2007/5/21, jamie <<a href="mailto:jamiemcc@blueyonder.co.uk">jamiemcc@blueyonder.co.uk</a>>:<div><span class="gmail_quote"></span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On Mon, 2007-05-21 at 15:10 +0200, Mikkel Kamstrup Erlandsen wrote:<br>><br>><br>> 2007/5/21, Mikkel Kamstrup Erlandsen <<a href="mailto:mikkel.kamstrup@gmail.com">mikkel.kamstrup@gmail.com</a>>:<br>> 2007/5/19, Jos van den Oever <
<a href="mailto:jvdoever@gmail.com">jvdoever@gmail.com</a>>:<br>> Hi all,<br>><br>> I was rereading the xesam-query.xsd and have the<br>> following remarks about it.
<br>><br>> - There is no target namespace. This must be fixed<br>> before it becomes final.<br>> - The attribute 'type' on <query> is redundant since
<br>> one also use a<br>> selection for it.<br>><br>> Why is the type attrib redundant? Fx searching for stuff of<br>> type (or "category" in our current terminology) Document and
<br>> Xesam:Content.Creator=Jos needs it AFAI can see.<br>><br>><br>> The namespace should be fixed, that is certain. I believe the<br>> wiki page also state that... It would just be nice to know if
<br>> we get those project pages on fdo or not before we settle on a<br>> namespace.<br>><br>> I have some other "diff"-proposals for the search spec<br>> somewhere, I'll post them in a sec...
<br>><br>><br>> Here are my proposed updates to the search spec (in addition to Jos'<br>> and Fabrice's):<br>><br>> 1) rename iface to org.freedesktop.xesam.Search (capitalize the last<br>> word "search") according to dbus interface naming conventions
<br>><br>> 2) spec out that client should take care to close their session<br>> objects, but that it it the servers responsibility to clear them up<br>> anyway. This can be done by logging caller name in NewSession and
<br>> listening to NameOwnerChanged on org.freedesktop.DBus on<br>> old_name->empty string.<br><br>make sense<br><br>><br>> 3) Only allow UTF-8 XML in query language<br><br>of course<br><br>><br>> 4) There is no way to do the 'w' switch from user search language in
<br>> xml query language. I propose to do this via a <string> attribute.<br><br>whats w switch?</blockquote><div><br>Try and match words inside strings in a sane way. Fx. "case" matches "CamelCase"w.
<br>Note that this feature would be an extension to the basic selectors of course, so it wouldn't be required to implement a xesam-compliant engine.<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
><br>> 5) And a bigger issue, that is a tough call... Should we use real dbus<br>> objects for sessions/searches? According to Havoc Pennington it<br>> doesn't take a roundtrip to the bus to register a new object, so
<br>> spawning lots of search objects shouldn't flood the bus (not with<br>> object-registration requests at least).<br><br>yes the overhead is negligible - its not like an app is going to<br>construct 100's of search objects all at once
</blockquote><div><br>I just had seconds thoughts on this. One thing is that the spawning of the Search object on the server doesn't touch the bus, but what about when the client sets up a proxy to it?<br><br>The current way where only handles are used is guaranteed to not have extra dbus roundtrips and I really think this is an important feature.
<br> </div>Maybe we wont handle 100's of searches, but I could easily imagine myself doing 50 searches/sec or whatever the engine could pull off. This could fx be in a scenario where I would like to extract meta-information about the index. Such as information clusters.
<br><br>Cheers,<br>Mikkel<br></div><br>