Can you go into more detail on this?
Firefox uses private data locally every time you start typing in the address bar. What leads someone to object to this particular use (specifically for privacy reasons), but to leaving history on in the first place?
The one advantage to the local parsing compared to serverside is that there isn't a cloud full of everybody's browser history ready to be hacked or resold. That's a real benefit, I don't want to undersell that, but it's not the entirety of the story.
As for your example: Firefox using your browser history to autocomplete urls as you type in the address bar is innocuous because that information never goes anywhere but to that user; it's solely for the user's benefit. Firefox using that same browser history to tell advertisers which subjects the user might be interested in is not innocuous; it's broadcasting information to third parties that the user might prefer be kept private. The user's data is no longer being used for the user's benefit, but for the advertiser's.
The Pocket code tells Mozilla what ads are being shown and which are clicked on. Since they know that the criteria for ad X is that phrase Y or Z is in the history, then they are gathering data on the type of content you are looking at on the web. This is only one step removed from the actual URLs.
Really? It definitely shouldn't do that.
> and which are clicked on
That's pretty hard to avoid, though. How much of a problem it is depends on how precise the targeting is. But since you're showing interest by clicking the ad, there's not necessarily much of a privacy issue.