<p>Yeah, I was thinking of wt access possibly being simpler and faster by name, but you're right this will need testing and may well be a problem for committed trees.</p>
<div class="gmail_quote">On May 6, 2011 3:02 PM, "John Arbash Meinel" <<a href="mailto:john@arbash-meinel.com">john@arbash-meinel.com</a>> wrote:<br type="attribution">> -----BEGIN PGP SIGNED MESSAGE-----<br>
> Hash: SHA1<br>> <br>> On 05/06/2011 02:57 PM, Martin Pool wrote:<br>>> That's a good explanation of it, and I'd like to separate those three<br>>> things too. Accessing more tree contents by name rather than id may well<br>
>> improve performance too.<br>>> <br>> <br>> Right now data is stored in the repository very much by file-id.<br>> Probably just fallthrough from how we decided to reference it. But<br>> accessing *repository* content by path will always cause us to<br>
> dereference it from an Inventory (which is *also* stored based on<br>> file-id, so it is pretty worst-case behavior).<br>> <br>> If we do chose to switch, then we'll also want to switch our storage<br>> behavior.<br>
> <br>> John<br>> =:-><br>> <br>> -----BEGIN PGP SIGNATURE-----<br>> Version: GnuPG v1.4.10 (GNU/Linux)<br>> Comment: Using GnuPG with Mozilla - <a href="http://enigmail.mozdev.org/">http://enigmail.mozdev.org/</a><br>
> <br>> iEYEARECAAYFAk3D8UYACgkQJdeBCYSNAAPMpACgpbuqz7U7+X5XyPu6GsrTxGqO<br>> mukAn0dwerQYovzx59XJNvhbatPvV7c2<br>> =9FHt<br>> -----END PGP SIGNATURE-----<br>> <br></div>