Initial Kindle PW5 support #8856

NiLuJe · 2022-02-27T21:54:59Z

Requires koreader/koreader-base#1457

(Oh, hello there, "mystery" device ;p).

Rejig frontlight warmth API to more closely match the existing API, and, hopefully, clarify some of its quirks, and reduce boilerplate and duplicate code in platform implementations.
Tweak Kindle:setDateTime to prefer using the platform's custom script, as in interacts better with the stock UI. And make the fallbacks handle old busybox versions better.
Add Kindle PW5 support ;).
Add warmth support to the Kindle platform.
Random TextBoxWidget cleanups: make sure we immediately free destroyed instances.
FrontLightWidget: Refactor to make it slightly less obnoxious to grok and update; i.e., separate layout from update, and properly separate brightness from warmth handling. Move to simpler widgets instead of reinventing the wheel.
TextBoxWidgets: Implement setText to match TextWidget's API, as some callers may be using the two interchangeably (i.e., Button).
NaturalLightWidget: Make sure we pass a string to InputText
InputText: Add debug guards to catch bad callers not passing strings ;).

This change is

NiLuJe · 2022-02-27T21:58:16Z

Pending a minor refactor of the FL widget because that thing's a mess.

Stuff requiring further investigation:

Driver really likes optimizing DU updates out, so flashing_ui barely flashes ^^.
Exiting KOReader currently exhibits symptoms similar to what Kindle: don't kill kb service on start #8122 fixed for earlier FW versions (namely, long UI stalls). Haven't had time to look into it yet, and low priority as rebooting works. (It's still a Kindle, though, so, rebooting still takes an assload of time).

yparitcher · 2022-02-28T15:38:01Z

(Oh, hello there, "mystery" device ;p).

Did you find an exploit to jailbreak, or was entry via serial? :)

NiLuJe · 2022-02-28T15:53:31Z

I'm not the exploit finder, and it won't be published until Amazon manages to ship a patched FW, because responsible disclosure and all that ;).

But, no, it's a good old chain of sneaky *nix trickery, quite fun :).

AFAIK, serial is a no-go on recent devices, secure boot and all that jazz :/.

yparitcher · 2022-02-28T16:08:15Z

time to buy one now before they ship with updated firmware :) as much as i hate thier software, amazon does come out with some of the best hardware.

Frenzie · 2022-02-28T16:40:04Z

Disagree, 8 GB storage with no (micro)-SD is grossly underspecced and nigh unusable.

Edit: to be clear, 32 GB on the Kobo Libra 2 is also an affront. But, you know, it's not 8 GB.

NiLuJe · 2022-02-28T17:43:32Z

The Signature has 32GB storage, IIRC (it's wildly too expensive, though ;)).

Frenzie · 2022-03-06T10:08:01Z

frontend/ui/widget/textboxwidget.lua

+function TextBoxWidget:setText(text)
+    if text == self.text then
+        return
+    end
+
+    self.text = text
+
+    if self.use_xtext then
+        self:_measureWithXText()
+    else
+        self:_evalCharWidthList()
+    end
+    self:_splitToLines()
+
+    self:update()
+end
+dbg:guard(TextBoxWidget, "setText",
+    function(self, text)
+        assert(type(text) == "string",
+            "Wrong text type (expected string)")
+    end)


What's this sneaky stuff? :-P

To match TextWidget's API, because Button might call it since the multi-line support pr ;).

(Also, I needed it too ;p).

poire-z · 2022-03-06T10:29:58Z

frontend/ui/widget/textboxwidget.lua

+function TextBoxWidget:setText(text)
+    if text == self.text then
+        return
+    end
+
+    self.text = text
+
+    if self.use_xtext then
+        self:_measureWithXText()
+    else
+        self:_evalCharWidthList()
+    end
+    self:_splitToLines()
+
+    self:update()


Not sure I would trust this. Seems a few more things may be done in :init() like dealing with height, that doing this quick fake-init() might mess ?

yparitcher · 2022-03-06T15:33:28Z

Disagree, 8 GB storage with no (micro)-SD is grossly underspecced and nigh unusable.

