Patches for 4.28.17925 #90
Assignees
Labels
firmware release
Category: Updating patches for a new release.
Comments
|
On a mildly related subject: this one explained the optional -TFW suffix: it's for when an update ships a touch panel FW update. ┌─(niluje@illyria:pts/13)────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────(/var/tmp/niluje)─┐
└─(0.10:34%:15:41:%)── diff -Nuarp rcS etc/init.d/rcS ──(Thu, Jul 15)─┘
--- rcS 2021-07-15 15:41:59.525885955 +0200
+++ etc/init.d/rcS 2021-07-13 17:27:25.000000000 +0200
@@ -117,6 +117,7 @@ if [ $? != 0 ] || [ $FS_CORRUPT == 1 ] |
write_uboot_env $PLATFORM $UBOOT_RECOVERY
reboot
+ exit
fi
mount -t vfat -o noatime,nodiratime,shortname=mixed,utf8 /dev/mmcblk0p3 /mnt/onboard
@@ -159,6 +160,7 @@ if [ $CAN_UPGRADE == 1 ] && [ -e /mnt/on
write_uboot_env $PLATFORM $UBOOT_MMC
sync
reboot
+ exit
fi
(
@@ -168,22 +170,6 @@ fi
/etc/init.d/on-animator.sh
) &
-if [ $PRODUCT == europa ] && [ -e /lib/firmware/elants_i2c.bin ]; then
- TOUCH_FW_VERSION=`cat /sys/class/i2c-adapter/i2c-0/0-0010/elan_ktf/fw_version`
- if [ "$TOUCH_FW_VERSION" != "6417" ]; then
- echo > /sys/devices/platform/soc/twi0/i2c-0/0-0010/elan_ktf/update_fw
-
- killall -9 on-animator.sh
- sleep 1;
-
- zcat /etc/images/$PREFIX\ghostbuster.raw.gz | /usr/local/Kobo/pickel showpic
- zcat /etc/images/$PREFIX\my_first_time.raw.gz | /usr/local/Kobo/pickel showpic
- sync
- sleep 2;
- poweroff
- fi
-fi
-
(
/usr/local/Kobo/pickel disable.rtc.alarm |
|
Yep, that makes sense. It also explains why this update is only delivered to devices running 17826 (the last TouchFW update). If there are more like this, I'll need to update the firmware page to properly support and display update dependencies. |
|
FWIW, I had to switch away from the (And by "get it", I mean manually ^^). |
pgaskin
added a commit
to pgaskin/kobopatch-testdata
that referenced
this issue
Jul 17, 2021
|
@pgaskin You gave the impression on MR that you were waiting for me to do something with my patches ... whereas I believe I'm waiting for you. I need the resource map text output file for 4.28.17925 before I can start. |
I'd recommend starting with #86, since I expect the recent versions will be similar, and it's easier to keep track of it incrementally. I just started doing the resources, then I'll have the libnickel patches for the last 4 versions done within the next hour or two. |
|
Resource extraction: ./qrc2zip --output "nickel.17925.qInitResources_resources.zip" --recursive --verbose "nickel" 1 $((0x12205d0 - 0x0010000)) $((0x00272a0 - 0x0010000)) $((0x12200e0 - 0x0010000)) || { echo "Error: qrc2zip failed." 1>&2; exit 1; }
./qrc2zip --output "nickel.17925.qInitResources_translations.zip" --recursive --verbose "nickel" 1 $((0x15d89a0 - 0x0010000)) $((0x12209b8 - 0x0010000)) $((0x15d8718 - 0x0010000)) || { echo "Error: qrc2zip failed." 1>&2; exit 1; }
./qrc2zip --output "nickel.17925.qInitResources_styles.zip" --recursive --verbose "nickel" 1 $((0x15f71c0 - 0x0010000)) $((0x15d8ab8 - 0x0010000)) $((0x15f4910 - 0x0010000)) || { echo "Error: qrc2zip failed." 1>&2; exit 1; }
./qrc2zip --output "nickel.17925.qInitResources_certificates.zip" --recursive --verbose "nickel" 1 $((0x161e710 - 0x0010000)) $((0x15f7d40 - 0x0010000)) $((0x161d6c0 - 0x0010000)) || { echo "Error: qrc2zip failed." 1>&2; exit 1; }Output$ git diff --no-index 17826 17925diff --git a/17826/translations/trans_ca_ES.qm b/17925/translations/trans_ca_ES.qm
index 74cc9ae..2109473 100644
