shell
I've always found it awkward when complex shell commands involve parsing text output with AWK. Basically we're converting structured data into text, then parsing the text again to get back at the structured data.
PowerShell is a Windows shell which avoids that problem by allowing you to pipe typed objects between shell commands instead of just text.
Are there any Unix shells which are object-based instead of text-based, like PowerShell?
openssl's stdin is a byte stream.
The contents of $user is a sequence of non-0 bytes (which may or may not form valid characters in UTF-8 or other character set/encoding).
printf %s "$user"'s stdout is a byte stream.
printf %s "$user" | openssl dgst -sha256 –binary
Will connect printf's stdout with openssl's stdin. openssl's stdout is another byte stream.
Now, if you're inputing $user from the user from a terminal, The user will enter it by pressing keys on his keyboard. The terminal will send corresponding characters (as written on the key label) encoded in its configured character set. Usually, that character set will be based on the character set in the current locale. You can find what that is with locale charmap.
For instance, with a locale like fr_FR.iso885915@euro, and an xterm started in that locale, locale charmap will return ISO-8859-15. If the user enters stéphane as the username, that é will likely be encoded as the 0xe9 byte because that's how it's defined in the ISO-8859-15 character set.
If you want that é to be encoded as UTF-8 before passing to openssl, that's where you'd use iconv to convert that 0xe9 byte to the corresponding encoding in UTF-8 (two bytes: 0xc3 0xa9):
IFS= read -r user # read username from stdin as a sequence of bytes
# assumed to be encoded from characters as per the
# locale's encoding
printf %s "$user" |
iconv -t utf-8 | # convert from locale encoding to UTF-8
openssl dgst -sha256 –binary
I've had great results with this combination:
ps -p $$ | awk '$1 != "PID" {print $(NF)}'
On Tru64 (OSF/1), the output has parentheses around the shell. Tack on tr -d '()' to remove them.
ps -p $$ | awk '$1 != "PID" {print $(NF)}' | tr -d '()'
Appears to works on all shells, on Solaris 10 and RHEL 5.7 / 6.4. Didn't test other distros such as Debian, Ubuntu, or Mint, but I would think they should all work the same; I also have no idea if FreeBSD would work.
Best Answer
No, it is the other way around. There is no spoon^H^H^Hstructured data. There is only text.
A big part of the Unix philosophy is based on the idea of outputting text and accepting text as input. You might want to consider reading "The Art of Unix Programming", which has a nice explanation about this.
Don't get me wrong: I understand your point and I know what you are trying to get at. There are things like the interactive interpreters of Ruby and Python, which can be used as a shell, but they are not as friendly for basic tasks as Bash is. Try and change directory, for example.
Also, using objects in a shell is not all-that. If only your shell supports this, on Unix, you would be at a loss. All the standard Unix text manipulation tools would have to be altered, like grep, awk, sed, etc.
I think there has been an attempt to create something like this a few years back, but I can't remember the name and I haven't heard about it in a long time. It's probably not going to take off.