Process results and unit-test output of libcxx

Now "report" step combines result in a uniform way and processes unit test
results XML output. It works for sub-builds only started from the 'premerge'
pipeline, i.e. non-recursive. One downside is that now one has to wait until
all jobs have finished.

- Add instructions to setup python environment

- added option to do full report cycle but not call Phabricator

- use "annotations" to show build status. That lifts the need to filter ninja
  and other output (thus `ph_no_filter_output` param removed) and output
  everything. That is nice as script failures no longer lead to loss of logs.

- improved annotate() usability

- misc fixes

docs/playbooks.md

@@ -10,6 +10,20 @@
   
 # Playbooks
 
+## Development environment
+
+You need will need recent python 3 installed, e.g. follow this
+[installation guide](https://cloud.google.com/python/docs/setup?hl=en).
+To install required packages run:
+
+```shell script
+pip install -r ./scripts/requirements.txt
+```
+optional:
+```shell script
+pip install jupyterlab pandas seaborn #  for jupyter labs.  
+```  
+
 ## Testing scripts locally
 
 Build and run agent docker image `sudo ./containers/build_run.sh buildkite-premerge-debian /bin/bash`.
@@ -165,11 +179,11 @@ please refer to the source code for the details. These variables have `ph_` pref
 Most commonly used are:
 
 - `ph_scripts_refspec` ("master" by default): refspec branch of llvm-premerge-checks to use. This variable is also used in pipeline "bootstrap" in Buildkite interface.
+- `ph_dry_run_report`: do not report any results back to Phabricator.
 - `ph_no_cache`: (if set to any value) clear compilation cache before the build.
 - `ph_projects`: which projects to use, "detect" will look on diff to infer the projects, "default" selects all projects.
 - `ph_notify_email`: comma-separated list of email addresses to be notified when build is complete.
 - `ph_log_level` ("DEBUG", "INFO", "WARNING" (default) or "ERROR"): log level for build scripts. 
-- `ph_no_filter_output` (if set to any value): do not filter output of `ninja all` and other commands from buildkite log.
 - `ph_linux_agents`, `ph_windows_agents`: custom JSON constraints on agents. For example you might put one machine to a custom queue if it's errornous and send jobs to it with `ph_windows_agents="{{\"queue\": \"custom\"}}"`.
 - `ph_skip_linux`, `ph_skip_windows` (if set to any value): skip build on this OS.
 

scripts/add_phabricator_artifact.py

@@ -28,5 +28,6 @@ if __name__ == '__main__':
     args = parser.parse_args()
 
     logging.basicConfig(level=args.log_level, format='%(levelname)-7s %(message)s')
-    PhabTalk(os.getenv('CONDUIT_TOKEN')).maybe_add_url_artifact(args.phid, args.url, args.name)
+    dry = os.getenv('ph_dry_run_report') is not None
+    PhabTalk(os.getenv('CONDUIT_TOKEN'), dry_run_updates=dry).maybe_add_url_artifact(args.phid, args.url, args.name)
 

scripts/buildkite_utils.py

@@ -3,13 +3,14 @@ import os
 import re
 import subprocess
 import urllib.parse
-import shlex
 from typing import Optional
+from benedict import benedict
 
 import backoff
 import requests
 
 context_style = {}
+previous_context = 'default'
 styles = ['default', 'info', 'success', 'warning', 'error']
 
 
@@ -30,26 +31,30 @@ def upload_file(base_dir: str, file: str):
         return None
 
 
-def annotate(message: str, style: str = 'default', context: str = 'default', append: bool = True):
+def annotate(message: str, style: str = 'default', context: Optional[str] = None, append: bool = True):
     """
     Adds an annotation for that currently running build.
     Note that last `style` applied to the same `context` takes precedence.
     """
+    global previous_context, styles, context_style
     if style not in styles:
         style = 'default'
+    if context is None:
+        context = previous_context
+    previous_context = context
     # Pick most severe style so far.
     context_style.setdefault(context, 0)
     context_style[context] = max(styles.index(style), context_style[context])
     style = styles[context_style[context]]
     if append:
         message += '\n\n'
-    r = subprocess.run(f"buildkite-agent annotate {shlex.quote(message)}"
+    cmd = ['buildkite-agent', 'annotate', message, '--style', style, '--context', context]
-                       f' --style={shlex.quote(style)}'
+    if append:
-                       f" {'--append' if append else ''}"
+        cmd.append('--append')
-                       f" --context={shlex.quote(context)}", shell=True, capture_output=True)
+    r = subprocess.run(cmd, capture_output=True)
     logging.debug(f'annotate call {r}')
     if r.returncode != 0:
-        logging.warning(message)
+        logging.warning(r)
 
 
 def feedback_url():
@@ -63,18 +68,24 @@ class BuildkiteApi:
         self.token = token
         self.organization = organization
 
