[OE-core] [PATCH V3 1/5] systemd: add 'less' to its RRECOMMENDS

[Bernhard Reutner-Fischer]

On 23 January 2014 09:23:18 ChenQi <Qi.Chen at windriver.com> wrote:

> On 01/22/2014 06:18 PM, Bernhard Reutner-Fischer wrote:
> > On 22 January 2014 11:08, ChenQi <Qi.Chen at windriver.com> wrote:
> >> On 01/22/2014 06:00 PM, Bernhard Reutner-Fischer wrote:
> >>> On 20 January 2014 09:35, ChenQi <Qi.Chen at windriver.com> wrote:
> >>>> On 01/20/2014 04:06 PM, Bernhard Reutner-Fischer wrote:
> >>>>> On 20 January 2014 03:23:02 Chen Qi <Qi.Chen at windriver.com> wrote:
> >>>>>
> >>>>>> The less command provided by busybox cannot interpret the control
> >>>>>> character, resulting in obscure output of the journalctl command
> >>>>>> in a systemd based image.
> >>>>>
> >>>>> Can you send me the input containing this control character, please?
> >>>>>
> >>>>> Perhaps we can do something sensible about it..
> >>>>> TIA,
> >>>>
> >>>> For more details, please refer to
> >>>> https://bugzilla.yoctoproject.org/show_bug.cgi?id=5723.
> >>>> You can also add you comments there.
> >>> Can you please try:
> >>> http://marc.info/?l=busybox&m=139038437907397&w=2
> >>> and let me know if this fixes your issue?
> >>>
> >>> thanks,
> >>>
> >>>
> >> Hi,
> >>
> >> I cannot open the link.
> >> Could you please send me your patches as attachment? I'll try them out
> >> tomorrow and let you know the results.
> > http://marc.info/?l=busybox&m=139038437907397&q=raw
> > or, just for you:
> > http://busybox.net/~aldot/tmp/0001-less-Add-R.patch
> >
> > thanks,
>
> Hi Bernhard,
>
> I just tried this patch out.
>
> 1) Test 1
> Command: journalctl
> Result: control characters not interpreted
> Log:
> Jan 23 07:35:49 qemux86 kernel: [[1;39mACPI: RSDP 000f1610 00014 (v00 BOCHS )[[
> 0m
> Jan 23 07:35:49 qemux86 kernel: [[1;39mACPI: RSDT 0fffe450 00034 (v01 BOCHS  BX
> PCRSDT 00000001 BXPC 00000001)[[0m
> Jan 23 07:35:49 qemux86 kernel: [[1;39mACPI: FACP 0fffff80 00074 (v01 BOCHS  BX
> PCFACP 00000001 BXPC 00000001)[[0m
> Jan 23 07:35:49 qemux86 kernel: [[1;39mACPI: DSDT 0fffe490 01137 (v01 BXPC
> BXDSDT 00000001 INTL 20120913)[[0m
> Jan 23 07:35:49 qemux86 kernel: [[1;39mACPI: FACS 0fffff40 00040[[0m
> Jan 23 07:35:49 qemux86 kernel: [[1;39mACPI: SSDT 0ffff700 00838 (v01 BOCHS  BX
> PCSSDT 00000001 BXPC 00000001)[[0m
> Jan 23 07:35:49 qemux86 kernel: [[1;39mACPI: APIC 0ffff610 00078 (v01 BOCHS  BX
> PCAPIC 00000001 BXPC 00000001)[[0m
> Jan 23 07:35:49 qemux86 kernel: [[1;39mACPI: HPET 0ffff5d0 00038 (v01 BOCHS  BX
> PCHPET 00000001 BXPC 00000001)[[0m
> Jan 23 07:35:49 qemux86 kernel: ACPI: Local APIC address 0xfee00000
> Jan 23 07:35:49 qemux86 kernel: [[1;39m0MB HIGHMEM available.[[0m
> Jan 23 07:35:49 qemux86 kernel: [[1;39m255MB LOWMEM available.[[0m
>
> 2) Test 2
> Command: export SYSTEMD_PAGER="less -R"; journalctl
> Result: control characters interpreted
>
> I know the behavior of `less' command from busybox is different from that 
> of the 'less' package. And we can't ask too much.
> But do you think it's possible to make 'journalctl' work correctly without 
> setting SYSTEMD_PAGER to "less -R"? I'm wondering why busybox's 'less' 
> doesn't use the '-R' behaviour as its default. If so, we would not need 
> this '-R' option at all.

Apparently Denys does not like the idea of having -R in there at all but 
even if it gets in, defaulting to -R is not acceptable I think. We could, 
however, add parsing the LESS env var as the big less does. I guess that 
would add another 30-40 bytes to the 300-ish for the -R itself, though.
Plain -r was much smaller, but I dropped it due to paranoia, fwiw.

To some extent I do agree with Denys that journalctl should only print ANSI 
colors if it's stdout isatty.

Thanks,

Sent with AquaMail for Android
http://www.aqua-mail.com