mkerrisk / man-pages Goto Github PK

The 't' flag is text mode versus binary mode 'b'.

In the section Ranges, the last example given states that:

... and "[--0]" matches the three characters '-', '.', '0', since '/' cannot be matched.)

Upon trial,

ls -a | grep [--0] does not return anything containing a ..

I would change it myself, but I have very little experience with GitHub (and maybe too much experience with semantics).

man2/pciconfig_{write,read,iobase}.2

man3/{db,btree,dbopen,hash,mpool,recno}.3

deprecated by libxcrypt
man3/crypt{,_r}.3

sccs is no longer used: ( from the posix man pages )
/man1p/{admin,delta,get,prs,rmdel,sact,sccs,unget,val,what}.1p

The man page for the command "date" needs updated to clarify a "+" is needed before the format arguments.

For example it lists: date [OPTION]... [+FORMAT]

But no example later on that specifies "+" needs included before the listed format options. Moving the "+" outside the brackets and providing an example could further clarify.

I believe this is the correct location but I can't seem to locate the man page for this command to push corrections.

Yesterday I had to timestamp messages received from a TCP socket. I found that there is SO_TIMESTAMP, but it uses struct timeval which AFAIK is obsolete.

Then I found in the web some pages that talk about SO_TIMESTAMPNS, but I couldn't find clear documentation about it, and there is absolutely nothing in the man pages.

So I had to try. I tried SO_TIEMSTAMPNS, and it does work, so I think it should be documented.

EDIT: I've found there is a 4-years old FIXME:

Of the nameserver keyword, the man page states,

If there are multiple servers, the resolver library queries them in the order listed.

and,

(The algorithm used is to try a name server, and if the query times out, try the next, until out of name servers, then repeat trying all the name servers until a maximum number of retries are made.)

These are implementation details of glibc and relying on this behaviour introduces a number of subtle problems that commonly trip people up.

I think glibc's behaviour should be documented in glibc's documentation, and the text of the man page should be change to something like:

If there are multiple servers, they must all provide the same "view" of the DNS.

and

(Behaviour varies between different resolvers regarding how multiple servers are interpreted.)

Or perhaps the parenthetical remark should be removed altogether.

It is possible for mkdir(2) to return EINVAL in certain circumstances. However, the man page mkdir(2) doesn't mention that.

I am hitting it on overlayfs. Other instances reported:

http://stackoverflow.com/questions/10015377/mkdir-pathname-with-symbol
http://lists.opensuse.org/opensuse-bugs/2009-05/msg01771.html
(and many others)

So, it needs to be documented in man page for mkdir(2).

Why is there no entry for the X windowing system?

Not really an issue, just a question. The stdarg functions man pages point to the man page for stdarg.h. There's a man page for stdio.h too but the stdio functions also have their own man pages. Why are these two the only header files with man pages?

man 7 signal, the below part has an error:

• "Input" socket interfaces, when a timeout (SO_RCVTIMEO) has been set on the socket using setsockopt(2): accept(2), recv(2), recvfrom(2), recvmmsg(2) (also with a non-NULL timeout argument), and recvmsg(2).

• "Output" socket interfaces, when a timeout (SO_RCVTIMEO) has been set on the socket using setsockopt(2): connect(2), send(2), sendto(2), and sendmsg(2).

I suppose that for Output socket interfaces, the option should be SO_SNDTIMEO not SO_RCVTIMEO.
Please fix it.

In the example program on the man page, we are doing semget() call but checking for the shmget() failure.

Link to the section that should be corrected

In the example provided in the Description section, #include <xys.h> should be #include <xyz.h> to agree with the description.

Can we have some basic commands available to read here in this repo's README.md?

The website is currently outdated. It shows an outdated version of man-pages (5.13, released in 2021). Could you add a note saying that on the homepage?

A related PR wants to mention that this Github mirror is outdated too.

Can you clarify that when using ADJ_TAI that you need to fill in TAI-UTC, so currently +37, and not UTC-TAI as published by BIPM and currently -37?

It would also be useful to describe what happens to the clocks. I assume that CLOCK_REALTIME doesn't change, and that CLOCK_TAI will be CLOCK_REALTIME + buf.constant. I also assume that buf.tai will be set to buf.constant on return.