I reverted this in r238048. I am working on an idea to try to get most of the savings and still be bug compatible with the current ICF implementation.
On 22 May 2015 at 16:02, Nico Weber <[email protected]> wrote: > On Wed, May 20, 2015 at 7:41 PM, Rafael Espíndola > <[email protected]> wrote: >> >> > This regressed binary size of Chromium on Linux with 1300 kB. Our >> > theory is that the linker [1] is dead-stripping more effectively with >> > unique section names for some reason. >> >> It is ICF, which is why it didn't reproduce with the clang binary. >> >> I reduced this to https://sourceware.org/bugzilla/show_bug.cgi?id=18440 >> >> I will think about it a bit. > > > Have you had a chance to think about this? Should this not be done by > default until an ld fix is deployed in most places? > >> >> >> Cheers, >> Rafael > > _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
