## Summary
* **What is the goal of this PR?**
Replace simple threshold-based grayscale quantization with ordered
dithering using a 4x4 Bayer matrix. This eliminates color banding
artifacts and produces smoother gradients on e-ink display.
* **What changes are included?**
- Add 4x4 Bayer dithering matrix for 16-level threshold patterns
- Modify `grayscaleTo2Bit()` function to accept pixel coordinates and
apply position-based dithering
- Replace simple `grayscale >> 6` threshold with ordered dithering
algorithm that produces smoother gradients
## Additional Context
* Bayer matrix approach: The 4x4 Bayer matrix creates a repeating
pattern that distributes quantization error spatially, effectively
simulating 16 levels of gray using only 4 actual color levels (black,
dark gray, light gray, white).
* Cache invalidation: Existing cached `cover.bmp` files will need to be
deleted to see the improved rendering, as the converter only runs when
the cache is missing.
## Summary
In using my build of
https://github.com/daveallie/crosspoint-reader/pull/130 I realized that
we need a "open" button hint above the second button in the File browser
## Additional Context
* Add any other information that might be helpful for the reviewer
(e.g., performance implications, potential risks, specific areas to
focus on).
## Summary
* **What is the goal of this PR?** Fixes a bug where text disappears
after approximately 25 pages in long chapters during EPUB indexing.
* **What changes are included?**
- Removed the `MAX_LINES = 1000` hard limit in
`ParsedText::computeLineBreaks()`
- Added safer infinite loop prevention by checking if `nextBreakIndex <=
currentWordIndex` and forcing advancement by one word when stuck
## Additional Context
* **Root cause:** The `MAX_LINES = 1000` limit was introduced to prevent
infinite loops, but it truncates content in long chapters. For example,
a 93KB chapter that generates ~242 pages (~9,680 lines) gets cut off at
~1000 lines, causing blank pages after page 25-27.
* **Solution approach:** Instead of a hard line limit, I now detect when
the line break algorithm gets stuck (when `nextBreakIndex` doesn't
advance) and force progress by moving one word at a time. This preserves
the infinite loop protection while allowing all content to be rendered.
* **Testing:** Verified with a Korean EPUB containing a 93KB chapter -
all 242 pages now render correctly without text disappearing.
Using QRCode library from pio to generate the QR code.
Done:
- Display QR code for URL in network mode
- minor fixes of layout
- Display QR for URL in AP mode
- Display QR for AP in AP mode
---------
Co-authored-by: Dave Allie <dave@daveallie.com>
## Summary
* **What is the goal of this PR?**
Add a "Continue Reading" feature to improve user experience when
returning to a previously opened book.
* **What changes are included?**
- Add dynamic "Continue: <book name>" menu item in Home screen when a
book was previously opened
- File browser now starts from the folder of the last opened book
instead of always starting from root directory
- Menu dynamically shows 3 or 4 items based on reading history:
- Without history: `Browse`, `File transfer`, `Settings`
- With history: `Continue: <book>`, `Browse`, `File transfer`,
`Settings`
## Additional Context
* This feature leverages the existing `APP_STATE.openEpubPath` which
already persists the last opened book path
* The Continue Reading menu only appears if the book file still exists
on the SD card
* Book name in the menu is truncated to 25 characters with "..." suffix
if too long
* If the last book's folder was deleted, the file browser gracefully
falls back to root directory
* No new dependencies or significant memory overhead - reuses existing
state management
## Summary
This creates a `renderer.drawButtonHints` to make all of the "hints"
over buttons to match the home screen.
## Additional Context
* Add any other information that might be helpful for the reviewer
(e.g., performance implications, potential risks, specific areas to
focus on).
---------
Co-authored-by: Dave Allie <dave@daveallie.com>
Replace linear O(n) search with binary search O(log n) for unicode
interval lookup. Korean fonts have many intervals (~30,000+ glyphs), so
this improves text rendering performance during page navigation.
## Summary
* **What is the goal of this PR?** (e.g., Fixes a bug in the user
authentication module, Implements the new feature for
file uploading.)
Replace linear `O(n)` glyph lookup with binary search `O(log n)` to
improve text rendering performance during page navigation.
* **What changes are included?**
- Modified `EpdFont::getGlyph()` to use binary search instead of linear
search for unicode interval lookup
- Added early return for empty interval count
## Additional Context
* Add any other information that might be helpful for the reviewer
(e.g., performance implications, potential risks, specific areas to
focus on).
