Whilst tinkering with history settings, I keep seeing references to the 'history comment character' in the man pages. Is this different to using # for commenting in Bash?
I can't seem to find an answer as to what actual character code is for the history comment character and how it differs to a conventional bash comment?
This is how I understand it:
There is the "normal" comment character, #, which can't be changed. It indicates the start of a comment in scripts and in interactive sessions (unless the interactive_comments shell option is disabled), and everything from that character on is ignored.
The history comment character can be set using histchars; the default value is !^#.
! is used to indicate history expansion (like "repeat previous command": !!)
^ is to substitute something in the previous command:
$ echo foo foo
foo foo
$ ^foo^bar^ # final ^ is optional
bar foo
# is the history comment character.
The history comment character is used:
In ~/.bash_history, to mark timestamps as comments when HISTTIMEFORMAT is set:
#1597532894
echo foo
#1597532908
histchars='!^@'
@1597532918
vim ~/.bash_history
Notice how the timestamp prefix switches to @ after the second command.
To skip history substitution for the remaining words on a line:
$ echo foo
foo
$ echo !$ @ !$ # !$
foo @ foo
$ histchars='!^@'
$ echo foo
foo
$ echo !$ @ !$ # !$
foo @ !$
!$ is "the last word of the previous command". With the default settings (history comment character is #), echo !$ @ !$ # !$ expands to echo foo @ foo # !$ and prints foo @ foo; everything after # is ignored. When switching the history comment character to @, the second !$ is not expanded to foo any longer.
When would you use that? I honestly don't know, and it really only applies to interactive shell sessions, where the default of # make total sense, and anything else would be highly surprising behaviour to me. I can't find anything in the Bash release notes to explain what the intended purpose of being able to separately set the history comment character is.