<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Tue, Feb 14, 2017 at 10:13 PM, Matt Turner <span dir="ltr"><<a href="mailto:mattst88@gmail.com" target="_blank">mattst88@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><span class="gmail-">On Tue, Feb 14, 2017 at 3:52 PM, Jason Ekstrand <<a href="mailto:jason@jlekstrand.net">jason@jlekstrand.net</a>> wrote:<br>
> The new version does a SHA1 sum of the timestamp together with the<br>
> device PCI ID. This fixes a theoretical bug where, if you moved a drive<br>
> between machines, you could end up trying to use a pipeline cache from<br>
> one platform with a different platform. Now we'll reject any cache that<br>
> is targeted at different hardware.<br>
<br>
</span>I suspect that my patches from this morning brought this to the front<br>
of your mind. I would really rather you just make such a suggestion to<br>
me as a review comment than sending a conflicting patch before I have<br>
time to send v2.<br>
</blockquote></div><br></div><div class="gmail_extra">Actually, they were completely coincidental. What brought this to mind was some work that I'm doing on a future feature that requires a bit more UUIDing of things. I realized there was a conflict but I figured that switching from mtime to build-id and factoring in PCI ID were orthogonal things. I knew when I sent the patch that I would have to rebase it. I'm sorry if it came off as passive-aggressive or something.<br><br></div><div class="gmail_extra">In case I haven't said so, your figuring out the build-id stuff is very much appreciated! It's way better than mtime!<br><br></div><div class="gmail_extra">--Jason<br></div></div>