[patch/rfc] Implement ptrace target vector inheritance

[Andrew Cagney]

> Hello,
> 
> This came from an thread between myself and MarkK (I can't find the url).
> 
> The current native code constructs its target vector using a convoluted combination of #ifdefs and global functions.
> 
> This patch replaces all that with what is effectively a runtime inheritance structure:
> 
>     inf-ptrace is-a inf-child is-a target
> 
> I've updated/tested NetBSD/PPC, with a tweak to nbsdppc-nat. that further extends the above with:
> 
>     nbsdppc-nat is-a inf-ptrace
> 
> there were no regressions.
> 
> The intent is for all natives to eventually migrate to this.
> 
> Comments?
> 
> Right now I'm more interested in high-level stuff - the idea and the strategy.  I'm sure the implementation will evolve as more natives come on board, and as I eliminate a few remaining globals.
> 
> I'll leave this for a week,

I've checked in the attached, per suggestions.  I've left out the new 
commands (keeping new functionality separate from refactoring).  I've 
also left a few things like call_ptrace in infptrace.c.

committed,
Andrew

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: diffs
URL: <http://sourceware.org/pipermail/gdb-patches/attachments/20040912/43603722/attachment.ksh>