On 1/8/21 1:44 PM, Michael Kerrisk (man-pages) wrote:

On 1/8/21 12:49 PM, Alejandro Colomar (man-pages) wrote:

quoted

On 1/8/21 12:29 PM, Michael Kerrisk (man-pages) wrote:

quoted

Hi Alex,

On Thu, 7 Jan 2021 at 18:11, Alejandro Colomar (man-pages)
[off-list ref] wrote:

quoted

Hi Michael,

On 1/6/21 1:51 PM, Michael Kerrisk (man-pages) wrote:

quoted

Hi Alex,

On 1/5/21 10:56 PM, Alejandro Colomar (man-pages) wrote:

quoted

Hi Michael,

While having a look at your latest commits,
I saw that there's a bit of inconsistency in escaping whitespace
(existing previous to the commit):

Sometimes it's [!\ (], and sometimes it's [! (].

Thanks for prodding me about this.

Yes, it's inconsistent. And given the use of .nf/.fi
around the text blocks, escaping the spaces serves no
purpose. So I made a more radical fix; see commit
5c10d2c5e299011e34adb568737acfc8920fc27c

Nice!

After your following commit (422d5327a88fa89394100bafad69b21e50b26399),
I found that there are many such cases.  Just [[grep -rnI '\\ ' man?]]
and you'll see.  Some of them are valid, but a lot of them aren't.

Can you point me at some examples?

A sample:

man3/envz_add.3:61:.RI ( *envz ,\  *envz_len )
<< this should be two lines
man3/xdr.3:184:.IR "long\ *" .
<< unnecessary
man3/scandir.3:277:.IR "const void\ *" .
<< unnecessary and self-inconsistent
man5/proc.5:1350:.RI ( "readelf\ \-l" ).
<< unnecessary
man7/symlink.7:424:.I "rm\ \-r slink directory"
<< unnecessary and self-inconsistent
man7/feature_test_macros.7:492:.IR "cc\ \-std=c99" ).
<< unnecessary
man8/ld.so.8:313:.IR "readelf\ \-n"
<< unnecessary

Maybe I'm missing something?

So, the point here is that suppose a line break falls badly and we end up with

...................... const void
* ...............................

or

......................... readelf
-n ..............................

or
.............................. cc
-std=c99 ........................

or

.............................. 16
MB ..............................

The general problem here is that a small piece of a unit (usually at
the end) is broken onto a new line, making  things a bit more
difficult to read. That sort of thing is ugly, I think. That's
why there is the "\\ ".

Ahhh, true.
Thanks!

Maybe there are a few redundant cases. And maybe there are a few
borderline cases. For example, maybe in envz_add.3, "\\ " is
not strictly necessary (though I'm inclined to keep it).

There may still be a few misplaced "\\ " escapes (I just fixed a few),
but many of these really are needed, I think.

Cheers,

Michael

-- 
Alejandro Colomar
Linux man-pages comaintainer; https://www.kernel.org/doc/man-pages/
http://www.alejandro-colomar.es/