- Performance implications: Fonts with many unicode intervals benefit
the most. Korean fonts have ~30,000+ glyphs across multiple intervals,
but any font with significant glyph coverage (CJK, extended Latin,
emoji, etc.) will see improvement.
- Complexity: from `O(n)` to `O(log n)` where n = number of unicode
intervals. For fonts with 10+ intervals, this reduces lookup iterations
significantly.
- Risk: Low - the binary search logic is straightforward and the
intervals are already sorted by unicode codepoint (required for the
original early-exit optimization).
## Summary
* Use single unified cache file for book spine, table of contents, and
core metadata (title, author, cover image)
* Use new temp item store file in OPF parsing to store items to be
rescaned when parsing spine
* This avoids us holding these items in memory
* Use new toc.bin.tmp and spine.bin.tmp to build out partial toc / spine
data as part of parsing content.opf and the NCX file
* These files are re-read multiple times to ultimately build book.bin
## Additional Context
* Spec for file format included below as an image
* This should help with:
* #10
* #60
* #99
## Summary
* Prevent SD card error causing boot loop
* We need the screen and fonts to be initialized to show the full screen
error message
* Prior to this change, trying to render the font would crash the
firmware and boot loop it
## Summary
* Handle 16x16 MCU blocks in JPEG decoding
* We were only correctly handling 8x8 blocks, which means that we did
not correctly support a lot of JPGs leading to an interlacing style on
the images
## Additional Context
* Fixes https://github.com/daveallie/crosspoint-reader/issues/118
## Summary
* **What is the goal of this PR?** Adds WiFi Access Point (AP) mode
support for File Transfer, allowing the device to create its own WiFi
network that users can connect to directly - useful when no existing
WiFi network is available. And in my experience is faster when the
device is right next to your laptop (but maybe further from your wifi)
* **What changes are included?**
- New `NetworkModeSelectionActivity` - an interstitial screen asking
users to choose between:
- "Join a Network" - connects to an existing WiFi network (existing
behavior)
- "Create Hotspot" - creates a WiFi access point named
"CrossPoint-Reader"
- Modified `CrossPointWebServerActivity` to:
- Launch the network mode selection screen before proceeding
- Support starting an Access Point with mDNS (`crosspoint.local`) and
DNS server for captive portal behavior
- Display appropriate connection info for both modes
- Modified `CrossPointWebServer` to support starting when WiFi is in AP
mode (not just STA connected mode)
## Additional Context
* **AP Mode Details**: The device creates an open WiFi network named
"CrossPoint-Reader". Once connected, users can access the file transfer
page at `http://crosspoint.local/` or `http://192.168.4.1/`
* **DNS Captive Portal**: A DNS server redirects all domain requests to
the device's IP, enabling captive portal behavior on some devices
* **mDNS**: Hostname resolution via `crosspoint.local` is enabled for
both AP and STA modes
* **No breaking changes**: The "Join a Network" option preserves the
existing WiFi connection flow
* **Memory impact**: Minimal - the AP mode uses roughly the same
resources as STA mode
## Summary
* Adds support for OTA
* Gets latest firmware bin from latest GitHub release
* I have noticed it be a little flaky unpacking the JSON and
occasionally failing to start
## Summary
* HTML files are now static, streamed directly to the client without
modification
* For any dynamic values, load via JSON APIs
* For files page, we stream the JSON content as we scan the directory to
avoid holding onto too much data
## Additional details
* We were previously building up a very large string all generated on
the X4 directly, we should be leveraging the browser
* Fixes https://github.com/daveallie/crosspoint-reader/issues/94
Improves the duration for which the power button needs to be held - see
#53.
I left the measurement code for the calibration value in, as it will
likely change if we move the settings to NVS.
## Summary
* Give activities name and log when entering and exiting them
* Clearer logs when attempting to debug, knowing where users are coming
from/going to helps
## Summary
* Swap from `wasReleased` to `isPressed` when checking power button
duration
* In practice it makes the power down experience feel a lot snappier
* Remove the unnecessary 1000ms delay when powering off
## Additional Context
* A little discussion in here:
https://github.com/daveallie/crosspoint-reader/discussions/53#discussioncomment-15309707
## Summary
* adds cyrillic glyphs to pixel arial font, used as Small font in UI
## Additional Context
* with recent changes pixel arial font lost cyrillic glyphs
## Summary
* Fix incorrect justification of last line in paragraph
* `words` is changing size due to the slice, so `isLastLine` would
rarely be right, either removing justification mid-paragraph, or
including it in the last line.
