174583 - Remove nsIRDFDataSource { nsIEnumerator GetAllCommands(in nsIRDFResource aSource); }

Status: RESOLVED FIXED

(Core Graveyard :: RDF, defect)

Description

[timeless]

There aren't any real consumers.

Comment 1

[timeless]

Attachment: draft remove GetAllCommands

Comment 2

[timeless]

Here's a list of Modules, Files, and Owners.
If the owners or mozilla.org would sign off on this then I'll work on getting it
in at some point (depending on how important it is to mozilla.org).

XPApps <[e-mail]>
mozilla/xpfe/components/search/src/nsInternetSearchService.cpp {Internet Search}
<[e-mail]>
mozilla/xpfe/components/history/src/nsGlobalHistory.cpp {History}
<[e-mail]>
mozilla/xpfe/components/intl/nsCharsetMenu.cpp {Intl} <[e-mail]>
mozilla/xpfe/components/windowds/nsWindowDataSource.cpp <[e-mail]>

Mail <[e-mail]> <[e-mail]>

Comment 3

[Robert John Churchill]

r=rjc
for bookmarks, internet search, local search, history, http index, related
links,  window datasource and RDF changes

Comment 4

[Alec Flett]

Comment on attachment 102942 [details] [diff] [review]
draft remove GetAllCommands

no consumers? wacky. Seems like at the very least this should be using
nsISimpleEnumerator. 

In any case, sr=alecf for the footprint work. I wonder how many people even use
all the old command work that is in RDF? I wonder if we might consider moving
it to another interface (nsIRDFCommandDataSource?) so that not all consumers
would be forced to implement IsCommandEnabled/etc.

Those consumers that use it could QI to nsIRDFCommandDataSource or something

Comment 5

[Robert John Churchill]

> no consumers? wacky. Seems like at the very least this should be using
nsISimpleEnumerator.

History: Warren added GetAllCommands() which uses nsIEnumerator.  At some later
point, I believe that I added GetAllCmds() which uses nsISimpleEnumerator. It is
a good thing to remove the former and should have been done a long time ago
except that we were all busy with other things. As for adding yet another
interface: smells like "busy-work".

Comment 6

[Alec Flett]

the reason I don't think its busy work is that I believe that if 90% of rdf
datasources / consumers don't use the command system, then there's no need to
burden EVERY datasource with its use. It reduces code bloat.

Comment 7

[Robert John Churchill]

Alec, I suspect that we're talking about multiple bytes, not multiple K.  Just
having to add additional QIs for a new interface will cost some bloat (again,
bytes).  Feel free to measure.

Comment 8

[David :Bienvenu]

mail stuff looks ok, if indeed this interface is never called.

Comment 9

[Bradley Baetz (:bbaetz)]

Comment on attachment 102942 [details] [diff] [review]
draft remove GetAllCommands

r=bbaetz, given the various moa= in the bug

Comment 10

[timeless]

bbaetz commented in the wrong bug, but this is fixed (except for phoenix which
now has an out of sync cloned file).

Comment 11

[benc]

tever is not RDF QA anymore