in progress

tests/push-basic.sh

 #!/bin/bash
 
 # Copyright (c) 2012 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 set -e
 
 . ./test-lib.sh
 
@@ skipping 27 matching lines @@
 
   # Push a description to this URL.
   URL=$($GIT_CL_STATUS | sed -ne '/Issue number/s/[^(]*(\(.*\))/\1/p')
   curl --cookie dev_appserver_login="test@example.com:False" \
        --data-urlencode subject="test" \
        --data-urlencode description="foo-quux" \
        --data-urlencode xsrf_token="$(print_xsrf_token)" \
        $URL/edit
 
   test_expect_success "Base URL contains branch name" \
-      "curl -s $($GIT_CL_STATUS --field=url) | grep 'URL:' | grep -q '@master'"
+      "curl -s $($GIT_CL_STATUS --field=url) | grep 'URL:[[:space:]]*[^<]' | grep -q '@master'"

[iannucci, 2014/03/18 20:38:59]

OMFG NO... curling the html output is *horror*. This is literally trying to
parse data out of html with regular expressions.

We should be hitting <server>/api/<cl number> instead, which returns json. I
think it would be really easy to make git-cl-status return this URL instead (or
even add a git-cl-api command to dump json).

This is 100% guaranteed to break again.

 
   test_expect_success "git-cl push ok" \
     "$GIT_CL push -f"
 
   git checkout -q master > /dev/null 2>&1
   git pull -q > /dev/null 2>&1
 
   test_expect_success "committed code has proper description" \
       "git show | grep -q 'foo-quux'"
 
   test_expect_success "issue no longer has a branch" \
       "$GIT_CL_STATUS | grep -q 'work : None'"
 
   cd $GITREPO_PATH
   test_expect_success "upstream repo has our commit" \
       "git log master 2>/dev/null | grep -q 'foo-quux'"
 )
 SUCCESS=$?
 
 cleanup
 
 if [ $SUCCESS == 0 ]; then
   echo PASS
 fi