Received-SPF: pass (google.com: domain of linux-ext4-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Sat, 15 May 2021 10:22:39 +0200
From:   Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
To:     David Woodhouse <dwmw2@infradead.org>
Cc:     Linux Doc Mailing List <linux-doc@vger.kernel.org>,
        linux-kernel@vger.kernel.org, Jonathan Corbet <corbet@lwn.net>,
        Mali DP Maintainers <malidp@foss.arm.com>,
        alsa-devel@alsa-project.org, coresight@lists.linaro.org,
        intel-gfx@lists.freedesktop.org, intel-wired-lan@lists.osuosl.org,
        keyrings@vger.kernel.org, kvm@vger.kernel.org,
        linux-acpi@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
        linux-edac@vger.kernel.org, linux-ext4@vger.kernel.org,
        linux-f2fs-devel@lists.sourceforge.net,
        linux-hwmon@vger.kernel.org, linux-iio@vger.kernel.org,
        linux-input@vger.kernel.org, linux-integrity@vger.kernel.org,
        linux-media@vger.kernel.org, linux-pci@vger.kernel.org,
        linux-pm@vger.kernel.org, linux-rdma@vger.kernel.org,
        linux-sgx@vger.kernel.org, linux-usb@vger.kernel.org,
        mjpeg-users@lists.sourceforge.net, netdev@vger.kernel.org,
        rcu@vger.kernel.org
Subject: Re: [PATCH v2 00/40] Use ASCII subset instead of UTF-8 alternate
 symbols
Message-ID: <20210515102239.2ffd0451@coco.lan>
In-Reply-To: <61c286b7afd6c4acf71418feee4eecca2e6c80c8.camel@infradead.org>
References: <cover.1620823573.git.mchehab+huawei@kernel.org>
        <d2fed242fbe200706b8d23a53512f0311d900297.camel@infradead.org>
        <20210514102118.1b71bec3@coco.lan>
        <61c286b7afd6c4acf71418feee4eecca2e6c80c8.camel@infradead.org>
Followup-To: dri-devel@lists.freedesktop.org
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Em Fri, 14 May 2021 10:06:01 +0100
David Woodhouse <dwmw2@infradead.org> escreveu:

> On Fri, 2021-05-14 at 10:21 +0200, Mauro Carvalho Chehab wrote:
> > Em Wed, 12 May 2021 18:07:04 +0100
> > David Woodhouse <dwmw2@infradead.org> escreveu:
> >  =20
> > > On Wed, 2021-05-12 at 14:50 +0200, Mauro Carvalho Chehab wrote: =20
> > > > Such conversion tools - plus some text editor like LibreOffice  or =
similar  - have
> > > > a set of rules that turns some typed ASCII characters into UTF-8 al=
ternatives,
> > > > for instance converting commas into curly commas and adding non-bre=
akable
> > > > spaces. All of those are meant to produce better results when the t=
ext is
> > > > displayed in HTML or PDF formats.   =20
> > >=20
> > > And don't we render our documentation into HTML or PDF formats?  =20
> >=20
> > Yes.
> >  =20
> > > Are
> > > some of those non-breaking spaces not actually *useful* for their
> > > intended purpose? =20
> >=20
> > No.
> >=20
> > The thing is: non-breaking space can cause a lot of problems.
> >=20
> > We even had to disable Sphinx usage of non-breaking space for
> > PDF outputs, as this was causing bad LaTeX/PDF outputs.
> >=20
> > See, commit: 3b4c963243b1 ("docs: conf.py: adjust the LaTeX document ou=
tput")
> >=20
> > The afore mentioned patch disables Sphinx default behavior of
> > using NON-BREAKABLE SPACE on literal blocks and strings, using this
> > special setting: "parsedliteralwraps=3Dtrue".
> >=20
> > When NON-BREAKABLE SPACE were used on PDF outputs, several parts of=20
> > the media uAPI docs were violating the document margins by far,
> > causing texts to be truncated.
> >=20
> > So, please **don't add NON-BREAKABLE SPACE**, unless you test
> > (and keep testing it from time to time) if outputs on all
> > formats are properly supporting it on different Sphinx versions. =20
>=20
> And there you have a specific change with a specific fix. Nothing to do
> with whether NON-BREAKABLE SPACE is =E2=88=89 ASCII, and *certainly* noth=
ing to
> do with the fact that, like *every* character in every kernel file
> except the *binary* files, it's representable in UTF-8.
>=20
> By all means fix the specific characters which are typographically
> wrong or which, like NON-BREAKABLE SPACE, cause problems for rendering
> the documentation.
>=20
>=20
> > Also, most of those came from conversion tools, together with other
> > eccentricities, like the usage of U+FEFF (BOM) character at the
> > start of some documents. The remaining ones seem to came from=20
> > cut-and-paste. =20
>=20
> ... or which are just entirely redundant and gratuitous, like a BOM in
> an environment where all files are UTF-8 and never 16-bit encodings
> anyway.

