2017-01-08



Just realized that Michi writes a Unix diary  as  well.  :-)  [0]
There I found this entry:  [1]

     # 2016-12-22

     I added a new alias to my .profile today:

             function cdl {
                     cd $1
                     ls -h $1
             }

I have some comments:

The variable expansions  should  be  quoted,  so  it  works  with
filenames that contain whitespace as well.

Futhermore, it would be better to use:
     cd "$1" && ls -h
because the ls(1) should only be executed if the cd was  success-
ful. Otherwise you would get two error messages.  If you make the
ls(1) call conditional, then you can omit the directory argument,
because ls(1) either will be in the right directory or not be ex-
ecuted at all. (Without the `&&', ls(1) would  list  the  current
directory if the cd failed.)

And to eliminate the bashism, instead of ``function  cdl  {...}''
one should use ``cdl() {...}'' -- it's the same, just portable.


Since some weeks, I'm thinking about adding this function  to  my
standard set:

     md() {
             mkdir "$1" && cd "$1"
     }

This one is quite common. But when I think about  the  number  of
systems  I  would  have to deploy it, and if I think about how it
feels when I don't have it available (on all those other  systems
I  work on every once in a while) ... then I rather stick to Sven
Mascheck's opinion to (in the extreme) rather type `ls  -l'  than
wondering if `ll' is available.

I'm not that extreme though, I do have some aliases,  essentially
these:

     alias ls='ls -F'
     alias ll='ls -lF --time-style=long-iso'
     alias lh='ls -lhF --time-style=long-iso'
     alias la='ls -AF'

Only the `ls' and `ll' aliases I really need. The other  two  are
nice to have.


On my development machines, I usually  have  ~/bin/rr,  which  is
some  kind  of recursive grep for searching the code base.  Some-
times I implement it with `egrep -r', sometimes (when I'm more on
the grep(1) should have no `-r' trip), I use:

     find "$@" -type f | sort | xargs -d '\n' grep -I "$re"

(I should try to rather use egrep(1) than grep(1) because of  the
speed as well as the more powerful regexps.)

I don't really care much  for  the  implementation  of  the  `rr'
script,  it just saves me some typing for the usual cases, mainly
I search for `identifier' or  `\<identifier\>'  or  `\<identifier
*('. For more advanced searches, I usually enter the grep command
directly.

On my email machine I additionally have a  whole  bunch  of  mmh-
related  scripts  (which are essentially on-liners), but that's a
single machine, so there's no need to move anything or to keep it
in sync.


[0]  http://www.micha.elmueller.net/unix-diary/
[1]  http://www.micha.elmueller.net/unix-diary/entries/2016-12-
19.txt


http://marmaro.de/lue/        markus schnalke <meillo@marmaro.de>