On Thu, 18 Apr 2013 05:30:28 +0200, Jonathan Wakely-4 [via gcc] <ml-node+s1065356n931358h45@xxxxxxxxxxxxx> wrote: > On 18 April 2013 03:52, Hibou57 wrote: >>> I will try with `--disable-sjlj-exceptions` >>With disable the `--disable-sjlj-exceptions`, it works, for the C >> languageso far. Seems so there is an issue with the SJLJ option. Should >> I report itas a bug? > > No, you should stop using options that you don't understand What? > , becausethey are causing your problems. You've probably not told the > runtimelinker how to find the dynamic libraries that you built to be > incompatible. This has nothing to do with the loader, but with the generation of a library. This is not a runtime link error, this is a link error. This occured while building `cc1`, not while running cc1. See the attached http://gcc.1065356.n5.nabble.com/file/n931332/logs-and-others.tar.xz > > I build with --prefix and --enable-languages and usually nothing else. Actually, it don't for me (and I'm not the only one). > > Enable features if you *know* you need them, not by guessing at whatyou > think you need. I know what's SJLJ. It compiled fine with SJLJ exception using previous versions GCC of in the past, but now it don't. I'm aware SJLJ is not part of the LSB, but if it's provided as an option, it should be working as expected. Sorry, I won't care about your “No” above, and will wait for others suggestions. If there are none, I will submit it as a bug, as you seems to not understand the issue, the reason you gave to not submit it is not relevant. -- Yannick Duchêne -- View this message in context: http://gcc.1065356.n5.nabble.com/Inconsistant-exception-API-libgcc-s-and-cc1-don-t-agree-tp931332p931359.html Sent from the gcc - Help mailing list archive at Nabble.com.