Edit: to be clear, 32 GB on the Kobo Libra 2 is also an affront. But, you know, it's not 8 GB.

To me the most important thing is to fit in my pocket :)
Otherwise it is is not portable,
so a bigger screen while still fitting in my pocket, kinda outweighs storage concerns.
no one else make a pocketable > 6" screen

Frenzie · 2022-03-06T16:16:04Z

Big pockets. ;-)

…

On Sun, Mar 6, 2022 at 4:33 PM yparitcher ***@***.***> wrote: Disagree, 8 GB storage with no (micro)-SD is grossly underspecced and nigh unusable. Edit: to be clear, 32 GB on the Kobo Libra 2 is also an affront. But, you know, it's not 8 GB. To me the most important thing is to fit in my pocket :) Otherwise it is is not portable, so a bigger screen while still fitting in my pocket, kinda outweighs storage concerns. no one else make a pocketable > 6" screen — Reply to this email directly, view it on GitHub <#8856 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABRQBPSKDAWTDL3WGKUSSLU6TF5LANCNFSM5PPP3H7Q> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you commented.Message ID: ***@***.***>

poire-z · 2022-03-06T20:06:35Z

frontend/ui/widget/textboxwidget.lua

+    self:_computeDimensions()
+    self:_updateLayout()
+    if self.editable then
+        self:moveCursorToCharPos(self.charpos or 1)
+    end
+    self.dimen = Geom:new(self:getSize())


I'm a bit unconfortable with these kind of changes. It might be all right, but after a few minutes looking around, I can't say I'm 100% sure - it's less readable that how it was before (_computeDimensions actually renders the text, so this naming is hiding abit the whole core of what's really happening).
And what about this self.dimen here that won't get updated by your setText() if you give it larger text that makes a few more lines ?

I'd be more confortable if you just did in your setText() : free() and then :init(). That's what we do in InputText:initTextBox() - it would feel less risky, and the wee additional recomputing this would do vs. your splitting would be bearable for the rare users of this new setText().

But I like catching design/implementation bugs with stuff like that ;o).

NiLuJe · 2022-03-07T00:00:50Z

Note to self: Mostly done testing this on the KOReader side (as in, haven't looked into the driver merging behavior nor the stock UI hangs).

Note to self: check if those aren't simply wait_for_any timeouts.

Note to self: and/or weird interaction with the new React UI, because there be lotsa new services I know nothing about running.

So, just have to check that the warmth API tweaks didn't horribly break on Kobo ;o).

poire-z · 2022-03-07T00:29:46Z

frontend/ui/widget/textboxwidget.lua

@@ -151,6 +152,26 @@ function TextBoxWidget:init()
        self.text_height = self.lines_per_page * self.line_height_px
    end

+    self:_computeTextDimensions()


Still, we're doing more than "computing text dimensions": we render the text...

And now that you're not even using this TextBoxWidget in FrontlightWidget, it feels unnecessary to hack this module and leave in there some code paths we're not even taking.

That's kind of implied by the verb (same as _measureText) ;).

There's nothing really left of concern in init vs this, AFAICT.

And Button can take this path, so I'll take something that appeared to work just fine over a crash ;).

There's nothing really left of concern in init vs this

Except readability (not having to jump around to see what's happening) and messing my git history :)

And Button can take this path, so I'll take something that appeared to work just fine over a crash

Which you could have solved with a dumber setText() doing :free() + :init().

frontend/ui/widget/textboxwidget.lua

frontend/device/kindle/device.lua

NiLuJe · 2022-03-08T18:37:24Z

Note to self: check if those aren't simply wait_for_any timeouts.

Huh. Good hunch ;).

