[kbd] Test failures in kbd-2.0.0 on ppc

[kbd] Test failures in kbd-2.0.0 on ppc

[Ken Moffat]

[-- Attachment #1: Type: text/plain, Size: 2746 bytes --]

 I'm in the middle of trying to build a new system on my old mac G5
(ppc64, but using linux32 to compile userspace) and the kbd tests
failed with:

Making check in tests
make[1]: Entering directory `/building/kbd-2.0.0/tests'
make  libkeymap-init libkeymap-kmap libkeymap-keys libkeymap-parse libkeymap-charset dumpkeys-fulltable dumpkeys-mktable dumpkeys-bkeymap
make[2]: Entering directory `/building/kbd-2.0.0/tests'
make[2]: `libkeymap-init' is up to date.
make[2]: `libkeymap-kmap' is up to date.
make[2]: `libkeymap-keys' is up to date.
make[2]: `libkeymap-parse' is up to date.
make[2]: `libkeymap-charset' is up to date.
sed \
	-e 's,2.0.0,2.0.0,g' \
	-e 's,@DATADIR@,.,g' \
	-e 's,@BUILDDIR@,.,g' \
	<dumpkeys-fulltable.in >./dumpkeys-fulltable
touch -r dumpkeys-fulltable.in ./dumpkeys-fulltable
chmod 755 ./dumpkeys-fulltable
sed \
	-e 's,2.0.0,2.0.0,g' \
	-e 's,@DATADIR@,.,g' \
	-e 's,@BUILDDIR@,.,g' \
	<dumpkeys-mktable.in >./dumpkeys-mktable
touch -r dumpkeys-mktable.in ./dumpkeys-mktable
chmod 755 ./dumpkeys-mktable
sed \
	-e 's,2.0.0,2.0.0,g' \
	-e 's,@DATADIR@,.,g' \
	-e 's,@BUILDDIR@,.,g' \
	<dumpkeys-bkeymap.in >./dumpkeys-bkeymap
touch -r dumpkeys-bkeymap.in ./dumpkeys-bkeymap
chmod 755 ./dumpkeys-bkeymap
make[2]: Leaving directory `/building/kbd-2.0.0/tests'
make  check-TESTS
make[2]: Entering directory `/building/kbd-2.0.0/tests'
Running suite(s): libkeymap
100%: Checks: 3, Failures: 0, Errors: 0
PASS: libkeymap-init
Running suite(s): libkeymap
100%: Checks: 2, Failures: 0, Errors: 0
PASS: libkeymap-kmap
Running suite(s): libkeymap
100%: Checks: 5, Failures: 0, Errors: 0
PASS: libkeymap-keys
Running suite(s): libkeymap
100%: Checks: 7, Failures: 0, Errors: 0
PASS: libkeymap-parse
Running suite(s): libkeymap
100%: Checks: 2, Failures: 0, Errors: 0
PASS: libkeymap-charset
PASS: dumpkeys-fulltable
failed
FAIL: dumpkeys-mktable
failed
FAIL: dumpkeys-bkeymap
=========================================
2 of 8 tests failed
Please report to gladkov.alexey@gmail.com
=========================================
make[2]: *** [check-TESTS] Error 1
make[2]: Leaving directory `/building/kbd-2.0.0/tests'
make[1]: *** [check-am] Error 2
make[1]: Leaving directory `/building/kbd-2.0.0/tests'
make: *** [check-recursive] Error 1

 This is with ./configure --prefix=/usr --disable-vlock
--enable-libkeymap (although I'm not quite sure what benefit
libkeymap actually provides).

 The temp.* files were still there, and cat'ing them has trashed
that tty on this old system, so .bz2 versions are attached.

 Still building the system (based on recent LFS), the machine is
horrendously slow so it's going to be a while before it (hopefully)
gets to a usable state.