Binary files a/17826/translations/trans_ca_ES.qm and b/17925/translations/trans_ca_ES.qm differ
diff --git a/17826/translations/trans_da_DK.qm b/17925/translations/trans_da_DK.qm
index 15cb9f2..0b7665b 100644
Binary files a/17826/translations/trans_da_DK.qm and b/17925/translations/trans_da_DK.qm differ
diff --git a/17826/translations/trans_de_DE.qm b/17925/translations/trans_de_DE.qm
index 426b75a..e9decaf 100644
Binary files a/17826/translations/trans_de_DE.qm and b/17925/translations/trans_de_DE.qm differ
diff --git a/17826/translations/trans_en_US.qm b/17925/translations/trans_en_US.qm
index 22f7dde..6a90566 100644
Binary files a/17826/translations/trans_en_US.qm and b/17925/translations/trans_en_US.qm differ
diff --git a/17826/translations/trans_es_ES.qm b/17925/translations/trans_es_ES.qm
index 1fe9cd2..4f9b580 100644
Binary files a/17826/translations/trans_es_ES.qm and b/17925/translations/trans_es_ES.qm differ
diff --git a/17826/translations/trans_es_MX.qm b/17925/translations/trans_es_MX.qm
index 29d4f50..13d9215 100644
Binary files a/17826/translations/trans_es_MX.qm and b/17925/translations/trans_es_MX.qm differ
diff --git a/17826/translations/trans_fi_FI.qm b/17925/translations/trans_fi_FI.qm
index ef8c842..e615ad6 100644
Binary files a/17826/translations/trans_fi_FI.qm and b/17925/translations/trans_fi_FI.qm differ
diff --git a/17826/translations/trans_fr_FR.qm b/17925/translations/trans_fr_FR.qm
index f6daeb1..15f38ba 100644
Binary files a/17826/translations/trans_fr_FR.qm and b/17925/translations/trans_fr_FR.qm differ
diff --git a/17826/translations/trans_it_IT.qm b/17925/translations/trans_it_IT.qm
index 427e30a..d229368 100644
Binary files a/17826/translations/trans_it_IT.qm and b/17925/translations/trans_it_IT.qm differ
diff --git a/17826/translations/trans_ja_JP.qm b/17925/translations/trans_ja_JP.qm
index 7f518c2..e2a477c 100644
Binary files a/17826/translations/trans_ja_JP.qm and b/17925/translations/trans_ja_JP.qm differ
diff --git a/17826/translations/trans_nb_NO.qm b/17925/translations/trans_nb_NO.qm
index 3e74635..c41237f 100644
Binary files a/17826/translations/trans_nb_NO.qm and b/17925/translations/trans_nb_NO.qm differ
diff --git a/17826/translations/trans_nl_NL.qm b/17925/translations/trans_nl_NL.qm
index 249f3e2..4e985cc 100644
Binary files a/17826/translations/trans_nl_NL.qm and b/17925/translations/trans_nl_NL.qm differ
diff --git a/17826/translations/trans_pt_BR.qm b/17925/translations/trans_pt_BR.qm
index f27e7c5..da99d9d 100644
Binary files a/17826/translations/trans_pt_BR.qm and b/17925/translations/trans_pt_BR.qm differ
diff --git a/17826/translations/trans_pt_PT.qm b/17925/translations/trans_pt_PT.qm
index b61c920..29a8e79 100644
Binary files a/17826/translations/trans_pt_PT.qm and b/17925/translations/trans_pt_PT.qm differ
diff --git a/17826/translations/trans_sv_SE.qm b/17925/translations/trans_sv_SE.qm
index bcda43b..c2c2bca 100644
Binary files a/17826/translations/trans_sv_SE.qm and b/17925/translations/trans_sv_SE.qm differ
diff --git a/17826/translations/trans_tr_TR.qm b/17925/translations/trans_tr_TR.qm
index 5117028..12ce0f3 100644
Binary files a/17826/translations/trans_tr_TR.qm and b/17925/translations/trans_tr_TR.qm differ
diff --git a/17826/translations/trans_zh_CN.qm b/17925/translations/trans_zh_CN.qm
index 475ad05..1ee3c4b 100644
Binary files a/17826/translations/trans_zh_CN.qm and b/17925/translations/trans_zh_CN.qm differ
diff --git a/17826/translations/trans_zh_TW.qm b/17925/translations/trans_zh_TW.qm
