Maho NAKATA wrote:
Hi Stefan, and *
I could provide DEV300_m33 with Stefan's patch. Very stable.
No hack for regcomp anymore.
Thank you very much,
-- Nakata Maho http://accc.riken.jp/maho/
This news really makes me happy. :)
Is there a relevant issue and/or CWS through which we should integrate
the patch (also for the gcc_linux_powerpc bridge, I would say)? If not,
let me know, and I will take care.
However, if anybody happens to know whether Mac OS X offers a function
to do instruction memory synchronization, it would be even better to
simply call that function instead of fiddling with magic hand crafted
assembler instructions.
On Solaris SPARC,
<http://qa.openoffice.org/issues/show_bug.cgi?id=76808> "Use
sync_instruction_memory(3C) in Solaris cpp_uno bridges" got integrated
in DEV300m33, replacing hand crafted assembler magic for instruction
memory synchronization with a function provided by Solaris. In the
past, we did have mysterious build breakers on Solaris SPARC, and I am
now thinking that those might have been due to errors in the instruction
memory synchronization. Ause, are there still any hacks in the code to
work around those mysterious problems (IIRC, you did put hacks into the
code to address that)? If yes, would be interesting to know whether
builds would work reliably now without those hacks.
-Stephan
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
