From 1ef08402893b0a97f2d018f36ca2f7b83925ba06 Mon Sep 17 00:00:00 2001 From: JC-000 <3798556+JC-000@users.noreply.github.com> Date: Thu, 15 Oct 2026 13:30:53 -0500 Subject: [PATCH 1/4] m3 demo README: image is built from f3df8261, branch has moved; update upstream status pr/3-tls-r1 fast-forwarded to 6271e7e2 (ESP TLS code moved to software/wifi/tls/; no behaviour change). The prerelease and image are unchanged. UPSTREAM_PRS now names GideonZ/1541ultimate#995, draft #997 and draft #996 (#994); none merged. Co-Authored-By: Claude Opus 5.5 (1M context) --- tools/package/build_m3_demo.sh | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/tools/package/build_m3_demo.sh b/tools/package/build_m3_demo.sh index 8889767..50ff50b 100755 --- a/tools/package/build_m3_demo.sh +++ b/tools/package/build_m3_demo.sh @@ -55,19 +55,20 @@ MAKE_ARGS=(BACKEND=uci-m3 "HTTPS_HOST=$DEMO_HOST" "HTTPS_PATH=$DEMO_PATH") # UPSTREAM_PRS are plain constants, not env knobs, so the README is a # function of the commit: a later commit swaps them. FIRMWARE="TLS client sockets for the Ultimate 64 / U64 Elite, from - JC-000/1541ultimate branch pr/3-tls-r1 (f3df8261), prebuilt image in - prerelease tls-preview-f3df8261. Not upstream; proposed on - GideonZ/1541ultimate#952." + JC-000/1541ultimate branch pr/3-tls-r1; the prebuilt image, in + prerelease tls-preview-f3df8261, is built from f3df8261. Not + upstream; proposed on GideonZ/1541ultimate#952." FIRMWARE_SOURCE="https://github.com/JC-000/1541ultimate/tree/pr/3-tls-r1" FIRMWARE_PRERELEASE="https://github.com/JC-000/1541ultimate/releases/tag/tls-preview-f3df8261" FIRMWARE_IMAGE_SHA="15e9ff7cd8a11daf1f4698d574480381d475db7754e1d276d9aa95e8013ded3c" # The build the "Measured" figures were taken on (the image above). MEASURED_FW="firmware 3.15 git f3df8261 (the image above, ESP module 1.8)" -# Change it once an upstream PR implements $21-$25. +# Change it once an upstream PR implementing $21-$25 is merged. UPSTREAM_PRS="not upstream yet. GideonZ/1541ultimate#952 is the proposal issue; this demo follows its proposed command numbering (\$21 OPEN_TLS, - \$23 TLS_INFO, \$25 TLS_RELEASE). No upstream PR implements these - commands yet (GideonZ/1541ultimate#953 adds GET_RANDOM only)." + \$23 TLS_INFO, \$25 TLS_RELEASE). The commands are open upstream as + issue #995 and draft PR #997, which builds on draft PR #996 (ESP-IDF + 6.0.3, issue #994); none is merged. #953 adds GET_RANDOM only." # Earlier demo images (this script's history, JC-000/c64-https 82a549a to # 9eabd50) were validated on these builds; neither is public. EARLIER_RUNS="esp-tls/m3 f2e46946 and esp-tls/m3-master 3a1ff9ff" From 29743498d049954cc693bf33aa2bf224a5c172d4 Mon Sep 17 00:00:00 2001 From: JC-000 <3798556+JC-000@users.noreply.github.com> Date: Thu, 15 Oct 2026 13:30:53 -0500 Subject: [PATCH 2/4] README: #174 is fixed; make -n/-q/-t no longer delete objects Co-Authored-By: Claude Opus 5.5 (1M context) --- README.md | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/README.md b/README.md index f6b4991..10ed8cf 100644 --- a/README.md +++ b/README.md @@ -450,10 +450,9 @@ object and the PRG before make builds its file database. Two things to know anyway: -- The compare runs at parse time, so on master `make -n` (and `-q`, and `-t`) - with different flags is **not** side-effect-free — it deletes the objects - without rebuilding them. That is issue #174, and a fix is in flight; check - whether it has landed before relying