[PATCH 0/7] Fix some bugs in abspath.c

From: Michael Haggerty <mhagger@alum.mit.edu>

I really just wanted to tidy up filter_refs(), but I've been sucked into a cascade of recursive yak shaving. This is my first attempt to pop the yak stack.

I want to use real_path() for making the handling of GIT_CEILING_DIRECTORIES more robust, but I noticed that it is broken for some cases that will be important in this context. This patch series adds some new tests of that function and fixes the ones that are broken, including a similar breakage in absolute_path(). It applies to the current master.

Please note that both absolute_path("") and real_path("") used to return the current directory followed by a slash. I believe that this was a bug, and that it is more appropriate for both functions to reject the empty string. The evidence is as follows:

* If this were intended behavior, presumably the return value would
  *not* have a trailing slash.

* I couldn't find any callers that appeared to depend on the old
  behavior.

* The test suite runs without errors after the change.

But I didn't do a thorough code review to ensure that no callers ever rely on the old behavior. The most likely scenario would be if one of these functions were used to parse something like $PATH, where the empty string denotes the current directory, but I didn't see any callers that seemed to be doing anything like this.

Michael Haggerty (7):
  t0000: verify that absolute_path() fails if passed the empty string
  absolute_path(): reject the empty string
  t0000: verify that real_path() fails if passed the empty string
  real_path(): reject the empty string
  t0000: verify that real_path() works correctly with absolute paths
  real_path(): properly handle nonexistent top-level paths
  t0000: verify that real_path() removes extra slashes

 abspath.c        | 12 ++++++++++--
 t/t0000-basic.sh | 31 ++++++++++++++++++++++++++++++-
 2 files changed, 40 insertions(+), 3 deletions(-)