-    @backoff.on_exception(backoff.expo, Exception, max_tries=3, logger='', factor=3)
     def get_build(self, pipeline: str, build_number: str):
-        authorization = f'Bearer {self.token}'
         # https://buildkite.com/docs/apis/rest-api/builds#get-a-build
-        url = f'https://api.buildkite.com/v2/organizations/{self.organization}/pipelines/{pipeline}/builds/{build_number}'
+        return benedict(self.get(f'https://api.buildkite.com/v2/organizations/{self.organization}/pipelines/{pipeline}/builds/{build_number}').json())
-        response = requests.get(url, headers={'Authorization': authorization})
+
+    @backoff.on_exception(backoff.expo, Exception, max_tries=3, logger='', factor=3)
+    def get(self, url: str):
+        authorization = f'Bearer {self.token}'
+        response = requests.get(url, allow_redirects=True, headers={'Authorization': authorization})
         if response.status_code != 200:
-            raise Exception(f'Builkite responded with non-OK status: {response.status_code}')
+            raise Exception(f'Buildkite responded with non-OK status: {response.status_code}')
-        return response.json()
+        return response
 
 
 def format_url(url: str, name: Optional[str] = None):
     if name is None:
         name = url
     return f"\033]1339;url='{url}';content='{name}'\a\n"
+
+
+def strip_emojis(s: str) -> str:
+    return re.sub(r':[^:]+:', '', s).strip()

scripts/clang_format_report.py

@@ -22,6 +22,7 @@ import pathspec
 import unidiff
 
 from phabtalk.phabtalk import Report, Step
+from buildkite_utils import annotate
 
 
 def get_diff(base_commit) -> Tuple[bool, str]:
@@ -93,8 +94,8 @@ def run(base_commit, ignore_config, step: Optional[Step], report: Optional[Repor
         report.add_artifact(os.getcwd(), patch_file, 'clang-format')
     if not success:
         step.success = False
-        step.messages.append(
+        annotate(f'clang-format: Please format your changes with clang-format by running `git-clang-format HEAD^`'
-            'Please format your changes with clang-format by running `git-clang-format HEAD^` or applying patch.')
+                 f'or applying the attached patch.', style='error')
     logging.debug(f'report: {report}')
     logging.debug(f'step: {step}')
 

scripts/clang_tidy_report.py

@@ -22,7 +22,7 @@ from typing import Optional
 import pathspec
 
 import ignore_diff
-from buildkite_utils import format_url
+from buildkite_utils import annotate
 from phabtalk.phabtalk import Report, Step
 
 
@@ -100,11 +100,9 @@ def run(base_commit, ignore_config, step: Optional[Step], report: Optional[Repor
         report.add_artifact(os.getcwd(), p, 'clang-tidy')
     if errors_count + warn_count != 0:
         step.success = False
-        url = format_url("https://github.com/google/llvm-premerge-checks/blob/master/docs/clang_tidy.md"
+        url = "https://github.com/google/llvm-premerge-checks/blob/master/docs/clang_tidy.md#review-comments."
-                         "#review-comments.", "why?")
+        annotate(f'clang-tidy found {errors_count} errors and {warn_count} warnings. {inline_comments} of them were '
-        step.messages.append(
+                 f'added as review comments [why?]({url})', style='error')
-            f'clang-tidy found {errors_count} errors and {warn_count} warnings. {inline_comments} of them are added '
-            f'as review comments {url}')
     logging.debug(f'report: {report}')
     logging.debug(f'step: {step}')
 

scripts/phabtalk/phabtalk.py

@@ -17,8 +17,6 @@ Interactions with Phabricator.
 """
 
 import logging
-import socket
-import time
 from typing import Optional, List, Dict
 import uuid
 
@@ -32,9 +30,9 @@ class PhabTalk:
     """
 
     def __init__(self, token: Optional[str], host: Optional[str] = 'https://reviews.llvm.org/api/',
-                 dryrun: bool = False):
+                 dry_run_updates: bool = False):
         self._phab = None  # type: Optional[Phabricator]
-        if not dryrun:
+        self.dry_run_updates = dry_run_updates
         self._phab = Phabricator(token=token, host=host)
         self.update_interfaces()
 
