Git bash and sqlcmd-CodePudding

I'm using a shell script in git bash to call sqlcmd to run some sql scripts. The script names are based on the git branch name, so the command is sqlcmd -E -S mySQLServer -d myDB "$branchsql"

It works fine from the command line, but I want to repeat it for several git branches, so I have a script that calls this script for a list of branches:

While read branch
do
  . C:/sqlScript.sh $branch
done < "$1"

The file with the list of branches is passed in $1

What happens is that is reads the first branch from the list, but never moves on to the next one. It repeatedly executes sqlScript.sh with the same value in $branch.

If I change sqlScript to just echo $1, everything works as expected. When I call sqlcmd, the first branch only is passed. So why does sqlcmd mess things up ?

CodePudding user response：

Drop the leading dot, invoke C:/sqlScript.sh ... rather than . C:/sqlScript.sh ...

. scripts.sh is short for source script.sh ,: it will execute the commands listed in script.sh in the current shell. If you have a command such as exit it will exit the current shell.

A regular invocation will start a separate shell, which won't mess with your current one.

CodePudding user response：

Thanks, but executing either way has the same result. If I execute sqlcmd in the script I call, I get this effect of only reading the first line from the list of branches. If I just do say Git checkout and git pull, it works. If I add sqlcmd, it gets stuck on the first branch.