Congrats, this needs a complete rewrite of the the search function which isn't done in 5 min :P

Looks like GNOME added entirely new widgets for that, such as
GNOME/evince@16314cd#diff-1e2c3508bdd3edce0a4a082c112eb2a3

I can't collect bountysource bounties due to lack of a paypal account, so I am not going to be diverted to this by the bounty. Process would be to find all the evince commits and backport them, which could potentially be as big a job as the libunique->GtkApplication transition in Caja was. That job took me literally months.

I took a look at it 2 years ago.

• min 20-40 commits to backport
• gnome commits require a switch to GtkHeaderbars, which isn't MATE style. So simple back porting won't work.

Btw.
Thanks for donating, but paypal fee is very high when transferring money outside the states.
A least a developer will get 1 $ from your donation ;)
Any way, thank you very much.

The way I see it, the bounty itself is not incentive enough to do the work, but it gives it a bit more visibility, which could encourage other people to add to the bounty, or another random contributor to do a drive-by pull request.

If anything, it bumps the issue to the top of the notification list, so there's value in that too.

So this shows that backporting all those GNOME commits would also require redoing the widgets to use the GNOME 2/MATE toolbars and menubars rather than GtkHeaderBars. In other words, an even bigger job than that backend work I did for Caja back in 2016. In that go-around, GNOME kept all of the UI changes to separate commits from backend changes (as there was no relation between them), but in this case they could not as new GUI widgets had to be added, and those have to match the rest of evince.

Too bad git isn't capable enough to resolve all the conflicts that crop up from taking a current GNOME repo and trying to revert all the UI change commits-and even then, new code would be required to set up the UI for widgets that didn't exist in the GNOME 2 days.