On 6/10/20 11:47 AM, Jerin Jacob wrote:
> On Wed, Jun 10, 2020 at 1:57 PM Andrew Rybchenko
> <arybche...@solarflare.com> wrote:
>>
>> On 6/10/20 8:22 AM, Jerin Jacob wrote:
>>> On Tue, Jun 9, 2020 at 7:27 PM Bruce Richardson
>>> <bruce.richard...@intel.com> wrote:
>>>>
>>>> On Tue, Jun 09, 2020 at 04:40:28PM +0300, Andrew Rybchenko wrote:
>>>>> On 6/9/20 1:00 PM, Ananyev, Konstantin wrote:
>>>>>>
>>>>>>>
>>>>>>> On Mon, Jun 08, 2020 at 12:17:23PM -0700, Stephen Hemminger wrote:
>>>>>>>> On Mon, 8 Jun 2020 17:46:40 +0100 Bruce Richardson
>>>>>>>> <bruce.richard...@intel.com> wrote:
>>>>>>>>
>>>>>>>>> Rather than continuing to recommend an 80-char line limit, let's
>>>>>>>>> take a hint from the Linux kernel[1] and aim for an 100-char
>>>>>>>>> recommended limit instead.
>>>>>>>>>
>>>>>>>>> [1]
>>>>>>>>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=bdc48fa11e46f867ea4d75fa59ee87a7f48be144
>>>>>>>>>
>>>>>>>>> Signed-off-by: Bruce Richardson <bruce.richard...@intel.com> ---
>>>>>>>>> devtools/checkpatches.sh | 2 +-
>>>>>>>>> doc/guides/contributing/coding_style.rst | 2 +- 2 files changed, 2
>>>>>>>>> insertions(+), 2 deletions(-)
>>>>>>>>>
>>>>>>>>> diff --git a/devtools/checkpatches.sh b/devtools/checkpatches.sh
>>>>>>>>> index 158087f1c..4970ed830 100755 --- a/devtools/checkpatches.sh +++
>>>>>>>>> b/devtools/checkpatches.sh @@ -15,7 +15,7 @@
>>>>>>>>> VALIDATE_NEW_API=$(dirname $(readlink -f $0))/check-symbol-change.sh
>>>>>>>>> # Codespell can also be enabled by setting DPDK_CHECKPATCH_CODESPELL
>>>>>>>>> to a valid path # to a dictionary.txt file if dictionary.txt is not
>>>>>>>>> in the default location.
>>>>>>>>> codespell=${DPDK_CHECKPATCH_CODESPELL:-enable}
>>>>>>>>> -length=${DPDK_CHECKPATCH_LINE_LENGTH:-80}
>>>>>>>>> +length=${DPDK_CHECKPATCH_LINE_LENGTH:-100}
>>>>>>>>>
>>>>>>>>> # override default Linux options options="--no-tree" diff --git
>>>>>>>>> a/doc/guides/contributing/coding_style.rst
>>>>>>>>> b/doc/guides/contributing/coding_style.rst index
>>>>>>>>> 4efde93f6..1db3a7bbe 100644 ---
>>>>>>>>> a/doc/guides/contributing/coding_style.rst +++
>>>>>>>>> b/doc/guides/contributing/coding_style.rst @@ -21,7 +21,7 @@ The
>>>>>>>>> rules and guidelines given in this document cannot cover every
>>>>>>>>> situation, so * In the case of creating new files, the style should
>>>>>>>>> be consistent within each file in a given directory or module. *
>>>>>>>>> The primary reason for coding standards is to increase code
>>>>>>>>> readability and comprehensibility, therefore always use whatever
>>>>>>>>> option
>>>>>>> will make the code easiest to read.
>>>>>>>>>
>>>>>>>>> -Line length is recommended to be not more than 80 characters,
>>>>>>>>> including comments. +Line length is recommended to be not more than
>>>>>>>>> 100 characters, including comments. [Tab stop size should be
>>>>>>>>> assumed to be 8-characters wide].
