Precondition:
Ubuntu 24.04.3
Scanner Brother MFC-J6920DW
scantpaper 3.0.19
Symptom: After a page scan, there appears to be no way to abandon a bad scan. Neither the 'New' button (which has gained a +) nor the new menu selection will let you abandon the scan until after it has been saved.
It appears that you can select individual pages and cut them until the buffer is empty.
Symptom: If you scan a single page at 600 dpi color, (Book Cover) and then you load a different profile (black and white 300 dpi) Additional pages are scanned at the original profile configurations (Using gscan2pdf profiles)
Symptom: to the brother scanner. If the setting is flatbed and you have pages in the document feeder it will scan those instead. (probably brother specific) I have not checked if this behavior is common to gscan2pdf.
Status: I will start using scantpaper at this point but leave gscan2pdf installed for now
In order to give me more information to assess #2 and #3, please start scantpaper from the command line with the --log=log option, reproduce the problem, quit, and post the log file, which scantpaper should have compressed with xz.
In Gnome 46, the New button does not bring up any dialog, nor clears anything.
Ubuntu 24.04.2 Wayland
You can delete the resident scan by selecting all then Edit -> Delete
On Item 2, I discovered that if you change the profile and move back to the page options page it scans with the profile showing, not what you had set as a temporary option. Operator confusion. Please mark resolved.
Item 3 I have attached log
My read on your symptom
The code demonstrably sets source = FlatBed. If the device still feeds the ADF when paper is sitting in it, that’s the Brother MFC-J6920DW’s physical behaviour — the hardware prioritises the feeder regardless of the SANE source option. Since gscan2pdf also just sets the same source option (scantpaper is its rewrite), I’d expect gscan2pdf to behave identically — i.e. it’s not scantpaper-specific. That’s consistent with the “probably brother specific” hunch, but I can’t confirm from this log because it doesn’t actually capture a flatbed-selected-but-feeder-scanned run.
Please give me a different log file to demonstrate your problem with the new button.