Merge pull request ipython#4172 from minrk/backport-pr

add ability to check what PRs should be backported in backport_pr

tools/backport_pr.py

@@ -2,24 +2,37 @@
 """
 Backport pull requests to a particular branch.
 
-Usage: backport_pr.py branch PR
+Usage: backport_pr.py branch [PR]
 
 e.g.:
 
-    backport_pr.py 0.13.1 123
+    python tools/backport_pr.py 0.13.1 123
 
 to backport PR #123 onto branch 0.13.1
 
+or
+
+    python tools/backport_pr.py 1.x
+
+to see what PRs are marked for backport that have yet to be applied.
+
 """
 
 from __future__ import print_function
 
 import os
+import re
 import sys
+
 from subprocess import Popen, PIPE, check_call, check_output
 from urllib import urlopen
 
-from gh_api import get_pull_request, get_pull_request_files
+from gh_api import (
+    get_issues_list,
+    get_pull_request,
+    get_pull_request_files,
+    is_pull_request,
+)
 
 def find_rejects(root='.'):
     for dirname, dirs, files in os.walk(root):
@@ -84,9 +97,47 @@ def backport_pr(branch, num, project='ipython/ipython'):
 
     return 0
 
+backport_re = re.compile(r"[Bb]ackport.*?(\d+)")
+
+def already_backported(branch, since_tag=None):
+    """return set of PRs that have been backported already"""
+    if since_tag is None:
+        since_tag = check_output(['git','describe', branch, '--abbrev=0']).decode('utf8').strip()
+    cmd = ['git', 'log', '%s..%s' % (since_tag, branch), '--oneline']
+    lines = check_output(cmd).decode('utf8')
+    return set(int(num) for num in backport_re.findall(lines))
+
+def should_backport(labels):
+    """return set of PRs marked for backport"""
+    issues = get_issues_list("ipython/ipython",
+            labels=labels,
+            state='closed',
+            auth=True,
+    )
+    should_backport = set()
+    for issue in issues:
+        if not is_pull_request(issue):
+            continue
+        pr = get_pull_request("ipython/ipython", issue['number'], auth=True)
+        if not pr['merged']:
+            print ("Marked PR closed without merge: %i" % pr['number'])
+            continue
+        should_backport.add(pr['number'])
+    return should_backport
+
 if __name__ == '__main__':
-    if len(sys.argv) < 3:
+
+    if len(sys.argv) < 2:
         print(__doc__)
         sys.exit(1)
 
+    if len(sys.argv) < 3:
+        branch = sys.argv[1]
+        already = already_backported(branch)
+        should = should_backport("backport-1.1")
+        print ("The following PRs should be backported:")
+        for pr in should.difference(already):
+            print (pr)
+        sys.exit(0)
+
     sys.exit(backport_pr(sys.argv[1], int(sys.argv[2])))

tools/gh_api.py

@@ -119,12 +119,13 @@ def get_pull_request_files(project, num, auth=False):
 element_pat = re.compile(r'<(.+?)>')
 rel_pat = re.compile(r'rel=[\'"](\w+)[\'"]')
 
-def get_paged_request(url, headers=None):
+def get_paged_request(url, headers=None, **params):
     """get a full list, handling APIv3's paging"""
     results = []
+    params.setdefault("per_page", 100)
     while True:
-        print("fetching %s" % url, file=sys.stderr)
-        response = requests.get(url, headers=headers)
+        print("fetching %s with %s" % (url, params), file=sys.stderr)
+        response = requests.get(url, headers=headers, params=params)
         response.raise_for_status()
         results.extend(response.json())
         if 'next' in response.links:
@@ -133,28 +134,32 @@ def get_paged_request(url, headers=None):
             break
     return results
 
-def get_pulls_list(project, state="closed", auth=False):
-    """get pull request list
-    """
-    url = "https://api.github.com/repos/{project}/pulls?state={state}&per_page=100".format(project=project, state=state)
+def get_pulls_list(project, auth=False, **params):
+    """get pull request list"""
+    params.setdefault("state", "closed")
+    url = "https://api.github.com/repos/{project}/pulls".format(project=project)
     if auth:
         headers = make_auth_header()
     else:
         headers = None
-    pages = get_paged_request(url, headers=headers)
+    pages = get_paged_request(url, headers=headers, params=params)
     return pages
 
-def get_issues_list(project, state="closed", auth=False):
-    """get pull request list
-    """
-    url = "https://api.github.com/repos/{project}/pulls?state={state}&per_page=100".format(project=project, state=state)
+def get_issues_list(project, auth=False, **params):
+    """get issues list"""
+    params.setdefault("state", "closed")
+    url = "https://api.github.com/repos/{project}/issues".format(project=project)
     if auth:
         headers = make_auth_header()
     else:
         headers = None
-    pages = get_paged_request(url, headers=headers)
+    pages = get_paged_request(url, headers=headers, **params)
     return pages
 
+def is_pull_request(issue):
+    """Return True if the given issue is a pull request."""
+    return bool(issue.get('pull_request', {}).get('html_url', None))
+
 # encode_multipart_formdata is from urllib3.filepost
 # The only change is to iter_fields, to enforce S3's required key ordering
 

tools/github_stats.py

@@ -13,7 +13,7 @@
 
 from datetime import datetime, timedelta
 from subprocess import check_output
-from gh_api import get_paged_request, make_auth_header, get_pull_request
+from gh_api import get_paged_request, make_auth_header, get_pull_request, is_pull_request
 
 #-----------------------------------------------------------------------------
 # Globals
@@ -50,12 +50,6 @@ def issues2dict(issues):
         idict[i['number']] = i
     return idict
 
-
-def is_pull_request(issue):
-    """Return True if the given issue is a pull request."""
-    return bool(issue.get('pull_request', {}).get('html_url', None))
-
-
 def split_pulls(all_issues, project="ipython/ipython"):
     """split a list of closed issues into non-PR Issues and Pull Requests"""
     pulls = []