index 2a5791a..9181082 100644
Binary files a/17826/translations/trans_zh_TW.qm and b/17925/translations/trans_zh_TW.qm differhttps://krc.storage.pgaskin.net/nickel.17925.qInitResources_certificates.zip |
|
@jackiew1, I've uploaded the resources for 17820+. Note that there are quite a few new resources in 17820 (#88 (comment)). I'm going to do the initial libnickel patch updates within the next hour and a bit so the files are ready for the nickel changes. I'll start with 17623, update those for 17820-17925, test them on 17925 (on my Clara), then push them. |
As of 17826 @ eab3714. The same as usual. |
pgaskin
added a commit
that referenced
this issue
Jul 18, 2021
Notable changes: - "ePub uniform font scale" has been temporarily removed since it will require a rewrite for 17820+.
pgaskin
added a commit
that referenced
this issue
Jul 18, 2021
pgaskin
added a commit
that referenced
this issue
Jul 18, 2021
|
As far as I'm concerned fw versions 4.28.xxxx prior to this one are beta versions and no-one should need a kobopatch pack for them. I don't intend to waste any more time on patches for them. I'm not even convinced 4.28.17925 is anything other than Kobo's most recent beta version. Having said all that I believe all my official nickel patches will run as-is on any of them. The Elipsa has the same screen pixels as the KA1 so I assume all the current Daylight settings will apply. The only patches I know of which did need attention were:
It looks like you've already rewritten the first and removed the second. I did investigate patch Unfortunately, the consequence for this patch is that if we now have to restrict ourselves to patching either part1 or part2 (rather than the 2 parts together) there is no longer quite enough cosmetic whitespace which can be reclaimed to accommodate each of the 4 alternate Replaces. I think I could rewrite the patch so that any of these 3 alts would still work:
That leaves the I don't know how many users use this option (enables ligatures/kerning in kepub) but as it's the only one I, personally, want I plan to use a workaround which gets rid of one of the Thoughts? |
I'm not sure either.
Note that this probably won't be an issue anymore with the next major kobopatch version (late next year at the earliest, though), since I'm planning to implement the ability to move strings to a new memory location, to allocate arbitrary amounts of space in the executable, and to support instruction replacements via trampolines.
That's probably a good idea, and it's also a better fit for the kobopatch way of doing things (multiple individual options in a PatchGroup to make it easier to use overrides and reduce the amount of copy/paste needed).
Yep, that's correct. It would be helpful if you could do a quick test on on some of your devices to determine the optimal multiplier, but it's not an issue if you can't since the current value shouldn't be any worse than it was before.
That's why I'm not testing on them directly. I decided to update the old ones so I could look at the changes incrementally (for both the firmware and the patches), and for completeness. Notice how the last 3 versions took around 7 minutes to update for. If you just want to update the latest one, I'll backport any changes as necessary.
That was my initial guess, but I when I looked into it, I found that dragon is the primary codename. I think it has more to do with the ppi than the resolution. @NiLuJe might have some relevant insight.
