buildah-bud tests: reenable skipped 'bud' tests

Sigh. Buildah PR https://github.com/containers/buildah/pull/3368
changed 'bud' to 'build' in tests. Podman #11585 well-intentionedly
did the same for run-buildah-bud tests ... but did so by *replacing*
'bud' with 'build', not by *adding* 'build' to the list of commands
handled by podman-build. Hence, all tests invoking 'run_buildah bud'
have been completely untested since then.

This remedies that, and deals with all the fallout. Principal among
which is the discovery that our exit-code changes are no longer
necessary: that thing we did where buildah exit status 1 or 2 became
podman exit status 125? That no longer applies. podman now exits
with the same status as buildah. This simplifies our diffs, and
lets us enable a bunch more tests.

Also:
 - in run-buildah-bud-tests script, run 'sudo --validate' early on.
   Reason: otherwise, the sudo step happens a few minutes after
   the script starts (after the git-pull), by which time the user
   may have stepped away to get coffee, then comes back ten or twenty
   minutes later to find a stupid sudo prompt and no tests run.

Signed-off-by: Ed Santiago <santiago@redhat.com>

3 files changed, 31 insertions, 21 deletions

diff --git a/test/buildah-bud/apply-podman-deltas b/test/buildah-bud/apply-podman-deltas
index 02a8fc222..03e3f8661 100755
--- a/test/buildah-bud/apply-podman-deltas
+++ b/test/buildah-bud/apply-podman-deltas
@@ -150,13 +150,7 @@ errmsg "checking authfile: stat /tmp/nonexistent: no such file or directory" \
 ###############################################################################
 # BEGIN tests that don't make sense under podman due to fundamental differences
 
-# TODO
-# Normally, when buildah exits 1 on error, podman exits 125.
-# These tests are the exception. They exit 1 under podman.
-skip "these tests exit 1 under podman, not 125" \
-     "bud with --add-host" \
-     "bud - invalid runtime flags test"
-
+# Fails with "Error: no context directory and no Containerfile specified"
 skip "does not work under podman" \
      "bud without any arguments should succeed"
 
@@ -167,8 +161,9 @@ skip "does not work under podman" \
 skip "FIXME FIXME FIXME: argument-order incompatible with podman" \
      "bud-squash-hardlinks"
 
-skip "FIXME FIXME FIXME: this passes on Ed's laptop, fails in CI??" \
-     "bud-multi-stage-nocache-nocommit"
+# Fails with "Error: context must be a directory: /path/to/Dockerfile"
+skip "podman-build fails with 'context must be a directory'" \
+     "bud with specified context should succeed if context contains existing Dockerfile"
 
 ###############################################################################
 # BEGIN tests which are skipped because they make no sense under podman-remote
@@ -235,4 +230,12 @@ skip_if_remote "FIXME FIXME FIXME: find a way to clean up their podman calls" \
                "bud with run should not leave mounts behind cleanup test" \
                "bud with custom files in /run/ should persist cleanup test"
 
+# Under podman-remote, the "Ignoring <stdin>:5:2: error: #error" message
+# is never seen. (Not even as stdout/stderr on the server; Ed checked).
+skip_if_remote "FIXME FIXME FIXME: 'Ignoring' warning is never seen" \
+               "bud with preprocessor error"
+
+# END   tests which are skipped due to actual podman or podman-remote bugs.
+###############################################################################
+
 exit $RC
diff --git a/test/buildah-bud/buildah-tests.diff b/test/buildah-bud/buildah-tests.diff
index 6fa36d904..399042240 100644
--- a/test/buildah-bud/buildah-tests.diff
+++ b/test/buildah-bud/buildah-tests.diff
@@ -1,15 +1,15 @@
-From 8a8fa1a75e0fa3261263afbc8c2504feb430df6a Mon Sep 17 00:00:00 2001
+From 6508e3df2a129554fdf8336d8a6f0cdcc6fd4832 Mon Sep 17 00:00:00 2001
 From: Ed Santiago <santiago@redhat.com>
 Date: Tue, 9 Feb 2021 17:28:05 -0700
 Subject: [PATCH] tweaks for running buildah tests under podman
 
