| Subject: | Re: [PATCH] Check precision rather than machine mode in STRIP_NOPS and STRIP_SIGN_NOPS |
|---|---|
| From: | Adam Nemet |
| Date: | Tue, 9 Jun 2009 16:08:54 -0700 |
Richard Guenther writes: > 2009/6/9 Adam Nemet <anemet@xxxxxxxxxxxxxxxxxx>: > > Richard Guenther writes: > >> Indeed, you should do the TYPE_PRECISION check only on > >> INTEGRAL_TYPE_P types and keep the mode check on the rest. > > > > Can't we also check POINTER_TYPE_P and OFFSET_TYPE? > > We can, but I don't think we ever have precision differences there. Nor have we mode differences, no? I thought we preferred precision over mode in the middle-end. Sorry I don't want to overpush this but that is the sense I got from Joseph. I verified that a mode attribute in connection with a pointer also updates the precision of the pointer type. Adam |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | [gcc-in-cxx] Merge from trunk, Ian Lance Taylor |
|---|---|
| Next by Date: | Re: [PATCH] Add new built-in: __builtin_unreachable(), Richard Henderson |
| Previous by Thread: | Re: [PATCH] Check precision rather than machine mode in STRIP_NOPS and STRIP_SIGN_NOPS, Richard Guenther |
| Next by Thread: | Re: [PATCH] Check precision rather than machine mode in STRIP_NOPS and STRIP_SIGN_NOPS, Richard Henderson |
| Indexes: | [Date] [Thread] [Top] [All Lists] |