on either behaviour. +- The compare runs at parse time, but `make -n`, `-q` and `-t` are exempt + (issue #174, fixed): with different flags they still `$(warning)` what a + real build would delete, and delete nothing. - `make clean` is still the right move if you want to be certain of a tree for any other reason, and it costs about a second. From 92ad38eaa3878e88f76f0075c483e79ad270d29d Mon Sep 17 00:00:00 2001 From: JC-000 <3798556+JC-000@users.noreply.github.com> Date: Thu, 15 Oct 2026 13:35:07 -0500 Subject: [PATCH 3/4] docs: address review of m3 provenance wording and the make -n note Co-Authored-By: Claude Opus 5.5 (1M context) --- README.md | 9 ++++++--- tools/package/build_m3_demo.sh | 24 +++++++++++++++--------- 2 files changed, 21 insertions(+), 12 deletions(-) diff --git a/README.md b/README.md index 10ed8cf..4bfdb3f 100644 --- a/README.md +++ b/README.md @@ -450,9 +450,12 @@ object and the PRG before make builds its file database. Two things to know anyway: -- The compare runs at parse time, but `make -n`, `-q` and `-t` are exempt - (issue #174, fixed): with different flags they still `$(warning)` what a - real build would delete, and delete nothing. +- The compare runs at parse time. `make -n`, `-q` and `-t` still run it and + `$(warning)` what a real build would delete; they are exempt only from + writing the stamp and deleting objects (issue #174, fixed). With changed + flags `make -n` still prints "Nothing to be done" and `make -q` exits 0: + both describe the old tree, so that stderr warning is the only signal. + A command-line `MAKEFLAGS=` or `MFLAGS=` is a parse-time error. - `make clean` is still the right move if you want to be certain of a tree for any other reason, and it costs about a second. diff --git a/tools/package/build_m3_demo.sh b/tools/package/build_m3_demo.sh index 50ff50b..f80f21b 100755 --- a/tools/package/build_m3_demo.sh +++ b/tools/package/build_m3_demo.sh @@ -49,26 +49,32 @@ MAKE_ARGS=(BACKEND=uci-m3 "HTTPS_HOST=$DEMO_HOST" "HTTPS_PATH=$DEMO_PATH") # What the firmware side published; quoted in the README, checked nowhere # here (the device is not involved in a build). -# FIRMWARE is the firmware side's own one-line description (agreed with -# them); the source branch, the prebuilt image and its sha256 follow. The +# FIRMWARE is the firmware side's one-line description, with the image +# provenance phrased as they suggested on 2026-10-15 ("the +# tls-preview-f3df8261 prerelease is built from f3df8261"); the image's +# source, the prebuilt image and its sha256 follow. The # image line is printed only when FIRMWARE_IMAGE_SHA is set. These and # UPSTREAM_PRS are plain constants, not env knobs, so the README is a # function of the commit: a later commit swaps them. FIRMWARE="TLS client sockets for the Ultimate 64 / U64 Elite, from - JC-000/1541ultimate branch pr/3-tls-r1; the prebuilt image, in - prerelease tls-preview-f3df8261, is built from f3df8261. Not - upstream; proposed on GideonZ/1541ultimate#952." -FIRMWARE_SOURCE="https://github.com/JC-000/1541ultimate/tree/pr/3-tls-r1" + JC-000/1541ultimate; the prebuilt image, in prerelease + tls-preview-f3df8261, is built from f3df8261. Development continues + on branch pr/3-tls-r1, the head of draft PR GideonZ/1541ultimate#997. + Not upstream; proposed on GideonZ/1541ultimate#952." +# The image's own source. The branch pr/3-tls-r1 has moved on