Clever Geek Handbook

Why does reading act differently in bash and dash?

This is an attempt to figure out how a readable utility works across multiple shells. Found the difference which seems like a mistake to me.

The results are that the dash keeps trailing space with read:

dash: <a b     >
bash: <a b>

In short: why does this code work differently in bash and dash read?

dash -c 'echo "    a b     " | { read var; echo "<$var>"; }'
bash -c 'echo "    a b     " | { read var; echo "<$var>"; }'
+3


source to share


1 answer


Here's an easier way to demonstrate your problem:

$ dash -c 'echo "a b     " | { read var; echo "<$var>"; }'
<a b     >
$ bash -c 'echo "a b     " | { read var; echo "<$var>"; }'
<a b>

This only happens when there are two or more fields, for example "a b "

, and not when there is only one, for example "a "

.

Here's what POSIX says (emphasis mine):



If there are fewer vars than fields, the last var should be set to a value containing the following elements:

  • Field corresponding to the last var in the normal assignment sequence described above

  • Separator (s) that follow the field matching the last var

  • Remaining fields and their IFS trailing space delimiters are ignored

dash

does not ignore IFS whitespace, so it appears to violate POSIX.

Rule bash

.

+5


source






More articles:


All Articles