Any objections to setting this 'true' by default? It seems we have three people okay with it. If none are brought up, I or anyone can switch it.
-Nathan On Sat, Feb 7, 2009 at 5:43 PM, Gregory Shimansky <[email protected]> wrote: > On 8 February 2009 Nathan Beyer wrote: >> Isn't "/usr/lib/libstdc++.so.6" generally a symlink to the specific >> version folder that's currently the default? > > Nope. It depends very much on the distro layout. The most likely place to find > libc is /lib where you can execute (you can actually execute this > lib) /lib/libc.so* and see the version. > >> Regardless, moving to v6 as the default seems to make sense to me - >> this seems to be the default on the latest distribution releases that >> I've seen. > > With this I agree. Version 6 has been the default for a time I can't remember > and those who still use v5 might as well switch after all. > >> On Sat, Feb 7, 2009 at 5:28 AM, Gregory Shimansky <[email protected]> > wrote: >> > On 7 February 2009 Mark Hindess wrote: >> >> Perhaps we should make use.libstdc++6=true the default? Or maybe even >> >> try a little harder to make a better guess at the default with a patch >> >> like the one appended? >> > >> > Hmm it is not this easy. The library libstdc++ belongs to gcc, and its >> > location depends on gcc version (gcc may be even installed in a >> > non-standard location). Guessing it is not easy without configure script. >> > Here's what I get when I execute "locate libstdc++.so.6" on Gentoo Linux: >> > >> > /usr/lib64/gcc/x86_64-pc-linux-gnu/3.4.6/32/libstdc++.so.6 >> > /usr/lib64/gcc/x86_64-pc-linux-gnu/3.4.6/32/libstdc++.so.6.0.3 >> > /usr/lib64/gcc/x86_64-pc-linux-gnu/3.4.6/libstdc++.so.6 >> > /usr/lib64/gcc/x86_64-pc-linux-gnu/3.4.6/libstdc++.so.6.0.3 >> > /usr/lib64/gcc/x86_64-pc-linux-gnu/4.1.2/32/libstdc++.so.6 >> > /usr/lib64/gcc/x86_64-pc-linux-gnu/4.1.2/32/libstdc++.so.6.0.8 >> > /usr/lib64/gcc/x86_64-pc-linux-gnu/4.1.2/libstdc++.so.6 >> > /usr/lib64/gcc/x86_64-pc-linux-gnu/4.1.2/libstdc++.so.6.0.8 >> > >> > The naming of path like "x86_64-pc-linux-gnu" is very platform dependent >> > as well. >> > >> > Someone might even want to use Intel compiler to build Harmony, I recall >> > icc used to work at some point of time and produced quite an optimized >> > code. >> > >> >> Index: working_classlib/make/properties.xml >> >> =================================================================== >> >> --- working_classlib/make/properties.xml (revision 741852) >> >> +++ working_classlib/make/properties.xml (working copy) >> >> @@ -313,6 +313,12 @@ >> >> </condition> >> >> <property name="hy.portlib.stubs" value="false" /> >> >> >> >> + <condition property="use.libstdc++6" value="true"> >> >> + <and> >> >> + <isset property="is.linux" /> >> >> + <available file="/usr/lib/libstdc++.so.6" type="file" /> >> >> + </and> >> >> + </condition> >> >> <condition property="hy.libstdc++.suffix" value=".libstdc++6"> >> >> <isset property="use.libstdc++6"/> >> >> </condition> >> >> Index: common_resources/make/properties.xml >> >> =================================================================== >> >> --- common_resources/make/properties.xml (revision 741852) >> >> +++ common_resources/make/properties.xml (working copy) >> >> @@ -69,6 +69,12 @@ >> >> <property name="hy.test.timeout" value="900000" /> >> >> <property name="hy.test.vmargs" value="" /> >> >> >> >> + <condition property="use.libstdc++6" value="true"> >> >> + <and> >> >> + <isset property="is.linux" /> >> >> + <available file="/usr/lib/libstdc++.so.6" type="file" /> >> >> + </and> >> >> + </condition> >> >> <condition property="hy.platform.suffix" value=".libstdc++6"> >> >> <isset property="use.libstdc++6"/> >> >> </condition> >> > >> > -- >> > Gregory > -- > Gregory >
