fetch2: Drop globbing supprt in file:// SRC_URIs

Globbing in file:// urls is terminally broken. Currently when its used, the
file checksum code is basically bypassed. This means changes to the source
files don't change the task checksum, the task doesn't rebuild when the
inputs change and things generally break.

To make globbing work generically, we'd have to scan the file system for
all possible matches to the glob and log whether they exist or not. We can't
simply log the files which exist, we have to also know which files could
later exist and influence the choice of file so we know when to reparse.

For a simple file://xxx/*, this could be done but for bigger patterns,
it becomes much more problemtic. We already support file://xxx/ in urls.

So, lets decide we'll not support globs in file://urls. Worse case users
can put files in a directory and reference that, moving files into place
if needed.

Remove all the glob special cases (see the comments if anyone doesn't
believe this is terminally broken) and error to the user if they have
such urls.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

1 files changed, 1 insertions, 1 deletions

diff --git a/lib/bb/cache.py b/lib/bb/cache.py
index b819a0c2d..9e0c931a0 100644
--- a/lib/bb/cache.py
+++ b/lib/bb/cache.py
@@ -636,7 +636,7 @@ class Cache(NoCache):
                 # Have to be careful about spaces and colons in filenames
                 flist = self.filelist_regex.split(fl)
                 for f in flist:
-                    if not f or "*" in f:
+                    if not f:
                         continue
                     f, exist = f.split(":")
                     if (exist == "True" and not os.path.exists(f)) or (exist == "False" and os.path.exists(f)):