[pid  2089] 18:42:05.604305 [b69d78dc] ioctl(19, MXCFB_SET_PREFETCH or MXCFB_SET_PWRDOWN_DELAY, 0) = 0 <0.000210>
[pid  2174] 18:42:05.615880 [b69d78dc] ioctl(19, MXCFB_SEND_UPDATE, {update_region={top=1492, left=567, width=1, height=1}, waveform_mode=WAVEFORM_MODE_AUTO => WAVEFORM_MODE_AUTO, update_mode=UPDATE_MODE_PARTIAL, update_marker=320, temp=TEMP_USE_AMBIENT, flags=0, dither_mode=EPDC_FLAG_USE_DITHERING_PASSTHROUGH, quant_bit=0, alt_buffer_data={phys_addr=NULL, width=0, height=0, alt_update_region={top=0, left=0, width=0, height=0}}, swipe_data={direction=MTK_SWIPE_DOWN, steps=0}, hist_bw_waveform_mode=MTK_WAVEFORM_MODE_DU, hist_gray_waveform_mode=MTK_WAVEFORM_MODE_GC16, ts_pxp=0, ts_epdc=0}) = 0 <0.001632>
[pid  2177] 18:42:05.618518 [b69d78dc] ioctl(19, MXCFB_WAIT_FOR_ANY_UPDATE_COMPLETE, {flag=FLAG_NONE <unfinished ...>
[pid  2174] 18:42:05.659475 [b69d78dc] ioctl(19, MXCFB_SEND_UPDATE, {update_region={top=1492, left=567, width=1, height=1}, waveform_mode=WAVEFORM_MODE_AUTO => WAVEFORM_MODE_AUTO, update_mode=UPDATE_MODE_PARTIAL, update_marker=321, temp=TEMP_USE_AMBIENT, flags=0, dither_mode=EPDC_FLAG_USE_DITHERING_PASSTHROUGH, quant_bit=0, alt_buffer_data={phys_addr=NULL, width=0, height=0, alt_update_region={top=0, left=0, width=0, height=0}}, swipe_data={direction=MTK_SWIPE_DOWN, steps=0}, hist_bw_waveform_mode=MTK_WAVEFORM_MODE_DU, hist_gray_waveform_mode=MTK_WAVEFORM_MODE_GC16, ts_pxp=0, ts_epdc=0}) = 0 <0.006690>
[pid  2089] 18:42:05.698332 [b69d78dc] ioctl(19, MXCFB_WAIT_FOR_UPDATE_SUBMISSION, 321) = 0 <0.000480>
[pid  2089] 18:42:07.046736 [b69d78dc] ioctl(19, MXCFB_WAIT_FOR_UPDATE_COMPLETE, {update_marker=321, collision_test=3064201216 => 3064201216}) = 0 <0.000073>
[pid  2089] 18:42:07.047140 [b69d78dc] ioctl(19, MXCFB_SEND_UPDATE, {update_region={top=0, left=0, width=1236, height=1648}, waveform_mode=MTK_WAVEFORM_MODE_GC16 => MTK_WAVEFORM_MODE_GC16, update_mode=UPDATE_MODE_FULL, update_marker=322, temp=TEMP_USE_AMBIENT, flags=0, dither_mode=EPDC_FLAG_USE_DITHERING_PASSTHROUGH, quant_bit=0, alt_buffer_data={phys_addr=NULL, width=0, height=0, alt_update_region={top=0, left=0, width=0, height=0}}, swipe_data={direction=MTK_SWIPE_DOWN, steps=0}, hist_bw_waveform_mode=MTK_WAVEFORM_MODE_DU, hist_gray_waveform_mode=MTK_WAVEFORM_MODE_GC16, ts_pxp=0, ts_epdc=0}) = 0 <0.000788>
[pid  2177] 18:42:07.691264 [b69d78dc] <... ioctl resumed> => markers=[]}) = 0 <2.072520>
[pid  2177] 18:42:07.692982 [b69d78dc] ioctl(19, MXCFB_WAIT_FOR_ANY_UPDATE_COMPLETE, {flag=FLAG_NONE => markers=[]}) = 0 <2.078578>

Not quite sure what the best way to handle that will be...

(Especially since that kernel facility seems to be a moving target, it might not behave in the same way on 5.14.2)

(We cannot support USBMS, so doing FM stuff on UsbPlugOut makes no sense).

