This is the mail archive of the
mailing list for the libstdc++ project.
Re: [libstdc++] Port-specific symbol exports, porting notes
On Mon, Jul 07, 2003 at 11:23:11PM -0500, Benjamin Kosnik wrote:
> >Maintaining two copies of the almost-same file would be a PITA.
> >Somebody wise in the ways of perl could probably automate the production
> >of linker-map.gnuold from linker-map.gnu, but that person is not me.
> It's probably not worth it at this stage to go through hoops for
> per-host symbol versioning, since it's not being used, even though it's
> something that is desirable in the long run.
Actually, I might try and use it for the instantiations that seem to be
needed on mips-elf, since it looks like no other platform needs those.
> >Also, note that if anybody wants to actually /use/ the per-port symbol
> >exporting functionality, they /have/ to use 2.14 for the parser fix.
> >Thoughts? Other ideas?
> Instead of defaulting to the new map, why don't you default to gnu-2.12,
> and then have gnu-2.14 when you see it can be used? It's just a semi
> colon. Really, it's a drag, but we have to put up with worse, and this
> will allow us to revisit this closer to the release.
It wasn't the semicolon thing, even though that continues to trip us up.
It's the new functionality. But we should be able to go with 2.older anyhow,
as long as the per-port functionality is not in actual use. (I'll put
the semicolons back but leave the marker hook for additional symbols.)
There may be reasons to require something higher than 2.12 anyhow. Did
you see Andrew's note about the assembler bug?
If ye love wealth greater than liberty, the tranquility of servitude greater
than the animating contest for freedom, go home and leave us in peace. We seek
not your counsel, nor your arms. Crouch down and lick the hand that feeds you;
and may posterity forget that ye were our countrymen. - Samuel Adams