[PATCH v3 1/3] bsps/xilinx_zynq: Add SPI driver for cadence-spi

[Chris Johns]

On 17/2/21 7:34 am, Gedare Bloom wrote:
> On Tue, Feb 16, 2021 at 10:44 AM <dufault at hda.com> wrote:
>>> On Feb 16, 2021, at 11:33 , Gedare Bloom <gedare at rtems.org> wrote:
>>>> The umlaut is part of the official brand.
>>>> That can be tricky, depending on the corporate policy.
>>>>
>>>>> I find it odd to avoid Unicode and not allow umlauts in 2021.
>>>>>
>>> We do allow unicode with justification. For names, I have no problem
>>> with unicode characters. We do need to confirm when it commits, that
>>> it comes back out the other side properly and doesn't become mangled
>>> by some ASCII conversion.
>>
>> Sebastian pointed out this file in an earlier Unicode discussion:
>>
>> testsuites/fstests/fsdosfsname01/create_files.cs
>>
>> It's full of Unicode.
>>
> And it is well-justified as testing.
> 
> We haven't really made anything official policy, but I think the rule
> is to avoid unicode unless there is a good reason for it.

We do not have a policy and I think we should.

I looked back and found the only discussion was about the sources for documentation:

https://lists.rtems.org/pipermail/devel/2019-February/052742.html

After that there is not a lot.

The doco thread prompted me to check patch 1/3 in this series and saved the
email. The text in it was a large long box of ascii encoding of some form. I
applied the patch using git as test and git is fine with this which is neat and
the unicode characters appeared in the git log as:

+/*
+ * Copyright (C) 2020 Jan Sommer, Deutsches Zentrum f<C3><BC>r Luft- und
Raumfahrt e. V. (DLR)
+ *

Less and vi both has similar things ...

 * Copyright (C) 2020 Jan Sommer, Deutsches Zentrum f\xc3\xbcr Luft- und
Raumfahrt e. V. (DLR)

The issue for terminal level work discussed in the doco source is present but in
terms of formal names I am fine with there being unicode characters. I think
having a formal name written in its proper form is what we should allow. The use
of unicode in comments or code is different and here we need to accept the low
common denominator which looks like me on a text terminal.

Chris