Untested.

It's a custom wrapper that will update the stock UI properly, too. (In addition to doing the date + hwclock dance)

Handle it via a sysfs poke, like on Cervantes.

Always honor the native scale, and dislpay it as-is

barely ever called on most platforms.

Might help w/ the ioctl timeouts?

NiLuJe · 2022-03-08T22:32:58Z

Note to self: check if those aren't simply wait_for_any timeouts.

Huh. Good hunch ;).

[pid  2089] 18:42:05.604305 [b69d78dc] ioctl(19, MXCFB_SET_PREFETCH or MXCFB_SET_PWRDOWN_DELAY, 0) = 0 <0.000210>
[pid  2174] 18:42:05.615880 [b69d78dc] ioctl(19, MXCFB_SEND_UPDATE, {update_region={top=1492, left=567, width=1, height=1}, waveform_mode=WAVEFORM_MODE_AUTO => WAVEFORM_MODE_AUTO, update_mode=UPDATE_MODE_PARTIAL, update_marker=320, temp=TEMP_USE_AMBIENT, flags=0, dither_mode=EPDC_FLAG_USE_DITHERING_PASSTHROUGH, quant_bit=0, alt_buffer_data={phys_addr=NULL, width=0, height=0, alt_update_region={top=0, left=0, width=0, height=0}}, swipe_data={direction=MTK_SWIPE_DOWN, steps=0}, hist_bw_waveform_mode=MTK_WAVEFORM_MODE_DU, hist_gray_waveform_mode=MTK_WAVEFORM_MODE_GC16, ts_pxp=0, ts_epdc=0}) = 0 <0.001632>
[pid  2177] 18:42:05.618518 [b69d78dc] ioctl(19, MXCFB_WAIT_FOR_ANY_UPDATE_COMPLETE, {flag=FLAG_NONE <unfinished ...>
[pid  2174] 18:42:05.659475 [b69d78dc] ioctl(19, MXCFB_SEND_UPDATE, {update_region={top=1492, left=567, width=1, height=1}, waveform_mode=WAVEFORM_MODE_AUTO => WAVEFORM_MODE_AUTO, update_mode=UPDATE_MODE_PARTIAL, update_marker=321, temp=TEMP_USE_AMBIENT, flags=0, dither_mode=EPDC_FLAG_USE_DITHERING_PASSTHROUGH, quant_bit=0, alt_buffer_data={phys_addr=NULL, width=0, height=0, alt_update_region={top=0, left=0, width=0, height=0}}, swipe_data={direction=MTK_SWIPE_DOWN, steps=0}, hist_bw_waveform_mode=MTK_WAVEFORM_MODE_DU, hist_gray_waveform_mode=MTK_WAVEFORM_MODE_GC16, ts_pxp=0, ts_epdc=0}) = 0 <0.006690>
[pid  2089] 18:42:05.698332 [b69d78dc] ioctl(19, MXCFB_WAIT_FOR_UPDATE_SUBMISSION, 321) = 0 <0.000480>
[pid  2089] 18:42:07.046736 [b69d78dc] ioctl(19, MXCFB_WAIT_FOR_UPDATE_COMPLETE, {update_marker=321, collision_test=3064201216 => 3064201216}) = 0 <0.000073>
[pid  2089] 18:42:07.047140 [b69d78dc] ioctl(19, MXCFB_SEND_UPDATE, {update_region={top=0, left=0, width=1236, height=1648}, waveform_mode=MTK_WAVEFORM_MODE_GC16 => MTK_WAVEFORM_MODE_GC16, update_mode=UPDATE_MODE_FULL, update_marker=322, temp=TEMP_USE_AMBIENT, flags=0, dither_mode=EPDC_FLAG_USE_DITHERING_PASSTHROUGH, quant_bit=0, alt_buffer_data={phys_addr=NULL, width=0, height=0, alt_update_region={top=0, left=0, width=0, height=0}}, swipe_data={direction=MTK_SWIPE_DOWN, steps=0}, hist_bw_waveform_mode=MTK_WAVEFORM_MODE_DU, hist_gray_waveform_mode=MTK_WAVEFORM_MODE_GC16, ts_pxp=0, ts_epdc=0}) = 0 <0.000788>
[pid  2177] 18:42:07.691264 [b69d78dc] <... ioctl resumed> => markers=[]}) = 0 <2.072520>
[pid  2177] 18:42:07.692982 [b69d78dc] ioctl(19, MXCFB_WAIT_FOR_ANY_UPDATE_COMPLETE, {flag=FLAG_NONE => markers=[]}) = 0 <2.078578>

