Review fix for #2309: set_active, add_document and close_document emptied the place cursor but
kept the placeGun tool, so the next canvas click failed with 'nothing loaded in the place
cursor'. One helper, Session::unload_place_cursor, empties the cursor and its queue and goes back
to the Selection tool when the place cursor's tool is active; validate_place_cursor uses it too.
Test: changing_documents_unloads_the_place_cursor_and_its_tool.
Split out of #2073. Port of #1307's drops and #1666.
- Several images dropped together load the place cursor in the order they
came, each run of them where it comes among the other files, after the
images it already holds; click to place them one by one.
- One image dropped off a frame that takes it loads the place cursor instead
of being placed at once; on such a frame it still fills it.
- A drop is handled in its order, each file at its own pointer, and a dialog
(a multi-page PDF's Import Options, an alert) cancels the rest.
- On the web, files dropped together are read as one batch (none if one
can't be read), and a single dropped image loads the place cursor.
Split out of #2073. winit sends no pointer events during a drag from another
application on macOS, so a Finder drop had no pointer and never found the frame
under it. The host now gives the pointer's place on the screen
(`Services::pointer`, `PointerFn`: the system's points before the UI scale,
from `NSEvent.mouseLocation`), and `take_pending_drops` uses it for the files
that came without a pointer, divided by egui's zoom factor and made relative
to the window.
Split out of #2073. Port of #1307, #1392 (grid) and #1432 (thumbnail and count).
- File › Place takes several files; images load the place cursor in the
order chosen (`place.load` `files`, `append`), after the ones it holds.
A multi-page PDF or an alert stops the batch, and a file that can't be
placed returns its error to `app.placeDialog`.
- Each click places the next image; arrow keys while dragging lay out a
grid (`place.grid`) and are handled even at a limit; Escape unloads the
rest (`place.cancel`). Undo of a loaded image's asset cancels the queue.
- The cursor shows how many images are loaded.
- On the web, Place picks several files and the inbox loads consecutive
images for the same document as one batch.
* Frames holding pasted-in images are hit inside the frame, take a placed graphic; documents open on a usable layer
A frame with items pasted into it (how a document from InDesign keeps a cropped photo) was hit
through its children like a group, so the whole uncropped image counted: a click on the empty
page beside such a photo selected it, and nothing could be deselected there. It is now hit by
its own area, as it is drawn clipped. Such a frame holding only graphics also takes a placed or
dropped graphic (Item::takes_graphic), which fills it as an empty or graphic frame is filled.
A document opened with new objects going to its top layer even when that layer was hidden and
locked (a dieline), so a placed graphic showed as an empty frame outline. It now opens on the
topmost layer that is shown and unlocked (Document::working_layer).
* Place: a graphic replacing the one a frame shows keeps the frame's crop and zoom
Placing or dropping into a frame that shows a graphic (in it, or as the one graphic pasted into
it) filled the frame afresh, so the crop and zoom chosen for the old picture were lost. The new
graphic now covers the old one's area, centred on it: an edited copy of the same picture lands
exactly where the old one was.
* Move the replace-keeps-crop and open-on-a-usable-layer changes to their own PRs
Review asked for them to be checked against reference app behaviour on their own. This PR keeps
the frame hit area and Item::takes_graphic. Reverts 8ea0bc16 and drops Document::working_layer.
* Move Document::working_layer and its test to #2267
A document opened with new objects going to its top layer even when that layer was hidden and
locked (a dieline), so a placed graphic showed as an empty frame outline. Document::working_layer
picks the topmost layer that is shown and unlocked (the top layer when there is none); DocState::new
and the fallback after the active layer is deleted use it.
* hyphenation: linear heuristic, cache misses outside the lock
The syllable heuristic built each word suffix as a String to test it, which is
quadratic in the word's length; it now uses prefix sums of vowels and UTF-8
lengths. Mapping a word's break points to glyphs scanned the word for each
point; both are indexed once per word. A word not in the hyphenation cache is
worked out without holding the cache's global mutex.
* hyphenation: compare the word ending with ends_with
A cached story copied every wrap neighbour's whole signature and item list into
its entry; stories on one spread share its wrapping items, so entries grew with
all of theirs together. A neighbour's signature now goes in as its hash and its
items are kept once each.
Anchored-frame placement asked for the host's composition once per frame, and
again at every level a host was reached from. Each host is now asked for once
per placement pass, and the cache and the uncached path remember the hosts they
resolved for the whole traversal.
- Footnote references on a line are deduplicated with a set instead of
searching the references collected so far.
- Nested styles stream the paragraph up to their delimiter with one character
of lookahead instead of copying the rest of the paragraph for each style,
and a delimiter's characters are indexed once per style.
- Shaping resolves which nested/GREP style is on top with one sweep over the
overlay boundaries instead of scanning every overlay for each stretch.
- Column rules group crossing lines by paragraph through an index.
- Table fragments index each row's top once instead of searching the rows
for every cell.
The warning from #2190 looked only at the attributes passed. A composite
font made by style.compositeFont.set isn't an installed family, so
applying one warned "Missing font: Mixed" even when its entries were
installed. And a fontStyle-only call, or a family change that kept a
style the new family lacks, said nothing.
After a call that sets or clears fontFamily or fontStyle, type.char now
tallies the fonts the formatted text resolves to, with the code
font.list uses (composite entries per character), and warns about the
missing ones and substitute styles. At a caret it checks the font the
next typing gets. tally_story becomes tally_range over a byte range,
and describe_fonts is the part of used_fonts that type.char reuses.
texjer
aiongg
Ikki Nishiharaandharanishi
Dylan M. Taylor
Hamza
JonRC