This is the mail archive of the gcc-patches@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: C++ lexical analysis rework



On Jan 16, 2005, at 11:15 PM, Eric Botcazou wrote:


Apple hasn't seen a problem with this yet. I'm still trying to figure
out why. Possibly, as Zack suggested, -gused takes care of most of the
cases where it might hurt.

It was me, actually. Note that I only said that -gused might hide the PCH
failures in the C++ testsuite

yes.


because the testcases are very simple and,
consequently, -gused might throw away all the debug info for them.

BINCL/EINCL stabs are not useful in -gused mode (because it is an alternative space optimization)


Maybe DBX_USE_BINCL is simply useless in practice.

They are very useful in non -gused mode (aka as -gfull). Which is default on most of the (if not all) platforms that use stabs. Even though -gused is default we, here at Apple, support -gfull mode also. There are many apps that benefit from -gfull mode. So this failures are important for us.


Running GDB testsuite with -gfull (or -g -fno-eliminate-unused-debug-symbols) will reveal actual damage.

-
Devang


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