[Libreoffice-commits] dev-tools.git: esc-reporting/esc-analyze.py

[Guilhem Moulin (via logerrit)]

 esc-reporting/esc-analyze.py |   11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

New commits:
commit 8804bcdcbb6715ed00cdea8b36faaca579f883af
Author:     Guilhem Moulin <guilhem at libreoffice.org>
AuthorDate: Thu Jul 23 19:11:25 2020 +0200
Commit:     Guilhem Moulin <guilhem at libreoffice.org>
CommitDate: Wed Sep 2 01:04:34 2020 +0200

    esc-analyze: remove 'fullid' field and use number change IDs instead.
    
    fullid needs to uniquely identify a changeset, which the Change-ID does
    not.  (Cherry-picks preserve Change-IDs, which might cause `gerrit set-reviewers $CHANGE_ID`
    to fail when the Change-ID matches several changes.)
    
    We used to specify changes by Change-ID and not by change number due to
    potential conflicts between change numbers and abbreviated commit IDs,
    but the ambiguity is fixed since gerrit 1.14.5.1, see
    
        https://bugs.chromium.org/p/gerrit/issues/detail?id=1707
        https://gerrit-review.googlesource.com/c/gerrit/+/71936
        https://gerrit.libreoffice.org/Documentation/cmd-set-reviewers.html
    
    I guess it'll be more robust to replace the old-style change numeric IDs
    with full IDs $PROJECT~$BRANCH~$CHANGE_ID, but that's a more intrusive
    change.
    
    Change-Id: Ibf5bb02ed1ab5479a0994dc3fbf1763c288edc23
    Reviewed-on: https://gerrit.libreoffice.org/c/dev-tools/+/99333
    Reviewed-by: Xisco Fauli <xiscofauli at libreoffice.org>
    Reviewed-by: Guilhem Moulin <guilhem at libreoffice.org>
    Tested-by: Guilhem Moulin <guilhem at libreoffice.org>

diff --git a/esc-reporting/esc-analyze.py b/esc-reporting/esc-analyze.py
index dabca7c..1176c32 100755
--- a/esc-reporting/esc-analyze.py
+++ b/esc-reporting/esc-analyze.py
@@ -687,10 +687,7 @@ def analyze_reports():
         row['status'] = 'NEW'
       xDate = datetime.datetime.strptime(row['updated'], '%Y-%m-%d %H:%M:%S.%f000')
       ownerEmail = util_check_mail(row['owner']['name'], row['owner']['email'])
-      # while web is happy with the unique project~branch~changeID label, commandline interface
-      # only accepts ambiguous changeID, doesn't help, so fullid is not really fullid, but at least
-      # less prone to conflicts than just changeset-number that also can easily prefix-match commit-hashes
-      entry = {'id': key, 'fullid': row['change_id'], 'name': row['owner']['name'], 'email': ownerEmail, 'title': row['subject']}
+      entry = {'id': key, 'name': row['owner']['name'], 'email': ownerEmail, 'title': row['subject']}
 
       if row['status'] != 'ABANDONED':
         if ownerEmail is None:
@@ -738,7 +735,7 @@ def analyze_reports():
           else:
             automateList['gerrit']['to_abandon_comment'][entry['id']] = patchset
         if cntReview == 0 and not statList['people'][ownerEmail]['isCommitter']:
-            tmpListToReview.append({'id': entry['id'], 'fullid': entry['fullid'], 'patchset': patchset})
+            tmpListToReview.append({'id': entry['id'], 'patchset': patchset})
 
     for rowTmp in tmpListToReview:
       if gerritData['patch'][rowTmp['id']]['project'] == 'online':
@@ -763,8 +760,8 @@ def analyze_reports():
         except Exception as e:
           pass
       x = statList['people'][reviewEmail]
-      automateList['gerrit']['to_review'][rowTmp['fullid']] = {'name': statList['people'][reviewEmail]['gerrit']['reviewName'],
-                                                               'patchset': rowTmp['patchset'], 'id': rowTmp['id']}
+      automateList['gerrit']['to_review'][rowTmp['id']] = {'name': statList['people'][reviewEmail]['gerrit']['reviewName'],
+                                                           'patchset': rowTmp['patchset'], 'id': rowTmp['id']}
 
     for key, row in bugzillaData['bugs'].items():
       if not 'cc' in row: