| Age | Commit message (Expand) | Author |
|---|
| 2004-06-09 | In FUNCTION_ARG_ADVANCE(), do not pad the register position counter if the | Miod Vallat |
| 2004-06-09 | More meaningful formulation of the __va_stk initialization in | Miod Vallat |
| 2004-05-19 | When passing a variable of 32 bytes or larger, as argument to a function, | Miod Vallat |
| 2004-05-19 | In __builtin_saveregs(), initialize __va_stk to a correct value. | Miod Vallat |
| 2004-05-17 | Make FUNCTION_ARG_ADVANCE() really match FUNCTION_ARG() register usage. | Miod Vallat |
| 2004-05-04 | Do not disable -fcaller-saves anymore. | Miod Vallat |
| 2004-04-08 | In GO_IF_LREGITIMATE_ADDRESS(), use RTX_OK_FOR_BASE_P() to check for a | Miod Vallat |
| 2004-04-06 | When optimizing a logical arithmetic operation, which operands are both | Miod Vallat |
| 2004-03-19 | Disable caller-save (which is enabled at -O2 or -Os) as it suffers from a | Miod Vallat |
| 2004-02-27 | Teach gcc the proper "set" syntax for __attribute__ ((alias)) to work | Miod Vallat |
| 2003-10-01 | Introduce a new target switch, -mmemcpy (and -mno-memcpy). | Miod Vallat |
| 2003-08-08 | On some architectures, being too optimistic in preserve_subexpressions_p(), | Miod Vallat |
| 2003-08-07 | Disable the so-called optimized memcpy functions for small (less than a few | Miod Vallat |
| 2003-08-01 | A working gcc 2.95/m88k compiler, for some low standard value of working. | Miod Vallat |
| 2001-01-29 | 2.95.3, test release 2 | Marc Espie |
| 2000-09-24 | Synch with FSF stable tree | Marc Espie |
| 2000-06-05 | Synch with 2.95.3 (pre-rel) | Marc Espie |
| 1999-08-23 | Update to gcc 2.95.1, official release. | Marc Espie |
| 1999-07-22 | 990718, very close to release. | Marc Espie |
| 1999-05-26 | Configuration fragments for arches that have not been committed to egcs | Marc Espie |
| 1999-05-26 | egcs projects compiler system | Marc Espie |