## Additional Context
* Introduced in #73
## Summary
* Unset openEpubPath on boot and set once epub fully loaded
## Additional Context
* If an epub was crashing when loading, it was possible to get the
device stuck into a loop. There was no way to get back to the home
screen as we'd always load you back into old epub
* Break this loop by clearing the stored value when we boot, still
jumping to the last open epub, but only resetting that value once the
epub has been fully loaded
## Summary
- Add basic JPG image support
- Map JPG back to 2-bit BMP output
- Can be used to later render the BMP file from disk or directly pass to
output if wanted
- Give the 3 passes over the data needed to render grayscale content,
putting it on disk is preferred to outputting it multiple times
## Additional Context
- WIP, looking forward to BMP support from
https://github.com/daveallie/crosspoint-reader/pull/16
- Addresses some of #11
## Summary
* Paginate book list
* Avoid out of bounds rendering of long book titles, truncate with
ellipsis instead
## Additional Context
* Should partially help with
https://github.com/daveallie/crosspoint-reader/issues/75 as it was
previously rendering a lot of content off screen, will need to test with
a large directory
## Summary
* Extract EPUB TOC into temp file before parsing
* Streaming ZIP -> XML parser uses up a lot of memory as we're
allocating inflation buffers while also holding a few copies of the
buffer in different forms
* Instead, but streaming the inflated file down to the SD card (like we
do for HTML parsing, we can lower memory usage)
## Additional Context
* This should help with
https://github.com/daveallie/crosspoint-reader/issues/60 and
https://github.com/daveallie/crosspoint-reader/issues/10. It won't
remove those class of issues completely, but will allow for many more
books to be opened.
Still a bit raw, but gets the time required to determine the size of
each chapter (for reading progress) down from ~25ms to 0-1ms.
This is done by keeping the zipArchive open (so simple ;)).
Probably we don't need to cache the spine sizes anymore then...
---------
Co-authored-by: Dave Allie <dave@daveallie.com>
## Summary
* Build out lines for pages when holding over 750 buffered words
* Should fix issues with parsing long blocks of text causing memory
crashes
## Problem
`drawBmp()` allocates two row buffers via `malloc()` but doesn't check
if allocations succeed. On low memory, this causes a crash when the NULL
pointers are dereferenced.
## Fix
Add NULL check after both `malloc()` calls. If either fails, log error
and return early.
Changed `lib/GfxRenderer/GfxRenderer.cpp`.
## Test
- Defensive addition only - no logic changes
- Manual device testing appreciated
## Problem
Three Epub getter functions can throw exceptions:
- `getCumulativeSpineItemSize()`: No bounds check before
`.at(spineIndex)`
- `getSpineItem()`: If spine is empty and index invalid, `.at(0)` throws
- `getTocItem()`: If toc is empty and index invalid, `.at(0)` throws
## Fix
- Add bounds check to `getCumulativeSpineItemSize()`, return 0 on error
- Add empty container checks to `getSpineItem()` and `getTocItem()`
- Use static fallback objects for safe reference returns on empty
containers
Changed `lib/Epub/Epub.cpp`.
## Test
- Defensive additions - follows existing bounds check patterns
- No logic changes for valid inputs
- Manual device testing appreciated
## Problem
`readFileToMemory()` allocates an output buffer via `malloc()` at line
120 but doesn't check if allocation succeeds. On low memory, the NULL
pointer is passed to `fread()` causing a crash.
## Fix
Add NULL check after `malloc()` for the output buffer. Follows the
existing pattern already used for `deflatedData` at line 141.
Changed `lib/ZipFile/ZipFile.cpp`.
## Test
- Follows existing validated pattern from same function
- Defensive addition only - no logic changes
## Problem
`invertScreen()`, `storeBwBuffer()`, and `restoreBwBuffer()` dereference
`frameBuffer` without NULL validation. If the display isn't initialized,
these functions will crash.
## Fix
Add NULL checks before using `frameBuffer` in all three functions.
Follows the existing pattern from `drawPixel()` (line 11) which already
validates the pointer.
Changed `lib/GfxRenderer/GfxRenderer.cpp`.
## Test
- Follows existing validated pattern from `drawPixel()`
- No logic changes - only adds early return on NULL
- Manual device testing appreciated
## Summary
- **What is the goal of this PR?**
Implements wireless EPUB file management via a built-in web server,
enabling users to upload, browse, organize, and delete EPUB files from
any device on the same WiFi network without needing a computer cable
connection.