 Signed-off-by: Ed Santiago <santiago@redhat.com>
 ---
- tests/helpers.bash | 69 ++++++++++++++++++++++++++++++++++++++++++++--
- 1 file changed, 66 insertions(+), 3 deletions(-)
+ tests/helpers.bash | 70 ++++++++++++++++++++++++++++++++++++++++++++--
+ 1 file changed, 67 insertions(+), 3 deletions(-)
 
 diff --git a/tests/helpers.bash b/tests/helpers.bash
-index e3087063..b3a8f5ee 100644
+index e3087063..178a486e 100644
 --- a/tests/helpers.bash
 +++ b/tests/helpers.bash
 @@ -51,6 +51,23 @@ EOF
@@ -60,13 +60,13 @@ index e3087063..b3a8f5ee 100644
  # There are various scenarios where we would like to execute `tests` as rootless user, however certain commands like `buildah mount`
  # do not work in rootless session since a normal user cannot mount a filesystem unless they're in a user namespace along with its
  # own mount namespace. In order to run such specific commands from a rootless session we must perform `buildah unshare`.
-@@ -247,8 +274,35 @@ function run_buildah() {
+@@ -247,8 +274,36 @@ function run_buildah() {
          --retry)         retry=3;        shift;;  # retry network flakes
      esac
  
 +    local podman_or_buildah=${BUILDAH_BINARY}
 +    local _opts="${ROOTDIR_OPTS} ${BUILDAH_REGISTRY_OPTS}"
-+    if [[ $1 == "build" || $1 == "build-using-dockerfile" ]]; then
++    if [[ $1 == "build" || $1 == "build-using-dockerfile" || $1 == "bud" ]]; then
 +        shift
 +        # podman defaults to --layers=true; buildah to --false.
 +        # If command line includes explicit --layers, leave it untouched,
@@ -82,10 +82,11 @@ index e3087063..b3a8f5ee 100644
 +            _opts=
 +        fi
 +
-+        # podman always exits 125 where buildah exits 1 or 2 (or, in the
-+        # case of git, 128, which is a bug in git, but I won't harp on that).
++        # Special case: there's one test that invokes git in such
++        # a way that it exits 128 (which IMO is a bug in git).
++        # podman exits 125 in that case.
 +        case $expected_rc in
-+            1|2|128)   expected_rc=125 ;;
++            128)   expected_rc=125 ;;
 +        esac
 +    fi
 +    local cmd_basename=$(basename ${podman_or_buildah})
@@ -97,7 +98,7 @@ index e3087063..b3a8f5ee 100644
  
      # If session is rootless and `buildah mount` is invoked, perform unshare,
      # since normal user cannot mount a filesystem unless they're in a user namespace along with its own mount namespace.
-@@ -262,8 +316,8 @@ function run_buildah() {
+@@ -262,8 +317,8 @@ function run_buildah() {
          retry=$(( retry - 1 ))
  
          # stdout is only emitted upon error; this echo is to help a debugger
@@ -108,7 +109,7 @@ index e3087063..b3a8f5ee 100644
          # without "quotes", multiple lines are glommed together into one
          if [ -n "$output" ]; then
              echo "$output"
-@@ -595,6 +649,15 @@ function skip_if_no_docker() {
+@@ -595,6 +650,15 @@ function skip_if_no_docker() {
    fi
  }
  
@@ -125,5 +126,5 @@ index e3087063..b3a8f5ee 100644
    daemondir=${TEST_SCRATCH_DIR}/git-daemon
    mkdir -p ${daemondir}/repo
 -- 
-2.35.1
+2.35.3
 
diff --git a/test/buildah-bud/run-buildah-bud-tests b/test/buildah-bud/run-buildah-bud-tests
index eb8de5618..4ff062496 100755
--- a/test/buildah-bud/run-buildah-bud-tests
+++ b/test/buildah-bud/run-buildah-bud-tests
@@ -93,6 +93,12 @@ fi
 # From here on out, any error is fatal
 set -e
 
+# Run sudo early, to refresh the credentials cache. This is a NOP under CI,
+# but might be appreciated by developers who run this script,  step away
+# during the git-checkout-buildah step, then come back twenty minutes later
+# to an expired sudo prompt and no tests have run.
+sudo --validate
+
 # Before pulling buildah (while still cd'ed to podman repo), try to determine
 # if this is a PR, and if so if it's a revendoring of buildah. We use this to
 # try to offer a helpful hint on failure.