| 1 |
On Tue, 2007-06-12 at 11:27 +0200, Fabian Groffen wrote: |
| 2 |
> I've hit this thought a few times now, but every time hacked around it |
| 3 |
> with assumptions. However, maybe it is a good idea to make the linker |
| 4 |
> (and perhaps also the compiler?) visible from ebuilds? Many "Darwin" |
| 5 |
> patches are linker patches. |
| 6 |
> |
| 7 |
> I was thinking of things like |
| 8 |
> |
| 9 |
> linker_GNU |
| 10 |
> linker_Darwin |
| 11 |
> linker_Solaris |
| 12 |
> |
| 13 |
> or? |
| 14 |
> |
| 15 |
> [[ ${LINKER} == binutils-2.17* ]] |
| 16 |
> [[ ${LINKER} == odcctools* ]] |
| 17 |
> |
| 18 |
> or any other mix inbetween? Does it make sense to anyone? Thoughts? |
| 19 |
|
| 20 |
What about having some intelligent functions in toolchain-funcs.eclass ? |
| 21 |
|
| 22 |
Because when GNU ld once works on platforms it does currently not, we |
| 23 |
might want to switch... |
| 24 |
|
| 25 |
I'm thinking of somewhat like that: |
| 26 |
|
| 27 |
tc-have-soname() { |
| 28 |
if [[ $($($(tc-getCC) -print-prog-name=ld) --version 2>/dev/null) == *GNU* ]]; then |
| 29 |
# ld used by CC is GNU ld, thus we have -soname |
| 30 |
return 0 |
| 31 |
fi |
| 32 |
case "${CHOST}" in |
| 33 |
*-hpux*) return 0 ;; # assume ldwrapper maps -soname to +h |
| 34 |
esac |
| 35 |
return 1 |
| 36 |
} |
| 37 |
|
| 38 |
and/or eventually: |
| 39 |
|
| 40 |
tc-soname() { |
| 41 |
if [[ $($(tc-getCC) -print-prog-name=ld) --version 2>/dev/null) == *GNU* ]]; then |
| 42 |
# ld used by CC is GNU ld, thus we have -soname |
| 43 |
echo "-soname=${1}" |
| 44 |
else |
| 45 |
case "${CHOST}" in |
| 46 |
*-hpux*) echo "+h ${1}" ;; # no mapping of -soname to +h in ldwrapper |
| 47 |
esac |
| 48 |
fi |
| 49 |
true |
| 50 |
} |
| 51 |
|
| 52 |
/haubi/ |
| 53 |
|
| 54 |
-- |
| 55 |
gentoo-alt@g.o mailing list |
Archives