- **What changes are included?**
- **New Web Server**
([`CrossPointWebServer.cpp`](src/CrossPointWebServer.cpp),
[`CrossPointWebServer.h`](src/CrossPointWebServer.h)):
- HTTP server on port 80 with a responsive HTML/CSS interface
- Home page showing device status (version, IP, free memory)
- File Manager with folder navigation and breadcrumb support
- EPUB file upload with progress tracking
- Folder creation and file/folder deletion
- XSS protection via HTML escaping
- Hidden system folders (`.` prefixed, "System Volume Information",
"XTCache")
- **WiFi Screen** ([`WifiScreen.cpp`](src/screens/WifiScreen.cpp),
[`WifiScreen.h`](src/screens/WifiScreen.h)):
- Network scanning with signal strength indicators
- Visual indicators for encrypted (`*`) and saved (`+`) networks
- State machine managing: scanning, network selection, password entry,
connecting, save/forget prompts
- 15-second connection timeout handling
- Integration with web server (starts on connect, stops on exit)
- **WiFi Credential Storage**
([`WifiCredentialStore.cpp`](src/WifiCredentialStore.cpp),
[`WifiCredentialStore.h`](src/WifiCredentialStore.h)):
- Persistent storage in `/sd/.crosspoint/wifi.bin`
- XOR obfuscation for stored passwords (basic protection against casual
reading)
- Up to 8 saved networks with add/remove/update operations
- **On-Screen Keyboard**
([`OnScreenKeyboard.cpp`](src/screens/OnScreenKeyboard.cpp),
[`OnScreenKeyboard.h`](src/screens/OnScreenKeyboard.h)):
- Reusable QWERTY keyboard component with shift support
- Special keys: Shift, Space, Backspace, Done
- Support for password masking mode
- **Settings Screen Integration**
([`SettingsScreen.h`](src/screens/SettingsScreen.h)):
- Added WiFi action to navigate to the new WiFi screen
- **Documentation** ([`docs/webserver.md`](docs/webserver.md)):
- Comprehensive user guide covering WiFi setup, web interface usage,
file management, troubleshooting, and security notes
- See this for more screenshots!
- Working "displays the right way in GitHub" on my repo:
https://github.com/olearycrew/crosspoint-reader/blob/feature/connect-to-wifi/docs/webserver.md
**Video demo**
https://github.com/user-attachments/assets/283e32dc-2d9f-4ae2-848e-01f41166a731
## Additional Context
- **Security considerations**: The web server has no
authentication—anyone on the same WiFi network can access files. This is
documented as a limitation, recommending use only on trusted private
networks. Password obfuscation in the credential store is XOR-based, not
cryptographically secure.
- **Memory implications**: The web server and WiFi stack consume
significant memory. The implementation properly cleans up (stops server,
disconnects WiFi, sets `WIFI_OFF` mode) when exiting the WiFi screen to
free resources.
- **Async operations**: Network scanning and connection use async
patterns with FreeRTOS tasks to prevent blocking the UI. The display
task handles rendering on a dedicated thread with mutex protection.
- **Browser compatibility**: The web interface uses standard
HTML5/CSS3/JavaScript and is tested to work with all modern browsers on
desktop and mobile.
---------
Co-authored-by: Dave Allie <dave@daveallie.com>
Adresses #53
Please check if we still need the code to "Give the user up to 1000ms to
start holding the power button, and must hold for
SETTINGS.getPowerButtonDuration()" - the power button should be pressed
already when waking up...
Also, decided to return before the delay to wait more to make the
behavior more immediate.
---------
Co-authored-by: Dave Allie <dave@daveallie.com>
## Problem
Three `fopen()` calls in ZipFile.cpp did not check for NULL before using
the file handle. If files cannot be opened, `fseek`/`fread`/`fclose`
receive NULL and crash.
## Fix
Added NULL checks with appropriate error logging and early returns for
all three locations:
- `getDataOffset()`
- `readFileToMemory()`
- `readFileToStream()`
## Testing
- Builds successfully with `pio run`
- Affects: `lib/ZipFile/ZipFile.cpp`
## Problem
- `getBookSize()` calls `getCumulativeSpineItemSize(getSpineItemsCount()
- 1)` which passes -1 when spine is empty
- `calculateProgress()` then divides by zero when book size is 0
## Fix
- Return 0 from `getBookSize()` if spine is empty
- Return 0 from `calculateProgress()` if book size is 0
## Testing
- Builds successfully with `pio run`
- Affects: `lib/Epub/Epub.cpp`