>>>>>>>>>
>>>>>>>>> .. note::
>>>>>>>>
>>>>>>>> I would even support going to 120 characters.
>>>>>>>>
>>>>>>>> Acked-by: Stephen Hemminger <step...@networkplumber.org>
>>>>>>>
>>>>>>> I think 100 is enough.
>>>>>>>
>>>>>>> In my case, I have a 1080p 24" monitor, and with two terminals
>>>>>>> side-by-side 100 characters just fits inside each vim window. Going to
>>>>>>> 120 would be fine for single terminal at a time, but I would find
>>>>>>> awkward for e.g. side-by-side diff comparison in meld etc.
>>>>>>
>>>>>> My preference would be to keep things as it is - 80 chars per line.
>>>>>> Having multiple different formatting styles in one source file looks
>>>>>> really awkward and make it hard to follow.
>>>>>
>>>>> +1
>>>>>
>>>> I wouldn't personally consider increasing the max line length as a style
>>>> change, but even if you consider it such I'd worry about rejecting style
>>>> changes on the basis that it may be different to what is there before. That
>>>> logic means that we can never, ever change any element of DPDK coding
>>>> style.
>>>>
>>>> I can see the issue with changes that require us to rework the style of
>>>> code in order to comply with the new style, but changing the max length
>>>> from 80 to 100 does not make 80-char lines incorrect and needing changes.
>>>
>>> Another point is: Other projects derived from the Linux kernel coding
>>> standard also
>>> getting migrated to the new coding standard. This change would be useful
>>> for:
>>> a) People works on multiple Linux coding standard derived projects
>>
>> Valid point, but not really strong.
>> I think that .editorconfig solves it.
>
> Yes, For adding the code. I meaning, Viewing the code there will be a
> disparity.
I hope you're not suggesting to reformat all existing code.
Otherwise the disparity will be there for a long-long time
anyway.
>>
>>> b) Some of the code such as 'base' and 'common' code for HW drivers
>>> are shared between multiple projects.
>>> Such code needs adjustment/change when pulling to the DPDK code base
>>> it it still follows 80 chars per line.
>>
>> Base and common code are not required to follow DPDK coding
>> style even now.
>
> I see, I dont think it is expressed in devtools/checkpatches.sh. I.e
> CI tools still flag as checkpatch issues.
Yes, it is an area which should be improved.
> Coming to original concern:(code disparity with existing code)
> Another option is, It is possible to change existing code to 100 lines
> with clang-format in an automatic fashion. But it will have a lot of
> changes.
> "C_Cpp.clang_format_style": "{ BasedOnStyle: LLVM,
> IndentWidth: 8, TabWidth: 8, UseTab: Always,
> AllowShortIfStatementsOnASingleLine: false, IndentCaseLabels: false,
> ColumnLimit: 100, AllowShortFunctionsOnASingleLine: false,
> AlwaysBreakAfterReturnType: AllDefinitions, ColumnLimit: 100,
> ConstructorInitializerAllOnOneLineOrOnePerLine: true,
> ConstructorInitializerIndentWidth: 8, ContinuationIndentWidth: 8,
> BreakBeforeBraces: Linux, AllowShortBlocksOnASingleLine: false,
> AlignConsecutiveAssignments: false, AlignEscapedNewlines: Right,
> AlignConsecutiveMacros : true, MaxEmptyLinesToKeep : 1,
> Cpp11BracedListStyle : true, AlignTrailingComments : true,
> ForEachMacros: ['TAILQ_FOREACH_SAFE', STAILQ_FOREACH',
> 'rte_graph_foreach_node', 'TAILQ_FOREACH', 'RTE_ETH_FOREACH_DEV']}",
No, no, no. Please, no. It will complicate backporting a lot,
it will break (over-complicate) git blame.
(I hope it was suggested just to be sure that it will not be
done).
