When I got home late last evening, I found that Troy and Greg had had a conversation about this in #xiphos. Troy theorized and Greg confirmed that Xiphos must be calling RenderText() too often; I hadn't noticed this -- it happens in a "bridge" routine in Xiphos' backend by which to call the engine's getEntryAttributes(). Oops, yes, the performance flaw was there. The need for RenderText() at that point is because of processing mouse hover/click on footnotes and xrefs, in turn because the verse in question might not be the current module key (so it needs to be brought into context with RenderText()), but it's not appropriate _always_ to do it. So I fixed that.
This leaves the proposed patch as a philosophical question instead of a performance necessity. Troy mentioned in last night's conversation with Greg that "it takes away the choice for frontend developers". Perhaps true, but on the other hand the lack of n=X content in *n/*x means every frontend has to go through this effort, if it's wanted. I note that PocketSword replaces *n/*x with small caps N/X (neither '*' nor n=X content, as a screen real estate conservation issue), and I doubt BibleTime uses it at all, so I'm wondering what choice is being taken away. The reason I made Xiphos do this a while back is in line with what DM said yesterday, that the information is there, so I'd like to think we should display it, but the frontends don't get to display it without being required to do this post-processing of the HTML output. Thoughts/theories/attitudes? _______________________________________________ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page