention to new noise (as indicated by a ">" prefix, because
diffs point from left-to-right, old-to-new), and understand what it means.
Because noise is only reported transiently, the first time you see it will
be your only chance to fix it.
> ==== Build noise differences (DEBUG) ====
>
> 1,20c1
> < current : array[-2] of int
> < current : const int
> < current : function() returning int
> < current : function(pointer to struct {pointer to struct {..}
> gh_depend
> s, pointer to struct rt_map {..} gh_ownlmp, pointer to struct lm_list {..}
> gh_ow
> nlml, unsigned int gh_refcnt, unsigned int gh_flags}, pointer to const char)
> ret
> urning int
> < current : function(unsigned long long, unsigned long long, unsigned
> shor
> t, pointer to union {union {..} inv_buf, union {..} cap_val_hw1_buf, union
> {
> ..} cap_val_sf1_buf}) returning pointer to const int
> < current : function(unsigned long, unsigned long, unsigned short,
> pointer
> to union {union {..} inv_buf, union {..} cap_val_hw1_buf, union {..}
> cap_va
> l_sf1_buf}) returning pointer to const int
> < current : function(unsigned short, unsigned char, int, pointer to
> union
> {array[12] of char buf}) returning pointer to const int
> < current : function(unsigned short, unsigned char, int, pointer to
> union
> {array[22] of char buf}) returning pointer to const int
> < current : int
> < previous: const int : "../common/symbols.c", line 98
> ... truncated.
Specifically, these diffs indicate that you fixed something, so now the
build isn't complaining any more.
--Mark
