ANN: New rtems-4.8 toolchain rpms available

Ralf Corsepius ralf.corsepius at rtems.org
Mon Jun 18 01:20:00 CDT 2007


On Sat, 2007-06-16 at 13:06 -0600, e.stiebler wrote:
> Ralf Corsepius wrote:
>  > On Fri, 2007-06-15 at 18:04 -0600, e.stiebler wrote:

> Is there a chance, that there is just something wrong with the rpm file ?

Well, there had been a bug in RH's rpm which had caused rpm to screw up
on sizes, so I would not want to exclude this possibility. However as
far as I can tell the culprit of this increase in sizes is precompiled
headers having been added to libstdc++'s headers in gcc-4.2.0 (grep for
*.gch files inside of the rpms.)

Deactivating building precompiled headers for libstdc++ (adding
--disable-libstdcxx-pch to configure when configuring gcc) brings
m68k-rtems-*-gcc-c++'s rpm sizes down to orders we've had with
gcc-4.1.*x (~25MB instead of ~150MB)

For the moment, I launched a new spin of m68k*-gcc*rpms with
--disable-libstdcxx-pch added, which should land in the public repos
within the next couple of hours.

However, I think, we need to decide on a better long term policy on
shipping precompiled libstdc++ headers. I am not sure about what to do
about them. 

ATM, I am inclined to think precompiled libstdc++ headery aren't really
useful for toolchains targetting embedded os such as RTEMS.

> The installed stuff in /opt/rtmes ...  is not as big a the .rpm file.
> (just had a quick look, probably I missed something)
:)

Ralf





More information about the rtems-users mailing list