Not quite sure what the best way to handle that will be...

(Especially since that kernel facility seems to be a moving target, it might not behave in the same way on 5.14.2)

Whelp, not forgetting to disable the "fast mode" thingy on exit appears to have helped, yay.

setWarmth / self.fl_warmth are "public" (-ish, as far as the second one is concerned), and always in the [0, 100] range. setWarmthHW is private, local to each implementation, and always uses the platform's native scale.

NiLuJe

Reviewed 9 of 16 files at r1, 5 of 8 files at r2, 3 of 4 files at r4, 1 of 1 files at r6, 8 of 8 files at r7, 3 of 3 files at r8, all commit messages.
Reviewable status: all files reviewed, 6 unresolved discussions (waiting on @Frenzie, @NiLuJe, @poire-z, and @yparitcher)

frontend/device/kindle/device.lua, line 64 at r6 (raw file):

Previously, yparitcher wrote…

This should probably be named isWifiOn or isWifiUp as it only checks if the wireless interface is up not if it is connected.

Done.

NiLuJe · 2022-03-08T22:56:08Z

Whelp, finally done with this :).

zwim · 2022-03-12T21:43:09Z

One question: What does the now not so mysterious device show on cat /sys/power/state?

NiLuJe · 2022-03-12T23:31:37Z

It might not be directly useful, for a couple reasons:

We leave power management to the stock system on Kindle.
Said daemon implements a two-tiered sleep state tracker, and on most recent devices, the deeper sleep state involves a custom-ish suspend to disk on a custom bit of storage hardware (or... something, never really looked into it).

That said, I'll give it a poke the next time it's plugged ;).

(What I can say is that it's hilariously more power efficient, even completely awake: a Sage/Elipsa in suspend drains more power than an idle PW5 at reasonable light levels (!). Computational spikes also drain way, way less power).

yparitcher · 2022-03-13T01:46:31Z

Just got my PW5 (5.14.0.1) It (just) fits my pocket :) waiting for the JB to try this wink wink ;)

InputText: Add debug guards to catch bad callers not passing strings ;). Fix koreader#8899

koreader/koreader-base#1457

* Rejig frontlight warmth API to more closely match the existing API, and, hopefully, clarify some of its quirks, and reduce boilerplate and duplicate code in platform implementations. * Tweak Kindle:setDateTime to prefer using the platform's custom script, as in interacts better with the stock UI. And make the fallbacks handle old busybox versions better. * Add Kindle PW5 support ;). * Add warmth support to the Kindle platform. * Random TextBoxWidget cleanups: make sure we immediately free destroyed instances. * FrontLightWidget: Refactor to make it slightly less obnoxious to grok and update; i.e., separate layout from update, and properly separate brightness from warmth handling. Move to simpler widgets instead of reinventing the wheel. * TextBoxWidgets: Implement `setText` to match TextWidget's API, as some callers may be using the two interchangeably (i.e., Button). * NaturalLightWidget: Make sure we pass a string to InputText * InputText: Add debug guards to catch bad callers not passing strings ;).

of granularity. Fix koreader#8913 Regression since koreader#8856

Fix #8913 Regression since #8856 Change tacks, to allow full granularity with the +/- buttons.

hius07 · 2022-04-08T05:52:01Z

frontend/device/kindle/device.lua

@@ -104,6 +126,7 @@ local Kindle = Generic:new{
    -- NOTE: We can cheat by adding a platform-specific entry here, because the only code that will check for this is here.
    isSpecialOffers = isSpecialOffers(),
    hasOTAUpdates = yes,
+    hasFastWifiStatusQuery = yes,


Good morning, thank you for that especially.

