This is an enhancement proposal for better syntax check of the scripts.

Currently only "bash -n" is run for the *.sh scripts which cannot detect many problems like misspelled variables, functions, commands and many more.

My personal reason is that I did misspell a variable (that was never commited to rear upstream) and I found it only by debugging a failure during my own testing.

I don't know a good bash syntax checker.

In particular for rear we would neeed a bash syntax checker that can check the overall syntax of all *.sh scripts.

Therefore as a first attempt to do a little better syntax check I like to propse hereby to check for words that exist only once in all *.sh scripts.

What I did right now as a very first proof of concept in a rear source directory:

$ rm /tmp/rear_all_scripts.sh

$ find . -name '*.sh' | xargs cat >>/tmp/rear_all_scripts.sh

$ cat /tmp/rear_all_scripts.sh | grep -v '^[[:space:]]*#' | tr -c -s '[:space:][:alnum:]_-' ' ' | tr -s '[:space:]' '\n' | sort | uniq -c | grep '^[[:space:]]*1 [[:alpha:]]' | cut -b9- >/tmp/rear_all_scripts.lonely_words

$ for w in $( cat /tmp/rear_all_scripts.lonely_words ) ; do echo $w ; find . -name '*.sh' | xargs egrep -h "[^[:alpha:]]$w[^[:alpha:]]|^$w[^[:alpha:]]" ; echo ------------------------------------------------------------------------------------------------------------------ ; done | grep -v '^[[:space:]]*#' | less

This way I found 1159 lonely words in all *.sh scripts.

Now I use my own brain and imagination to view that long "less" output for suspicious cases of lonely words that could be real bugs...