libdeprecated ? libfixbadcoding ?

On Thu, May 15, 2008 at 11:28 AM, Roger A. Faulkner
<Roger.Faulkner at sun.com> wrote:
>
>> Date: Mon, 12 May 2008 09:59:10 -1000
>> From: Joseph Kowalski <jek3 at sun.com>
>> Subject: Re: PSARC/2008/309 - expunge synonyms.h
>> To: John Plocher <John.Plocher at sun.com>
>> Cc: "Roger A. Faulkner" <Roger.Faulkner at sun.com>, psarc-ext at 
>> sac.sfbay.sun.com,
> Bart.Smaalders at sun.com, Rod.Evans at sun.com, Darrin.Johnson at sun.com
>>
>> John Plocher wrote:
>> >
>> > I'd go out on a limb and suggest that, since this is an obvious case of
>> > not following the instructions, and since we are quickly moving towards
>> > a major release classification for Nevada/ON, which gives one (and
>> > probably
>> > only one) honeymoon opportunity to clean up such things, we should simply:
>> I clearly heard that the same compatible rules we to be inplace for the
>> library
>> interfaces.  The semi-major release classification was for utilities
>> (and perhaps
>> administration - not clear to me).  These are library interfaces, hence
>> I think
>> we need to think in Minor release terms.  Maybe this will be clearer on
>> Wednesday.
>> >     Delete the header files as above,
>> They are in source.  Who cares?  They are implementation. Certainly not
>> architecture.
>> >     Create a LD_PRELOAD shim that restores these symbols for the
>> >     things that need it, and
>> I need to think about this more, but I don't think this works reliably if
>> third-party libraries are in the mix.  At this point, its just a nit.
>> >     Eliminate all of the SUNWprivate leading-underscore symbols
>> >     involved in the synonyms scheme.  Period.
>> Comma or semi-colon perhaps?  (See: how Major is Major)
>> > If something breaks, let it use the preload shim.  End of story.
>> The ARC shouldn't be designing this.  *If* there is a need for a
>> shim it should be part of the proposal, including any caveats.
>> Like I I said, I'm not sure we need a shim (TBD).
>> >
>> >   -John
>> - jek3
>
> Regardless of the merits of the Draconian approach (which I like!),
> there is merit to the idea of providing a preloadable shim library
> that provides the old leading-underscore symbols and jumps directly
> (via the PLT and GOT) to the corresponding non-underscore symbols.
>
> I created it yesterday and named it:
>
>    /lib/libshim.so.1
>    /lib/64/libshim.so.1
>
> So I'm adding this item to the official proposal.
>
> I welcome suggestions for a better name.
>
> Roger
>
> _______________________________________________
> opensolaris-arc mailing list
> opensolaris-arc at opensolaris.org
>



-- 
PGP Public Key 0x437AF1A1
Available on hkp://pgp.mit.edu

Reply via email to