This is the mail archive of the gcc-regression@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: bootstrap failure on Darwin



On Jul 24, 2006, at 1:47 PM, Daniel Jacobowitz wrote:


On Mon, Jul 24, 2006 at 01:40:07PM -0700, Geoffrey Keating wrote:
Bootstrap comparison failure!
./alloc-pool.o differs
./attribs.o differs

The following changes look most likely to be causing the problem, my
guess is the first one:

--------------------------------------------------------------------- ---
r115704 | drow | 2006-07-23 19:59:36 -0700 (Sun, 23 Jul 2006) | 4 lines
PR debug/27473
* dbxout.c (output_used_types_helper, output_used_types): New.
(dbxout_symbol): Call output_used_types.

I'm going to make a wild guess here... could you check and see if this has caused stabs to be output in a non-deterministic order, based on the hash table walk? A different ordering is used for dwarf2, since the used_types_hash walk doesn't create the DIEs, just mark them.

I just confirmed, that reverting Daniel's patch, fixes the regression.


Thanks,
Andrew Pinski


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]