user-friendliness sanity checks: catch common mistakes regarding $1/$2/$3.

.do files should never modify $1, and should write to *either* $3 or stdout,
but not both.  If they write to both, it's probably because they forgot to
redirect stdout to stderr, a very easy mistake to make but a hard one to
detect.

Now redo detects it for you and prints an informative message.
This commit is contained in:
Avery Pennarun 2010-11-22 04:40:54 -08:00
commit 6d767e2a65
8 changed files with 55 additions and 4 deletions

2
t/deps/.gitignore vendored
View file

@ -1,2 +1,4 @@
t1a
t2.count
overwrite
overwrite[123]

View file

@ -1 +1,2 @@
rm -f *~ .*~ *.count t1a
rm -f *~ .*~ *.count t1a overwrite overwrite[123]

4
t/deps/overwrite.do Normal file
View file

@ -0,0 +1,4 @@
redo overwrite1 2>&1 && exit 55
redo overwrite2 2>&1 && exit 56
redo overwrite3 2>&1 && exit 57
exit 0

2
t/deps/overwrite1.do Normal file
View file

@ -0,0 +1,2 @@
# this shouldn't be allowed; we're supposed to write to $3, not $1
echo >$1

5
t/deps/overwrite2.do Normal file
View file

@ -0,0 +1,5 @@
# this shouldn't be allowed; stdout is connected to $3 already, so if we
# replace it *and* write to stdout, we're probably confused.
echo hello world
rm -f $3
echo goodbye world >$3

8
t/deps/overwrite3.do Normal file
View file

@ -0,0 +1,8 @@
# we don't delete $3 here, we just truncate and overwrite it. But redo
# can detect this by checking the current file position of our stdout when
# we exit, and making sure it equals either 0 or the file size.
#
# If it doesn't, then we accidentally wrote to *both* stdout and a separate
# file, and we should get warned about it.
echo hello world
echo goodbye world >$3

View file

@ -1,3 +1,3 @@
redo test1 test2 ifchange-fail
redo test1 test2 ifchange-fail overwrite