ĸen
-- 
das eine Mal als Tragödie, dieses Mal als Farce

[-- Attachment #2: temp.lCPDUbQQc.bz2 --]
[-- Type: application/octet-stream, Size: 2449 bytes --]

[-- Attachment #3: temp.w8HGLWAwv.bz2 --]
[-- Type: application/octet-stream, Size: 1050 bytes --]

Re: [kbd] Test failures in kbd-2.0.0 on ppc

[Ken Moffat]

On Fri, Sep 20, 2013 at 01:27:12AM +0100, Ken Moffat wrote:
> =========================================
> 2 of 8 tests failed
> Please report to gladkov.alexey@gmail.com
> =========================================

 The situation is a bit more serious than I thought.  Mac keyboards
were generally dire and weird, plus I use a KVM switch to share one
keyboard and monitor between my desktop machines, so I've got a
regular x86 British keyboard.  This has always worked ok (apart from
the 101st key being not recognized - that's a kernel problem), and I
normally use my own keymap [ uk-utf ] to add a lot of things and to
make the compose settings for accents closer to the Xorg dead keys.

 With kbd-2.0 (whether compiled with or without the separate lib),
loadkeys returns a status of 0 but doesn't change the keymap.  Same
with the shipped uk map.  I've now dropped back to 1.15.5 here and
my own keymap is working fine.

 For anyone else who wants to test this, and who has a ppc with a
regular US keyboard, "loadkeys uk" ought to load the British map
(double quotes on shifted 2, at-sign on shifted comma, pound
(sterling) on shifted 3 instead of has-sign.  And if working,
"loadkeys us" ought to restore it.

 I repeat that this is with everything else up-to-date, maybe it
still works with old toolchains, I don't know.

ĸen
-- 
das eine Mal als Tragödie, dieses Mal als Farce

Re: [kbd] Test failures in kbd-2.0.0 on ppc

[Alexey Gladkov]

23.09.2013 04:31, Ken Moffat wrote:
> On Fri, Sep 20, 2013 at 01:27:12AM +0100, Ken Moffat wrote:
>> =========================================
>> 2 of 8 tests failed
>> Please report to gladkov.alexey@gmail.com
>> =========================================
> 
>  The situation is a bit more serious than I thought.  Mac keyboards
> were generally dire and weird, plus I use a KVM switch to share one
> keyboard and monitor between my desktop machines, so I've got a
> regular x86 British keyboard.  This has always worked ok (apart from
> the 101st key being not recognized - that's a kernel problem), and I
> normally use my own keymap [ uk-utf ] to add a lot of things and to
> make the compose settings for accents closer to the Xorg dead keys.
> 
>  With kbd-2.0 (whether compiled with or without the separate lib),
> loadkeys returns a status of 0 but doesn't change the keymap.  Same
> with the shipped uk map.  I've now dropped back to 1.15.5 here and
> my own keymap is working fine.
> 
>  For anyone else who wants to test this, and who has a ppc with a
> regular US keyboard, "loadkeys uk" ought to load the British map
> (double quotes on shifted 2, at-sign on shifted comma, pound
> (sterling) on shifted 3 instead of has-sign.  And if working,
> "loadkeys us" ought to restore it.
> 
>  I repeat that this is with everything else up-to-date, maybe it
> still works with old toolchains, I don't know.
> 

I've another bug report about the regression in the keymap parser:

https://bugs.archlinux.org/task/36689

-- 
Rgrds, legion

Re: [kbd] Test failures in kbd-2.0.0 on ppc

[Alexey Gladkov]

23.09.2013 04:31, Ken Moffat wrote:
> On Fri, Sep 20, 2013 at 01:27:12AM +0100, Ken Moffat wrote:
>> =========================================
>> 2 of 8 tests failed
>> Please report to gladkov.alexey@gmail.com
>> =========================================
> 
>  The situation is a bit more serious than I thought.  Mac keyboards
> were generally dire and weird, plus I use a KVM switch to share one
> keyboard and monitor between my desktop machines, so I've got a
> regular x86 British keyboard.  This has always worked ok (apart from
> the 101st key being not recognized - that's a kernel problem), and I
> normally use my own keymap [ uk-utf ] to add a lot of things and to
> make the compose settings for accents closer to the Xorg dead keys.

Can you send me your keymaps ?

-- 
Rgrds, legion

Re: [kbd] Test failures in kbd-2.0.0 on ppc

[Alexey Gladkov]

20.09.2013 04:27, Ken Moffat wrote:
> FAIL: dumpkeys-mktable
> FAIL: dumpkeys-bkeymap

Can you check this again on top of master branch ?

-- 
Rgrds, legion