This is the mail archive of the gcc@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]

PCH distcc improvements


Neil, Geoff:

This patch enables PCH work better in the presence of distcc (and things like it). First some brackground, PCH is turned off during -E and -save-temps compiles presently. For things like distcc to work well, where well is defined as not recompiling the same header that has been compiled with PCH over and over, it is desirable for PCH to be able to be turned on in these cases. This patch add a new flag -fpch-preprocess to do this.

The questions are, do people like the name of the option?

Should there be only one flag? If so, should it be flag_pch_preprocess, or CPP_OPTION(,pch_preprocess)?

Can I use flag_pch_preprocess in c-pch.c, or is there a better way to do this?

Is the use of cpphash.h in c-pch.h really bad? If so, what is the best way to avoid it (for injecting #include into the .i/.ii file)?

diffs against something (Apple 3.3 tree) close to 3.3:

Attachment: pch-distcc.diffs
Description: Binary data


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