[sane-devel] Coding conventions (was Re: Possible to rename network scanner?)

[Olaf Meeuwissen]

Hi Aaron, list,

Aaron Muir Hamilton writes:

> Some of these misleading indentation errors also look like genuine bugs.
> Here's one gem from backends/genesys_gl847.c which turned into a
> misleading indent when I ran it through clang-format:
>
> 660    while (val8 & REG41_FEBUSY);
> 661      {
> 662        usleep (10000);
> 663        status = sanei_genesys_get_status (dev, &val8);
> 664      };
>
> Pay close attention to the semicolons.

Things like this is why I prefer to address the compiler warnings one at
a time and ask for feedback where not sure what to do.

> As an aside, this is probably why the 1TBS is superiour to GNU-style.

Not familiar with 1TBS, so I looked it up[1]

 [1] https://en.wikipedia.org/wiki/Indent_style#Variant:_1TBS_.28OTBS.29

OK, got it.  Fine with me, though most of my own C/C++ code uses GNU.

We all have our preferences and I am not going to step on people's toes
trying to enforce a "one true style" for all of sane-backends.  Backend
maintainers can use whatever works for them in those files that make up
their backends.  All I would like to ask is to use whatever they prefer
consistently throughout a file.

If the file includes hints for common editors that's great but that is
typically limited to emacs and vim.  Adding an EditorConfig[2] entry
would be better as that ought to work for even more editors and IDEs.

 [2] http://editorconfig.org/

What I would like to apply throughout though is

 - UTF-8 encoding everywhere
 - no trailing whitespace at the end of line
 - no trailing empty lines are the end of file
 - new line at end of file

That shouldn't be too controversial but I'll stay away from the coding
style (and tab versus spaces) wars ;-)

> Aaron Muir Hamilton <aaron at correspondwith.me> writes:
>
>> Olaf Meeuwissen <paddy-hack at member.fsf.org> writes:
>>> # Never even mind Fedora 26 and Debian 9 on the horizon which'll trigger
>>> # a small "tsunami" of compiler warning fixes and autotools updates ...
>>
>> It's ugly work, but maybe it's time we normalize the formatting in some
>> of the files (especially ones which are abandoned-ish and don't have
>> open patches). A lot of the compiler warnings I see (I run Arch Linux,
>> so I'm on bleeding-edge-ish GCC most of the time) are things like
>> misleading indentation warnings. We don't need to invest in any
>> infrastructure for formatting IMHO, we could just do a once-over with
>> clang-format on some of the files.

As mentioned above, we all have our coding style preferences and some of
the developers may be more outspoken than others.  Since I like to keep
as many of them around as possible, I prefer to tread carefully with a
blanket re-indent of (some of the) code files.  If the indenting is more
or less (more rather than less) consistent, I don't see any need to redo
the indenting.

We have a number of builds[3] among which Fedora 25 should give results
close-ish to what you see.  Checkout out one of the recent logs[4].

 [3] https://gitlab.com/sane-project/backends/pipelines
 [4] https://gitlab.com/sane-project/backends/builds/16802918

I try to keep the ci-envs[5] up-to-date with *stable* releases, so
Fedora 26 (and Debian 9?) should be just around the corner.

 [5] https://gitlab.com/sane-project/ci-envs

Hope this helps,
--
Olaf Meeuwissen, LPIC-2            FSF Associate Member since 2004-01-27
 GnuPG key: F84A2DD9/B3C0 2F47 EA19 64F4 9F13  F43E B8A4 A88A F84A 2DD9
 Support Free Software                        https://my.fsf.org/donate
 Join the Free Software Foundation              https://my.fsf.org/join