Aura H2O AutoShelf beta

[lemor]

Hello @frostschutz
I used four years (!!!) your mod AutoShelf and in firmware 4.22.15190 it stopped working (maybe due to dosfstools updated).
Please update the mod! It's a shame if you abandon it.

[blackwalrus]

Originally Posted by lemor

Hello @frostschutz
I used four years (!!!) your mod AutoShelf and in firmware 4.22.15190 it stopped working (maybe due to dosfstools updated).
Please update the mod! It's a shame if you abandon it.

Yep, yep!

Also stopped working and I'm really looking forward to AutoShelf updates.

[frostschutz]

Thanks for the report. Kobo updated their version of Busybox, which now complained about a syntax error in the autoshelf script. Please try the 20200701 version and see if it works now.

Regarding other new features of recent firmware updates (SeriesID column) I haven't had the time to take a closer look at it yet, so this is a maintenance only release.

NOTE: Unfortunately as it turns out, the 20200701 is likewise broken on older firmwares. So only use it if you have the latest firmware, otherwise stick to older version. I'll have to find a solution that works for both...

[lemor]

Originally Posted by frostschutz

Thanks for the report. Kobo updated their version of Busybox, which now complained about a syntax error in the autoshelf script. Please try the 20200701 version and see if it works now.

Thank you for your quick response and working on the program.

Version 20200701 works halfway. That is, there is no picture, but the series appeared (Mark7). It would be great if the pictures on / off came back, as it is uncomfortable and embarrassing. Especially those who have not encountered this and simply do not understand that the mod works ...

[Semwize]

Yes, there are no pictures on / off, but the mod works. Just deja vu.

Originally Posted by frostschutz

Regarding other new features of recent firmware updates (SeriesID column) I haven't had the time to take a closer look at it yet, so this is a maintenance only release.

I fixed in line 205 for myself

SET SeriesID='$(escape "$series")', Series='$(escape "$series")', SeriesNumber='$(escape "$number")'

And it worked great. But I'm not sure if this is the best solution.

[frostschutz]

20200701b version fixes the firmware compatibility issue, now works for 4.6 - 4.22.15190 firmware

I'm very clueless why the image does not show since it's fine on my 1st gen H2O, and fbink should support either device...

Do either of you have telnet access and can test if the supplied version of fbink is operational at all?

/usr/local/AutoShelf/fbink hello world

should show hello world message on screen

if that works then it should be another error in my script but I have no idea right now what it could be

if it doesn't work then I guess I'm compiling fbink wrong somehow...

[Semwize]

Originally Posted by frostschutz

Do either of you have telnet access and can test if the supplied version of fbink is operational at all?

/usr/local/AutoShelf/fbink hello world

should show hello world message on screen

if that works then it should be another error in my script but I have no idea right now what it could be

if it doesn't work then I guess I'm compiling fbink wrong somehow...

I made it a lot easier. Pulled fbink from kobo-stuff-1.6.N-r17511

The picture has appeared. Everything is working. Tested version 20200701.

[frostschutz]

Thanks, so it's definitely an issue with my fbink binary... which for some reason, only seems to work for me *scratches head*

edit: I'm an idiot? the executable bit is missing

with any luck, sorted in 20200701c version

[Semwize]

Everything is working. Tested version 20200701.

Guess I spoke too soon. Not everything works, unfortunately.
When the series names contain Cyrillic, the database writes all sorts of nonsense and crashed. I crashed sqlite with ~3500 books, but it was my fault, I did not backup

@frostschutz, correct it please. Previously, everything worked just fine. And when I tested on a small amount, I did not think to check the Cyrillic alphabet. Yes, and did not expect this, frankly.

[frostschutz]

OK, this is a bug in busybox shell. Old version of busybox returned string length in bytes, now it returns string length in characters. However, when using substring indexing, it still expects a byte offset, which is now impossible to determine. Given the wrong byte offset, it ends up cutting into multibyte characters and mangling them.

It worked fine in the old busybox (using bytes everywhere), works fine in Bash (using characters everywhere), doesn't work in current busybox (mixed bytes and characters).

It works fine for ascii only files and folders, as characters and bytes are identical there.

To fix, I'll have to rewrite without relying on the shell string manipulation functions.

[Semwize]

Originally Posted by frostschutz

It works fine for ascii only files and folders, as characters and bytes are identical there.

The whole point of the program is lost, for those who use not ASCII.

[NiLuJe]

@frostschutz: Huh. I can confirm this with a bb built against its own i18n code (as Kobo's libc has broken locales).

(Haven't checked how Kobo's bb is built, as the repos still haven't been updated, and they don't build the bbconfig applet).

I'll check with a bb built against libc, for shit'n giggles, since I do fix the locales on my devices...

(But it does work with ZSH + libc + fixed locales; it counts codepoints).

┌─(ROOT@dahlia:pts/0)──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────(~)─┐
└─(0.06:22%:20:11:91%:#)── BLAH="héhé" && echo ${#BLAH} && echo ${BLAH:2} && echo ${BLAH:3}                                                                                                                                                                                                                                                            ──(Wed, Jul 01)─┘
4
hé
é

[NiLuJe]

Yep, works w/ libc routines:

[root@dahlia root]# BLAH="héhé" && echo ${#BLAH} && echo ${BLAH:2} && echo ${BLAH:3}
4
�hé
hé
[root@dahlia root]# ./busybox_locales ash
[admin@dahlia root]# BLAH="héhé" && echo ${#BLAH} && echo ${BLAH:2} && echo ${BLAH:3}
6
�hé
hé

Except said libc routines are broken by the lack of compiled locales by default .

[NiLuJe]

An easier workaround is a busybox built with its own unicode handling, but w/ CONFIG_FEATURE_CHECK_UNICODE_IN_ENV enabled, and launched without an UTF-8 locale in LANG/LC_ALL.

[root@dahlia root]# ./busybox_env ash
[admin@dahlia root]# BLAH="héhé" && echo ${#BLAH} && echo ${BLAH:2} && echo ${BLAH:3}
6
�hé
hé

[frostschutz]

I have found a workaround, but ran into some other issues too, so update will have to wait until tomorrow or day after.