Hey, good to see you (alive) ! How are you ?
(And where were/are you living? We all got very faimiliar with Ukraine geography lately :/)

Thank you, so-so, still migrating. Hope to appear here more frequently if I can.

Frenzie added this to the 2022.03 milestone Feb 27, 2022

Frenzie added the Kindle label Feb 27, 2022

NiLuJe force-pushed the kindle_pw5 branch from 52b5727 to 43c0cdf Compare February 28, 2022 17:40

NiLuJe force-pushed the kindle_pw5 branch from 10ff071 to 40fd360 Compare March 5, 2022 18:54

Frenzie reviewed Mar 6, 2022

View reviewed changes

poire-z reviewed Mar 6, 2022

View reviewed changes

poire-z reviewed Mar 7, 2022

View reviewed changes

frontend/ui/widget/textboxwidget.lua Outdated Show resolved Hide resolved

yparitcher reviewed Mar 7, 2022

View reviewed changes

frontend/device/kindle/device.lua Outdated Show resolved Hide resolved

NiLuJe added 10 commits March 8, 2022 22:48

Add a few FIXMEs, and kill an olf and meaningless TODO

4be332a

(We cannot support USBMS, so doing FM stuff on UsbPlugOut makes no sense).

Minimal PW5 support

e3fd477

Untested.

Unbreak input

6b75796

Kindle: Use setdate in setDateTime

3cac3e0

It's a custom wrapper that will update the stock UI properly, too. (In addition to doing the date + hwclock dance)

Explicitly mention that HW dithering doesn't do anything right now

8c11364

Kindle: Enable hasFastWifiStatusQuery

5bfddda

Handle it via a sysfs poke, like on Cervantes.

Finish PW5 port w/ warmth support

da8105c

Unbreak the FL widget

10978e3

Get rid of the weird "warmth must be [0...100]" insanity

9337acf

Always honor the native scale, and dislpay it as-is

NL: Step 0 doesn't require a block

908261a

NiLuJe added 4 commits March 8, 2022 22:48

Squish debug

e709ce2

Review pass

d80b9d8

Stress the fact that frontlightIntensityHW and frontlightWarmthHW are

84abe54

barely ever called on most platforms.

Luacheck'ed

b2d82c6

NiLuJe force-pushed the kindle_pw5 branch from 095efe8 to b2d82c6 Compare March 8, 2022 21:48

NiLuJe added 3 commits March 8, 2022 23:05

More reviewing

1183606

Disable "fast" mode on exit

d2a652e

Might help w/ the ioctl timeouts?

Oops, we need the fb fd to send an ioctl ;o).

913c2c5

NiLuJe added 3 commits March 8, 2022 23:38

That's already in the native scale

2b7489a

More explicit comment

c1b7601

setWarmth / self.fl_warmth are "public" (-ish, as far as the second one is concerned), and always in the [0, 100] range. setWarmthHW is private, local to each implementation, and always uses the platform's native scale.

Nope, can't enable wamrth on the KOA3 on my lonesome ;).

4c26331

NiLuJe commented Mar 8, 2022

View reviewed changes

NiLuJe marked this pull request as ready for review March 8, 2022 22:55

NiLuJe requested a review from pazos as a code owner March 8, 2022 22:55

NaturalLightWidget: Make sure we pass a string to InputText

fa27676

InputText: Add debug guards to catch bad callers not passing strings ;). Fix koreader#8899

Frenzie mentioned this pull request Mar 14, 2022

Changing RGB lights in kobo aura one now crashes, it didn't do this before #8899

Closed

Bump base

38d4e35

koreader/koreader-base#1457

NiLuJe merged commit 217a73f into koreader:master Mar 14, 2022

octavianx mentioned this pull request Mar 20, 2022

false alarm #8925

Closed

NiLuJe added a commit to NiLuJe/koreader that referenced this pull request Mar 21, 2022