Agreed.

>=20
> > > > While it is perfectly fine to use UTF-8 characters in Linux, and sp=
ecially at
> > > > the documentation,  it is better to  stick to the ASCII subset  on =
such
> > > > particular case,  due to a couple of reasons:
> > > >=20
> > > > 1. it makes life easier for tools like grep;   =20
> > >=20
> > > Barely, as noted, because of things like line feeds. =20
> >=20
> > You can use grep with "-z" to seek for multi-line strings(*), Like:
> >=20
> > 	$ grep -Pzl 'grace period started,\s*then' $(find Documentation/ -type=
 f)
> > 	Documentation/RCU/Design/Data-Structures/Data-Structures.rst =20
>=20
> Yeah, right. That works if you don't just use the text that you'll have
> seen in the HTML/PDF "grace period started, then", and if you instead
> craft a *regex* for it, replacing the spaces with '\s*'. Or is that
> [[:space:]]* if you don't want to use the experimental Perl regex
> feature?
>=20
>  $ grep -zlr 'grace[[:space:]]\+period[[:space:]]\+started,[[:space:]]\+t=
hen' Documentation/RCU
> Documentation/RCU/Design/Data-Structures/Data-Structures.rst
>=20
> And without '-l' it'll obviously just give you the whole file. No '-A5
> -B5' to see the surroundings... it's hardly a useful thing, is it?
>=20
> > (*) Unfortunately, while "git grep" also has a "-z" flag, it
> >     seems that this is (currently?) broken with regards of handling mul=
tilines:
> >=20
> > 	$ git grep -Pzl 'grace period started,\s*then'
> > 	$ =20
>=20
> Even better. So no, multiline grep isn't really a commonly usable
> feature at all.
>=20
> This is why we prefer to put user-visible strings on one line in C
> source code, even if it takes the lines over 80 characters =E2=80=94 to a=
llow
> for grep to find them.

Makes sense, but in case of documentation, this is a little more
complex than that.=20

Btw, the theme used when building html by default[1] has a search
box (written in Javascript) that could be able to find multi-line
patterns, working somewhat similar to "git grep foo -a bar".

[1] https://github.com/readthedocs/sphinx_rtd_theme

> > [1] If I have a table with UTF-8 codes handy, I could type the UTF-8=20
> >     number manually... However, it seems that this is currently broken=
=20
> >     at least on Fedora 33 (with Mate Desktop and US intl keyboard with=
=20
> >     dead keys).
> >=20
> >     Here, <CTRL><SHIFT>U is not working. No idea why. I haven't=20
> >     test it for *years*, as I din't see any reason why I would
> >     need to type UTF-8 characters by numbers until we started
> >     this thread. =20
>=20
> Please provide the bug number for this; I'd like to track it.

Just opened a BZ and added you as c/c.

> > Now, I'm not arguing that you can't use whatever UTF-8 symbol you
> > want on your docs. I'm just saying that, now that the conversion=20
> > is over and a lot of documents ended getting some UTF-8 characters
> > by accident, it is time for a cleanup. =20
>=20
> All text documents are *full* of UTF-8 characters. If there is a file
> in the source code which has *any* non-UTF8, we call that a 'binary
> file'.
>=20
> Again, if you want to make specific fixes like removing non-breaking
> spaces and byte order marks, with specific reasons, then those make
> sense. But it's got very little to do with UTF-8 and how easy it is to
> type them. And the excuse you've put in the commit comment for your
> patches is utterly bogus.