since. +FIRMWARE_SOURCE="https://github.com/JC-000/1541ultimate/tree/f3df8261" FIRMWARE_PRERELEASE="https://github.com/JC-000/1541ultimate/releases/tag/tls-preview-f3df8261" FIRMWARE_IMAGE_SHA="15e9ff7cd8a11daf1f4698d574480381d475db7754e1d276d9aa95e8013ded3c" # The build the "Measured" figures were taken on (the image above). MEASURED_FW="firmware 3.15 git f3df8261 (the image above, ESP module 1.8)" -# Change it once an upstream PR implementing $21-$25 is merged. +# Change it when any of GideonZ/1541ultimate #994-#997 changes state +# (merged, out of draft, closed, replaced). UPSTREAM_PRS="not upstream yet. GideonZ/1541ultimate#952 is the proposal issue; this demo follows its proposed command numbering (\$21 OPEN_TLS, \$23 TLS_INFO, \$25 TLS_RELEASE). The commands are open upstream as - issue #995 and draft PR #997, which builds on draft PR #996 (ESP-IDF - 6.0.3, issue #994); none is merged. #953 adds GET_RANDOM only." + GideonZ/1541ultimate issue #995 and draft PR #997, which builds on + draft PR #996 (ESP-IDF 6.0.3, issue #994); none is merged. + GideonZ/1541ultimate#953 adds GET_RANDOM only." # Earlier demo images (this script's history, JC-000/c64-https 82a549a to # 9eabd50) were validated on these builds; neither is public. EARLIER_RUNS="esp-tls/m3 f2e46946 and esp-tls/m3-master 3a1ff9ff" From 1864ad8f409fddb6a0b8c9d00b7b224f7c69c549 Mon Sep 17 00:00:00 2001 From: JC-000 <3798556+JC-000@users.noreply.github.com> Date: Thu, 15 Oct 2026 13:37:15 -0500 Subject: [PATCH 4/4] docs: scope the make -n/-q note to timestamps; full commit in FIRMWARE_SOURCE Co-Authored-By: Claude Opus 5.5 (1M context) --- README.md | 6 ++++-- tools/package/build_m3_demo.sh | 2 +- 2 files changed, 5 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index 4bfdb3f..3dae81a 100644 --- a/README.md +++ b/README.md @@ -453,8 +453,10 @@ Two things to know anyway: - The compare runs at parse time. `make -n`, `-q` and `-t` still run it and `$(warning)` what a real build would delete; they are exempt only from writing the stamp and deleting objects (issue #174, fixed). With changed - flags `make -n` still prints "Nothing to be done" and `make -q` exits 0: - both describe the old tree, so that stderr warning is the only signal. + flags `make -n` lists only what timestamps make stale (on an otherwise + fresh tree, "Nothing to be done") and `make -q` answers as if the flags + had not changed: both describe the old tree, so that stderr warning is + the only signal. A command-line `MAKEFLAGS=` or `MFLAGS=` is a parse-time error. - `make clean` is still the right move if you want to be certain of a tree for any other reason, and it costs about a second. diff --git a/tools/package/build_m3_demo.sh b/tools/package/build_m3_demo.sh index f80f21b..6660a7c 100755 --- a/tools/package/build_m3_demo.sh +++ b/tools/package/build_m3_demo.sh @@ -62,7 +62,7 @@ FIRMWARE="TLS client sockets for the Ultimate 64 / U64 Elite, from on branch pr/3-tls-r1, the head of draft PR GideonZ/1541ultimate#997. Not upstream; proposed on GideonZ/1541ultimate#952." # The image's own source. The branch pr/3-tls-r1 has moved on since. -FIRMWARE_SOURCE="https://github.com/JC-000/1541ultimate/tree/f3df8261" +FIRMWARE_SOURCE="https://github.com/JC-000/1541ultimate/tree/f3df8261b59ea5883cc5b6d9bb1eb4c79f8f0fd4" FIRMWARE_PRERELEASE="https://github.com/JC-000/1541ultimate/releases/tag/tls-preview-f3df8261" FIRMWARE_IMAGE_SHA="15e9ff7cd8a11daf1f4698d574480381d475db7754e1d276d9aa95e8013ded3c" # The build the "Measured" figures were taken on (the image above).