Although mknod is not a POSIX tool, it is widely available on nearly
all UNIX-like systems. It also can be implemented portably apart
from use of the makedev macros, which is already a requirement of
a couple other tools in sbase.
While we're at it, fix a few bugs:
- Include sys/sysmacros.h if makedev was not defined by sys/types.h
- The default mode should respect the user's umask, rather than
assuming it is 022.
- Clear the umask when -m is specified explicitly so that nodes can
be created with permissions wider than the user's umask.
- Utilize parsemode from libutil to support symbolic mode strings.
For sort(1) we need memmem(), which I imported from OpenBSD.
Inside sort(1), the changes involved working with the explicit lengths
given by getlines() earlier and rewriting some of the functions.
Now we can handle NUL-characters in the input just fine.
The assumption of NUL-terminated strings is actually quite a good one in
most cases. You don't have to worry about paths, because they may not
Same applies to arguments passed to you. Unless you have to unescape,
there is no way for you to receive a NUL.
There are two important exceptions though, and it's important that we
address them, or else we get unexpected behaviour:
1) All tools using unescape() have to be strict about delimlen.
Else they end up for instance unescaping
which in C's string-vision is an empty string.
2) All tools doing line wrenching and putting them out
again as lines again.
puts() will cut each line containing NULs off at the first
-s strip binary
-d create directory
-D create missing directories
-t DIR target directory
-m MODE permission bits
-o USER set owner
-g GROUP set group
Installed files are copied, and default mode is 755.
Signed-off-by: Mattias Andrée <firstname.lastname@example.org>
Looking at the old code, it became clear that the desired
functionality with the t-flag could not be added unless the
underlying data-structures were reworked.
Thus the only way to be successful was to rewrite the whole thing.
od(1) allows giving arbitrarily many type-specs per call, both via
-t x1o2... and -t x1 -t o2 and intermixed.
This fortunately is easy to parse.
Now, to be flexible, it should not only support types of integral
length. Erroring out like this is inacceptable:
$ echo -n "shrek"| od -t u3
od: invalid type string ‘u3’;
this system doesn't provide a 3-byte integral type
Thus, this new od(1) just collects the bytes until shortly before
printing, when the numbers are written into a long long with the
The bytes per line are just the lcm of all given type-lengths and >= 16.
They are equal to 16 for all types that are possible to print using
the old od(1)'s.
Endianness is of course also supported, needs some testing though,
especially on Big Endian systems.
The logic is simple, it's just a pain in the ass to fill the
Some lines had to be commented out, as glibc/musl apparently
have not fully implemented the mandatory variables for the
2013 corrigendum of POSIX 2008.
Also added a manpage and the necessary entries in README.
I also removed it from the TODO.
If this flag is not given, od(1) automatically replaces duplicate
adjacent lines with an '*' for each reoccurence.
If this flag is set, thus, no such filtering occurs.
In this case this would mean having to somehow keep the last printed
line in some backbuffer, building the next line and then doing the
necessary comparisons. This basically means that we duplicate the
functionality provided with uniq(1).
So instead of
$ od -t a > dump
you'd rather do
$ od -t a | uniq -f 1 -c > dump
Skipping the first field is necessary, as the addresses obviously differ.
Now, I was thinking hard why this flag even exists. If POSIX mandated
to add the address before the asterisk, so we know the offset of duplicate
occurrences, this would make sense. However, this is not the case.
Using uniq(1) also gives nicer output:
~ $ echo "111111111111111111111111111111111111111111111111" | od -t a -v | uniq -f 1 -c
3 0000000 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 0000060 nl
in comparison to
$ echo "111111111111111111111111111111111111111111111111" | od -t a
0000000 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Before working on od(1), I didn't even know it would filter out
duplicate adjacent lines like that. This is also a matter of
Concluding, the v-flag is implicitly set and users urged to just
use the existing tools provided by the system.
I don't think we would break scripts either. Firstly, it's rather
unlikely to have duplicate lines exactly matching the line-length of
od(1). Secondly, even if a script did that specifically, in the worst
case there would be a counting error or something.
Given od(1) is mostly used interactively, we can safely assume this
feature is for the benefit of the users.
Ditch this legacy POSIX crap!
Please enter the commit message for your changes. Lines starting
This client does not support the netascii mode. The default mode
is octet/binary and should be sufficient.
One thing left to do is to check the source port of the server
to make sure it doesn't change. If it does, we should ignore the
packet and send an error back without disturbing an existing
1) Remove the function prototypes. No need for them, as the
functions are ordered.
2) Add fieldseplen, so the length of the field-separator is not
calculated nearly each time skipcolumn() is called.
3) rename next_col to skip_to_next_col so the purpose is clear,
also reorder the conditional accordingly.
4) Put parentheses around certain ternary expressions.
5) BUGFIX: Don't just exit() in check(), but make it return something,
so we can cleanly fshut() everything.
6) OFF-POSIX: Posix for no apparent reason does not allow more than
one file when the -c or -C flags are given.
This can be problematic when you want to check multiple files.
With the change 5), rewriting check() to return a value, I went
off-posix after discussing this with Dimitris to just allow
arbitrary numbers of files. Obviously, this does not break scripts
and is convenient for everybody who wants to quickly check a big
amount of files.
As soon as 1 file is "unsorted", the return value is 1, as expected.
For convenience reasons, check()'s warning now includes the filename.
7) BUGFIX: Set ret to 2 instead of 1 when the fshut(fp, *argv) fails.
8) BUGFIX: Don't forget to fshut stderr at the end. This would improperly
return 1 in the following case:
$ sort -c unsorted_file 2> /dev/full
9) Other style changes, line length, empty line before return.
Make it clear that <blank> characters just are spaces or tabs and
not a special group which needs special treatment for wide characters.
Also, and that was the only problem here, correctly calculate the
offset given by the key definitions for the start- and end-characters
Mark the progress in the README and put parentheses around the missing
flags which are insane to implement for no real gain.
Use the *at functions instead of building paths manually. We do
still have path-building in recurse() and other areas, but the
long-term goal is to rid most interfaces of that for practical
and security reasons.
In this case, it's more or less trivial.
Also, refactor the manpage to be more consistent with the others.
BUGFIX: Return exit status 3 on error.
Mostly manpage-shuffling according to the changes in the corrigendum,
wording changes and more idiomatic expressions.
All this is finished up by marking the POSIX 2013 conformant tools
which is not available in older mandoc builds or nroff, but which
reflects what we actually did, so who cares?
This is a huge step and it's not far until we can release sbase 0.1.
I can't believe we've come this far! The idea is to look at the
2013 POSIX corrigendum for each tool and deep-test features before
making the first 0.1 release.
To keep the noise low, I'll do this in batches, not on a per-tool-
basis (as many of these are trivial to test).
In the meantime, I'll also think of a fitting STANDARDS section
for the non-POSIX tools. Now that the audits are pretty much done,
I can also have a more relaxed view on standards compliance instead
of having to dig through some uncleaned mess.
To mark this "new beginning", the README has gotten a liftover.
The POSIX 2008-column was more or less useless and as I expect the
checks to go along pretty quickly, I "reset" the compliance state
of all but the non-POSIX tools and will then go along and check every
single one of them in the next few days.
Apart from the few missing flags and audits, sbase should then be
ready to hit the world with the first release after 4 years of work.
Sort comes pretty much automatically, as no script relies on the
undefined behaviour of the input _not_ being sorted, we might as well
sort the sorted input already.
The only downside is memory usage, which can be an issue for large
The o-flag was trivial to implement.