Let's take one step back, in order to return to the intents of this
UTF-8, as the discussions here are not centered into the patches, but
instead, on what to do and why.

-

This discussion started originally at linux-doc ML.

While discussing about an issue when machine's locale was not set
to UTF-8 on a build VM, we discovered that some converted docs ended
with BOM characters. Those specific changes were introduced by some
of my convert patches, probably converted via pandoc.

So, I went ahead in order to check what other possible weird things
were introduced by the conversion, where several scripts and tools
were used on files that had already a different markup.

I actually checked the current UTF-8 issues, and asked people at
linux-doc to comment what of those are valid usecases, and what
should be replaced by plain ASCII.

Basically, this is the current situation (at docs/docs-next), for the
ReST files under Documentation/, excluding translations is:

1. Spaces and BOM

	- U+00a0 ('=C2=A0'): NO-BREAK SPACE
	- U+feff ('=EF=BB=BF'): ZERO WIDTH NO-BREAK SPACE (BOM)

Based on the discussions there and on this thread, those should be
dropped, as BOM is useless and NO-BREAK SPACE can cause problems
at the html/pdf output;

2. Symbols

	- U+00a9 ('=C2=A9'): COPYRIGHT SIGN
	- U+00ac ('=C2=AC'): NOT SIGN
	- U+00ae ('=C2=AE'): REGISTERED SIGN
	- U+00b0 ('=C2=B0'): DEGREE SIGN
	- U+00b1 ('=C2=B1'): PLUS-MINUS SIGN
	- U+00b2 ('=C2=B2'): SUPERSCRIPT TWO
	- U+00b5 ('=C2=B5'): MICRO SIGN
	- U+03bc ('=CE=BC'): GREEK SMALL LETTER MU
	- U+00b7 ('=C2=B7'): MIDDLE DOT
	- U+00bd ('=C2=BD'): VULGAR FRACTION ONE HALF
	- U+2122 ('=E2=84=A2'): TRADE MARK SIGN
	- U+2264 ('=E2=89=A4'): LESS-THAN OR EQUAL TO
	- U+2265 ('=E2=89=A5'): GREATER-THAN OR EQUAL TO
	- U+2b0d ('=E2=AC=8D'): UP DOWN BLACK ARROW

Those seem OK on my eyes.

On a side note, both MICRO SIGN and GREEK SMALL LETTER MU are
used several docs to represent microseconds, micro-volts and
micro-amp=C3=A8res. If we write an orientation document, it probably
makes sense to recommend using MICRO SIGN on such cases.

3. Latin

	- U+00c7 ('=C3=87'): LATIN CAPITAL LETTER C WITH CEDILLA
	- U+00df ('=C3=9F'): LATIN SMALL LETTER SHARP S
	- U+00e1 ('=C3=A1'): LATIN SMALL LETTER A WITH ACUTE
	- U+00e4 ('=C3=A4'): LATIN SMALL LETTER A WITH DIAERESIS
	- U+00e6 ('=C3=A6'): LATIN SMALL LETTER AE
	- U+00e7 ('=C3=A7'): LATIN SMALL LETTER C WITH CEDILLA
	- U+00e9 ('=C3=A9'): LATIN SMALL LETTER E WITH ACUTE
	- U+00ea ('=C3=AA'): LATIN SMALL LETTER E WITH CIRCUMFLEX
	- U+00eb ('=C3=AB'): LATIN SMALL LETTER E WITH DIAERESIS
	- U+00f3 ('=C3=B3'): LATIN SMALL LETTER O WITH ACUTE
	- U+00f4 ('=C3=B4'): LATIN SMALL LETTER O WITH CIRCUMFLEX
	- U+00f6 ('=C3=B6'): LATIN SMALL LETTER O WITH DIAERESIS
	- U+00f8 ('=C3=B8'): LATIN SMALL LETTER O WITH STROKE
	- U+00fa ('=C3=BA'): LATIN SMALL LETTER U WITH ACUTE
	- U+00fc ('=C3=BC'): LATIN SMALL LETTER U WITH DIAERESIS
	- U+00fd ('=C3=BD'): LATIN SMALL LETTER Y WITH ACUTE
	- U+011f ('=C4=9F'): LATIN SMALL LETTER G WITH BREVE
	- U+0142 ('=C5=82'): LATIN SMALL LETTER L WITH STROKE

Those should be kept as well, as they're used for non-English names.