If you don't think it should be included in the releases, feel free to put it as a comment here so it doesn't get lost and so it can be referenced if someone wants it specifically. |
No idea, really: I was already puzzled by how the Forma was handled on that front ;). |
I've just realised again how long "a quick test" takes with these patches 😃 I think the multiplier for the old I haven't got the patience to test on a different device tonight. I'll check out what happens on a KA1 (i.e. non-Dragon device) tomorrow. The KA1 didn't used to have a epub/kepub font-size mismatch. As far as I can tell new |
Hopefully that should be reduced whenever I get around to finishing the Kobo emulator...
I was actually referring to the new
👍
👍
Personally, I never understood the point of the font scaling option in that patch, but I think it's there since that patch predates the font scale ones by quite a bit. I'd vote for removing it. Are you planning to add patches for the other options to I'm going to do the final testing for the libnickel patches later tonight, then I'll make the release once you're ready. |
|
Yeah, +1 on removing the duplicate historical font scaling thingy ;). |
|
That's 3-nil for getting rid of it, then. In that case the old
which share the same PatchGroup and which don't need to be customised. I'll put them in For the missing |
|
I'm ready to release the patches. @jackiew1? |
|
I've now tested this fw on a KA1 and can confirm that when unpatched it has the same huge font-size mismatch between epub and kepub as seen on the AuraHD Dragon device. Applying patch When I first bought the KA1 (on launch day) this font-size mismatch didn't exist. When things changed I don't know because I've been using I'm not prepared to load this mark8 4.28.17925 on my beloved Forma but I might have a look later to see whether the official mark7 4.26.16704 also has the font-size mismatch when Release the patches if you really want to. Personally I'd wait until Kobo show some sign that they have a stable mark8 firmware. If they keep releasing new versions every few days I'm not going to be happy re-testing every time. I'm sure we've both got better things to do. As neither of us has an Elipsa (and no prospect of that changing, in my case) troubleshooting may be problematic. |
|
I've got access to one (for now at least), if you need a guinea pig ;). (I won't be available next week, though). |
Let's hope you won't regret volunteering 😄 |
Well, "later" turned out to be "quite soon". In fw 4.26.16704 the Forma has no font-size mismatch between epub/kepub when both |
It's due to the new method of calculating the dpi multiplier in 17820 (see #88 (comment)) which gets the DPI from the QScreen (which might be incorrect; @NiLuJe?).
Yep, that's OK. If it continues, I'll just update the patches (the usual updates require around 10-15 minutes excluding the automatic analysis time; it's not a multi-hour process anymore) without doing the full tests/releases for each version.
I'll do it tonight. |
|
Released: |
$ git diff --no-index src/versions/{4.28.17826,4.28.17925}diff --git a/src/versions/4.28.17826/libnickel.so.1.0.0.yaml/geoffr.yaml b/src/versions/4.28.17925/libnickel.so.1.0.0.yaml/geoffr.yaml
index 0d9e843..ec96194 100644
--- a/src/versions/4.28.17826/libnickel.so.1.0.0.yaml/geoffr.yaml
+++ b/src/versions/4.28.17925/libnickel.so.1.0.0.yaml/geoffr.yaml
@@ -491,5 +491,5 @@ Set font scale factor:
# Original value X=14, so original scale factor is (font size slider) / 46.
# Allowed replacement values for X: 0-15 (smaller X --> larger scale factor)
# Note: vmov.f32 s1, #15.0 before getReadingFontSizeScaleFactor
- - ReplaceInt: {Offset: 0x5E74A6, Find: 14, Replace: 4} # ePub
- - ReplaceInt: {Offset: 0xC5C29A, Find: 14, Replace: 4} # KePub
+ - ReplaceInt: {Offset: 0x5E73FE, Find: 14, Replace: 4} # ePub
+ - ReplaceInt: {Offset: 0xC5CB82, Find: 14, Replace: 4} # KePub
diff --git a/src/versions/4.28.17826/libnickel.so.1.0.0.yaml/pgaskin.yaml b/src/versions/4.28.17925/libnickel.so.1.0.0.yaml/pgaskin.yaml
index d70cfd1..e7de9a6 100644
--- a/src/versions/4.28.17826/libnickel.so.1.0.0.yaml/pgaskin.yaml
+++ b/src/versions/4.28.17925/libnickel.so.1.0.0.yaml/pgaskin.yaml
@@ -553,7 +553,7 @@ ePub uniform font scale:
working on 4.28.17820 since the dpi was changed to use QScreen instead of
144dpi on dragon and 96dpi on other devices (GeoffR's patch forced it to
always be 96dpi). This patch has not been tested on the Elipsa.