FrontLightWidget: Unbreak warmth bar on devices with more than 25 steps

80263d1

of granularity. Fix koreader#8913 Regression since koreader#8856

yparitcher mentioned this pull request Mar 22, 2022

PW4: Night Mode does not survive sleep #8937

Closed

Frenzie pushed a commit that referenced this pull request Mar 22, 2022

FrontLightWidget: Unbreak warmth on full-granularity devices (#8935)

d761b62

Fix #8913 Regression since #8856 Change tacks, to allow full granularity with the +/- buttons.

hius07 reviewed Apr 8, 2022

View reviewed changes

yparitcher mentioned this pull request Feb 17, 2023

fix broken escaping #10131

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Initial Kindle PW5 support #8856

Initial Kindle PW5 support #8856

NiLuJe commented Feb 27, 2022 •

edited

Loading

NiLuJe commented Feb 27, 2022 •

edited

Loading

yparitcher commented Feb 28, 2022

NiLuJe commented Feb 28, 2022

yparitcher commented Feb 28, 2022 via email

Frenzie commented Feb 28, 2022 •

edited

Loading

NiLuJe commented Feb 28, 2022

Frenzie Mar 6, 2022

NiLuJe Mar 6, 2022 •

edited

Loading

poire-z Mar 6, 2022

yparitcher commented Mar 6, 2022

Frenzie commented Mar 6, 2022 via email

poire-z Mar 6, 2022

This comment was marked as duplicate.

NiLuJe Mar 6, 2022

NiLuJe commented Mar 7, 2022 •

edited

Loading

poire-z Mar 7, 2022

NiLuJe Mar 7, 2022 •

edited

Loading

poire-z Mar 7, 2022

NiLuJe commented Mar 8, 2022 •

edited

Loading

NiLuJe commented Mar 8, 2022

NiLuJe left a comment

NiLuJe commented Mar 8, 2022

zwim commented Mar 12, 2022

NiLuJe commented Mar 12, 2022 •

edited

Loading

yparitcher commented Mar 13, 2022

hius07 Apr 8, 2022

poire-z Apr 8, 2022

hius07 Apr 8, 2022

Initial Kindle PW5 support #8856

Initial Kindle PW5 support #8856

Conversation

NiLuJe commented Feb 27, 2022 • edited Loading

NiLuJe commented Feb 27, 2022 • edited Loading

yparitcher commented Feb 28, 2022

NiLuJe commented Feb 28, 2022

yparitcher commented Feb 28, 2022 via email

Frenzie commented Feb 28, 2022 • edited Loading

NiLuJe commented Feb 28, 2022

Choose a reason for hiding this comment

NiLuJe Mar 6, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

yparitcher commented Mar 6, 2022

Frenzie commented Mar 6, 2022 via email

Choose a reason for hiding this comment

This comment was marked as duplicate.

Choose a reason for hiding this comment

NiLuJe commented Mar 7, 2022 • edited Loading

Choose a reason for hiding this comment

NiLuJe Mar 7, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

NiLuJe commented Mar 8, 2022 • edited Loading

NiLuJe commented Mar 8, 2022

NiLuJe left a comment

Choose a reason for hiding this comment

NiLuJe commented Mar 8, 2022

zwim commented Mar 12, 2022

NiLuJe commented Mar 12, 2022 • edited Loading

yparitcher commented Mar 13, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

NiLuJe commented Feb 27, 2022 •

edited

Loading

NiLuJe commented Feb 27, 2022 •

edited

Loading

Frenzie commented Feb 28, 2022 •

edited

Loading

NiLuJe Mar 6, 2022 •

edited

Loading

NiLuJe commented Mar 7, 2022 •

edited

Loading

NiLuJe Mar 7, 2022 •

edited

Loading

NiLuJe commented Mar 8, 2022 •

edited

Loading

NiLuJe commented Mar 12, 2022 •

edited

Loading