4. arrows and box drawing symbols:
	- U+2191 ('=E2=86=91'): UPWARDS ARROW
	- U+2192 ('=E2=86=92'): RIGHTWARDS ARROW
	- U+2193 ('=E2=86=93'): DOWNWARDS ARROW

	- U+2500 ('=E2=94=80'): BOX DRAWINGS LIGHT HORIZONTAL
	- U+2502 ('=E2=94=82'): BOX DRAWINGS LIGHT VERTICAL
	- U+2514 ('=E2=94=94'): BOX DRAWINGS LIGHT UP AND RIGHT
	- U+251c ('=E2=94=9C'): BOX DRAWINGS LIGHT VERTICAL AND RIGHT

Also should be kept.

In summary, based on the discussions we have so far, I suspect that
there's not much to be discussed for the above cases.

So, I'll post a v3 of this series, changing only:

	- U+00a0 ('=C2=A0'): NO-BREAK SPACE
	- U+feff ('=EF=BB=BF'): ZERO WIDTH NO-BREAK SPACE (BOM)

---

Now, this specific patch series address also this extra case:

5. curly commas:

	- U+2018 ('=E2=80=98'): LEFT SINGLE QUOTATION MARK
	- U+2019 ('=E2=80=99'): RIGHT SINGLE QUOTATION MARK
	- U+201c ('=E2=80=9C'): LEFT DOUBLE QUOTATION MARK
	- U+201d ('=E2=80=9D'): RIGHT DOUBLE QUOTATION MARK

IMO, those should be replaced by ASCII commas: ' and ".

The rationale is simple:=20

- most were introduced during the conversion from Docbook,
  markdown and LaTex;
- they don't add any extra value, as using "foo" of =E2=80=9Cfoo=E2=80=9D m=
eans
  the same thing;
- Sphinx already use "fancy" commas at the output.=20

I guess I will put this on a separate series, as this is not a bug
fix, but just a cleanup from the conversion work.

I'll re-post those cleanups on a separate series, for patch per patch
review.

---

The remaining cases are future work, outside the scope of this v2:

6. Hyphen/Dashes and ellipsis

	- U+2212 ('=E2=88=92'): MINUS SIGN
	- U+00ad ('=C2=AD'): SOFT HYPHEN
	- U+2010 ('=E2=80=90'): HYPHEN

	    Those three are used on places where a normal ASCII hyphen/minus
	    should be used instead. There are even a couple of C files which
	    use them instead of '-' on comments.

	    IMO are fixes/cleanups from conversions and bad cut-and-paste.

	- U+2013 ('=E2=80=93'): EN DASH
	- U+2014 ('=E2=80=94'): EM DASH
	- U+2026 ('=E2=80=A6'): HORIZONTAL ELLIPSIS

	    Those are auto-replaced by Sphinx from "--", "---" and "...",
	    respectively.

	    I guess those are a matter of personal preference about
	    weather using ASCII or UTF-8.

            My personal preference (and Ted seems to have a similar
	    opinion) is to let Sphinx do the conversion.

	    For those, I intend to post a separate series, to be
	    reviewed patch per patch, as this is really a matter
	    of personal taste. Hardly we'll reach a consensus here.

7. math symbols:

	- U+00d7 ('=C3=97'): MULTIPLICATION SIGN

	   This one is used mostly do describe video resolutions, but this is
	   on a smaller changeset than the ones that use "x" letter.

	- U+2217 ('=E2=88=97'): ASTERISK OPERATOR

	   This is used only here:
		Documentation/filesystems/ext4/blockgroup.rst:filesystem size to 2^21 =E2=
=88=97 2^27 =3D 2^48bytes or 256TiB.

	   Probably added by some conversion tool. IMO, this one should
	   also be replaced by an ASCII asterisk.

I guess I'll post a patch for the ASTERISK OPERATOR.
Thanks,
Mauro