- - BaseAddress: 0x5E74CE
+ - BaseAddress: 0x5E7426
- ReplaceBytes:
Offset: 0
FindInstBLX: {SymPLT: "QGuiApplication::primaryScreen()"} # unused
@@ -592,7 +592,7 @@ Customize ComfortLight settings:
# In an unnamed subroutine two layers into FrontLightPopupController::loadView
# (find it by going back from QTime::addSecs), the times for the dropdown are
# generated into a QVector<QPair<QString, QTime>> with a simple loop.
- - BaseAddress: 0xBAE960 # find the base of the unnamed subroutine with the x-ref to _ZN5QTimeC1Eiiii
+ - BaseAddress: 0xBAF248 # find the base of the unnamed subroutine with the x-ref to _ZN5QTimeC1Eiiii
#
# Change the initial hour / first bedtime dropdown item passed to the QTime
# constructor (mov r1, #21):
diff --git a/src/versions/4.28.17826/nickel.yaml/oren64.yaml b/src/versions/4.28.17925/nickel.yaml/oren64.yaml
index 16e57fe..5184928 100644
--- a/src/versions/4.28.17826/nickel.yaml/oren64.yaml
+++ b/src/versions/4.28.17925/nickel.yaml/oren64.yaml
@@ -27,7 +27,7 @@ New home screen subtitle custom font:
- Description: |
Increase subtitle font size, font color black. Works best when used
with with patch `New home screen increasing cover size`
- - BaseAddress: 0x15E1E38 # qss/GenericHomeWidget.qss
+ - BaseAddress: 0x15DED55 # qss/GenericHomeWidget.qss
- ReplaceZlib: {Find: "MetaDataLabel {\n color: #666666;\n qproperty-indent: 0;\n}", Replace: "MetaDataLabel {\n color: #000000;\n qproperty-indent: 0;\n}"}
- ReplaceZlib:
Find: "MetaDataLabel[qApp_deviceIsTrilogy=true] {\n font-size: 13px;\n}"
diff --git a/src/versions/4.28.17826/nickel.yaml/pgaskin.yaml b/src/versions/4.28.17925/nickel.yaml/pgaskin.yaml
index 1bcc091..9f05039 100644
--- a/src/versions/4.28.17826/nickel.yaml/pgaskin.yaml
+++ b/src/versions/4.28.17925/nickel.yaml/pgaskin.yaml
@@ -15,7 +15,7 @@ Remove forgot pin button from lock screen:
Removes the Forgot PIN -> Sign Out button from the lock screen.
If this is enabled and you forget your pin, you will need to
hard reset your Kobo.
- - BaseAddress: 0x15D1D42 # qss/PinCodeInputDialog.qss
+ - BaseAddress: 0x15D8400 # qss/PinCodeInputDialog.qss
- ReplaceZlib:
Find: "#lblForgotPin[qApp_deviceIsDragon=true],\n#lblSignOut[qApp_deviceIsDragon=true] {\n font-size: 26px;\n}"
Replace: "#lblForgotPin,#lblSignOut{qproperty-visible:false;}" |
|
Lab rat reporting in ;p. I applied my usual set on the Elipsa a few hours ago, nothing horribly blew up ;). |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
ePub uniform font scale? (@pgaskin)The text was updated successfully, but these errors were encountered: