Re: [PATCH] execveat.2: srcfix

["Michael Kerrisk (man-pages)" <mtk.manpages@gmail.com>]

Hi Alex,

On 1/2/21 10:40 PM, Alejandro Colomar (man-pages) wrote:
> Hi Michael,
> 
> I read everything this time ;)
>

:-)

[...]

>>>>> Still I didn't read past that :)
>>>
>>> Later I'll have a look past there :)
>>
>> That would be great!
> 
> adjtimex.2: compact

I decided against. It feels too crowded.

> getpeername.2: 78-col

It *is* with 78-col?

> kexec_load.2: compact?

Done.

> msgop.2: compact

It feels a bit too crowded.

> rt_sigqueueinfo.2: Join + compact

Done

> sgetmask.2: compact

Done

> subpage_prot.2: Join

Done

> vm86.2: compact

Done

> circleq.3, list.3, slist.3, tailq.3, stailq.3: group?

I've taken a shot at that. You may have improvements to
suggest, or even reorderings to suggest (as patches).

> des_crypt.3: group

Done

> drand48_r.3: Join

Done

> error.3: group

Done

> ether_aton.3: group

Done

> exec.3: consistency with commas; execvpe can be joined
Done; done

> fseeko.3: compact
> 
> resolver.3: group?

Done

> rewinddir.3: compact headers

Done.

> rpc: 78-col
> 	CLGET_SERVER_ADDR  struct sockaddr_in  // get server's addre
> 	ss

Fixed.

> rtnetlink.3: group or compact; 78-col

Group. But I don't see the 78-col problem?

> scandir.3:
> 	scandir(): move 'filter' and 'compar' 1 char to the right
> 	  so that parentheses are not aligned and confuse.
> 	  It still fits 1 char to the right but not to for a 78-col.
> 	group?
> 	scandirat(): ffix!!: namelist can be joined

Doine; done; done

> setaliasent.3: move more to the right

Done

> shm_open.3: compact

Done

> sigpause.3: compact

I prefer not. The APIs have the same name. A bit of space emphasizes that
they are different, I think.

> sigset.3: group

Done

> strstr.3: compact #define & #include

Done

> termios.3: group even more?

Done

> timegm.3: compact

Done

> unlocked_stdio.3: Join fread_unlocked(3)? Or not?

I think not.

But I did *add* a few blank lines here.

> xdr.3: wsfix: g/) (/s//)(/
> 	(See if there are any other pages with this
> 	 that I may haven't seen.)

Done.

Plus: error.3, ftw.3, glob.3, pthread_create.3, rpc.3


> man.7: group or compact

group

> rtnetlink.7: 78-col

78-col looks okay already?

> sigevent.7: s/) (/)(/

Done.

> 	If you move the comments a few chars to the right (3<=x<=6),
> 	  you will compact one line

I prefer to leave as is.

> Also, curiously execveat(2), which is the one that started all this,
> didn't look bad :p

True.

> So we'll have to grep for .nf/.fi too after this.

Well, I just fixed most of them. The following perhaps need further
consideration:

man1/iconv.1
man1/localedef.1
man1/time.1
man2/select_tut.2
man3/string.3
man4/sk98lin.4
man4/smartpqi.4
man7/man.7
man7/man-pages.7
man8/iconvconfig.8
man8/ldconfig.8
man8/ld.so.8
man8/zdump.8
man8/zic.8

The last two are imported pages, so should probably be ignored.
Perhaps none of the remainder really matter.

> Things to note for other patches:
> 
> isw*.3: Rewrite into one page similar to isalpha.3?
> 	Does it really need so many pages?

There sure is a lot of repetition across those pages...

> recno.3: Review: no APIs

It's a strange page, but I'm not sure that anything needs fixing.

> string.3: What is the criterion for functions to be there?
> 	Also, there are functions which are already documented
> 	  in their own pages (see strcpy(3))
> 	Some others don't appear there (see memcpy(3)
> 	  eventhough they are in string.h.

See also bstring(3)

bstring(3) and string(3) are ancient pages. I'm not entirely
convinced of their value. I suppose thay are useful in the sense
that you get a list of related functions. It is of course
anomalous that string(3) and brief function descriptions
while bstring(3) does not.

Cheers,

Michael


-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/