docs/known-issues.txt
changeset 2069 366bd2b22675
parent 2032 278df575fe4b
     1.1 --- a/docs/known-issues.txt	Thu Jul 22 23:33:17 2010 +0200
     1.2 +++ b/docs/known-issues.txt	Thu Aug 05 17:59:51 2010 +0200
     1.3 @@ -102,10 +102,10 @@
     1.4      libtool.m4: error: problem compiling FC test program
     1.5  
     1.6  Explanations:
     1.7 -  The gc build procedure tries to run a Fortran test to see if it has a
     1.8 +  The gcc build procedure tries to run a Fortran test to see if it has a
     1.9    working native fortran compiler installed on the build machine, and it
    1.10 -  can't find one. A native Fortran compielr is needed (seems to be) to
    1.11 -  build the Fortran frontend of the cross-compiler.
    1.12 +  can't find one. A native Fortran compiler is needed (seems to be neede)
    1.13 +  to build the Fortran frontend of the cross-compiler.
    1.14    Even if you don't want to build the Fortran frontend, gcc tries to see
    1.15    if it has one, but fails. This is no problem, as the Fortran frontend
    1.16    will not be built. There is nothing to be worry about (unless you do
    1.17 @@ -119,3 +119,21 @@
    1.18    None needed, it's a spurious error.
    1.19  
    1.20  --------------------------------
    1.21 +Symptoms:
    1.22 +  gcc barfs because it is "unable to detect the exception model".
    1.23 +
    1.24 +Explanations:
    1.25 +  On some architectures, proper stack unwinding (C++) requires that
    1.26 +  setjmp/longjmp (sjlj) be used, while on other architectures do not
    1.27 +  need sjlj. On some architectures, gcc is unable to determine whether
    1.28 +  sjlj are needed or not.
    1.29 +
    1.30 +Fix:
    1.31 +  None so far.
    1.32 +
    1.33 +Workaround:
    1.34 +  Trying setting use of sjlj to either 'Y' or 'N' (instead of the
    1.35 +  default 'M') in the menuconfig, option CT_CC_GCC_SJLJ_EXCEPTIONS
    1.36 +  labelled "Use sjlj for exceptions".
    1.37 +
    1.38 +--------------------------------