The reader draws a crisp BW page, then a 4-level grayscale antialiasing
pass with inverted plane polarity (white = level 0 = both planes clear)
and near-empty planes. The old model replaced the whole frame with those
planes, wiping the page to an inverted muddle.
Model it as e-ink actually behaves: keep a persistent surface, detect
grayscale vs BW by DTM2's set-fraction (background is the white
majority), and in grayscale treat level 0 as no-drive (keep pixel) so
only the sparse antialiased edges repaint over the retained BW page. BW
refreshes still repaint every pixel. Reader text now renders sharp.
A TB containing a guest store is miscompiled by the wasm backend: it
corrupts a pointer that a later load faults on. It reproduces in uzlib
inflate during the reader's font prewarm (browser panics with a rejected
Load access fault); native QEMU decompresses the same static flash font
fine, confirming a wasm32 codegen bug.
Force store-containing TBs onto the TCI interpreter as the working
workaround. Perf ceiling: those TBs no longer JIT-compile. Upgrade path:
localize the miscompiled op in tcg_wasm_out_qemu_st and drop this.
ssi-sd now enters token mode as it returns the final R1 byte, so the
write-data token that immediately follows is not discarded. SPI CMD13
returns the internal sd_r1 response (upstream fix) instead of sd_r2_s,
which the adapter converts to the two-byte SPI R2 status; this stops the
repeated "Unexpected response to cmd 13" loop on guest status polls.
Drops the experimental store-TB TCI forcing and sd-watch OOB logging now
that the raw FAT32 image path is stable.
Diagnostic: routes store-containing TBs through TCI and logs out-of-bounds
guest stores to isolate the SdSpiCard corruption. Also carries the FAT32
vvfat root/cluster fixes and ssi-sd command framing. Squash/split before
upstreaming.
Replace three separate Jira sync workflows (new_issues.yml,
issue_comment.yml, new_prs.yml) with a single sync-jira.yml following
the recommended pattern from espressif/sync-jira-actions.
Changes:
- Migrate from espressif/github-actions/sync_issues_to_jira@master
to espressif/sync-jira-actions@v1
- Add missing GITHUB_TOKEN permissions (contents:read, issues:write,
pull-requests:write)
- Add workflow_dispatch trigger for manual syncing
- Add dynamic run-name for better visibility in Actions UI
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Node.js 20 is deprecated on GitHub Actions runners and will be forced
to Node.js 24 starting June 2, 2026, with Node.js 20 removed entirely
on September 16, 2026.
Update all actions to versions that support Node.js 24:
- actions/checkout: v4/master -> v6
- actions/upload-artifact: v4 -> v7
- actions/download-artifact: v4 -> v8
- softprops/action-gh-release: v1 -> v2
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
1. Missing bare ee.vldbc.{8,16,32} instructions - Added 3 opcode table entries with addr_nop and updated translate_vldbc_s3 to handle addr_nop without erroring.
2. Missing ee.st.qacc_{h,l}.{h.32,l.128}.ip store instructions - Added 4 opcode table entries, 2 translate functions, and 2 helper functions (with declarations in helper.h) to implement the QACC store operations as mirrors of the existing load counterparts.
3. ee.src.q incorrect qs1->qs0 copy - The bare ee.src.q qa, qs0, qs1 was incorrectly copying qs1→qs0 after the funnel shift. Per the TRM, only the .qup variant does this copy. Fixed by making the copy conditional on op_type == vmul_qup.
4. Add ee.st.accx.ip to store 40-bit ACCX (zero-extended to 64-bit) to memory, 8-byte aligned.
5. Add ee.ld.ua_state.ip to load 128-bit UA_STATE from memory, 16-byte aligned.
6. Add ee.st.ua_state.ip to store 128-bit UA_STATE to memory, 16-byte aligned.
7. Add ee.ldxq.32: Indexed load: compute address from Q register s16 segment, load 32-bit into Q register segment.
8. Add ee.stxq.32: Indexed store: compute address from Q register s16 segment, store 32-bit from Q register segment.
Fix the Windows release binary failing to start due to missing
libiconv-2.dll. The root cause was the sed workaround using .dll.a
files (import libraries for dynamic linking) instead of .a files
(static libraries).
Changes:
- Fix sed command in configure-win.sh to use static libraries (.a)
instead of dynamic import libraries (.dll.a) for libiconv/libintl
- Add bundle-win-dlls.sh script that detects and copies any missing
DLLs as a safety net (reports none needed with proper static linking)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>