@@ -42,23 +40,16 @@ class PhabTalk:
     def update_interfaces(self):
         self._phab.update_interfaces()
 
-    @property
-    def dryrun(self):
-        return self._phab is None
-
     @backoff.on_exception(backoff.expo, Exception, max_tries=5, logger='', factor=3)
     def get_revision_id(self, diff: str) -> Optional[str]:
         """Get the revision ID for a diff from Phabricator."""
-        if self.dryrun:
-            return None
-
         result = self._phab.differential.querydiffs(ids=[diff])
         return 'D' + result[diff]['revisionID']
 
     def comment_on_diff(self, diff_id: str, text: str):
         """Add a comment to a differential based on the diff_id"""
-        print('Sending comment to diff {}:'.format(diff_id))
+        logging.info('Sending comment to diff {}:'.format(diff_id))
-        print(text)
+        logging.info(text)
         revision_id = self.get_revision_id(diff_id)
         if revision_id is not None:
             self._comment_on_revision(revision_id, text)
@@ -72,16 +63,16 @@ class PhabTalk:
             'value': text
         }]
 
-        if self.dryrun:
+        if self.dry_run_updates:
-            print('differential.revision.edit =================')
+            logging.info('differential.revision.edit =================')
-            print('Transactions: {}'.format(transactions))
+            logging.info('Transactions: {}'.format(transactions))
             return
 
         # API details at
         # https://secure.phabricator.com/conduit/method/differential.revision.edit/
         self._phab.differential.revision.edit(objectIdentifier=revision,
                                               transactions=transactions)
-        print('Uploaded comment to Revision D{}:{}'.format(revision, text))
+        logging.info('Uploaded comment to Revision D{}:{}'.format(revision, text))
 
     @backoff.on_exception(backoff.expo, Exception, max_tries=5, logger='', factor=3)
     def update_build_status(self, phid: str, working: bool, success: bool, lint: {}, unit: []):
@@ -114,11 +105,11 @@ class PhabTalk:
             }
             lint_messages.append(lint_message)
 
-        if self.dryrun:
+        if self.dry_run_updates:
-            print('harbormaster.sendmessage =================')
+            logging.info('harbormaster.sendmessage =================')
-            print('type: {}'.format(result_type))
+            logging.info('type: {}'.format(result_type))
-            print('unit: {}'.format(unit))
+            logging.info('unit: {}'.format(unit))
-            print('lint: {}'.format(lint_messages))
+            logging.info('lint: {}'.format(lint_messages))
             return
 
         self._phab.harbormaster.sendmessage(
@@ -126,16 +117,16 @@ class PhabTalk:
             type=result_type,
             unit=unit,
             lint=lint_messages)
-        print('Uploaded build status {}, {} test results and {} lint results'.format(
+        logging.info('Uploaded build status {}, {} test results and {} lint results'.format(
             result_type, len(unit), len(lint_messages)))
 
     @backoff.on_exception(backoff.expo, Exception, max_tries=5, logger='', factor=3)
     def create_artifact(self, phid, artifact_key, artifact_type, artifact_data):
-        if self.dryrun:
+        if self.dry_run_updates:
-            print('harbormaster.createartifact =================')
+            logging.info('harbormaster.createartifact =================')
-            print('artifactKey: {}'.format(artifact_key))
+            logging.info('artifactKey: {}'.format(artifact_key))
-            print('artifactType: {}'.format(artifact_type))
+            logging.info('artifactType: {}'.format(artifact_type))
-            print('artifactData: {}'.format(artifact_data))
+            logging.info('artifactData: {}'.format(artifact_data))
             return
         self._phab.harbormaster.createartifact(
             buildTargetPHID=phid,
@@ -144,6 +135,9 @@ class PhabTalk:
             artifactData=artifact_data)
 
     def maybe_add_url_artifact(self, phid: str, url: str, name: str):
+        if self.dry_run_updates:
+            logging.info(f'add ULR artifact "{name}" {url}')
+            return
         if phid is None:
             logging.warning('PHID is not provided, cannot create URL artifact')
             return
@@ -155,7 +149,6 @@ class Step:
         self.name = ''
         self.success = True
         self.duration = 0.0
-        self.messages = []
         self.reproduce_commands = []
 
     def set_status_from_exit_code(self, exit_code: int):

scripts/pipeline_create_branch.py

@@ -30,7 +30,7 @@ if __name__ == '__main__':
         'label': 'create branch',
         'key': 'create-branch',
         'commands': [
-            'pip install -r scripts/requirements.txt',
+            'pip install -q -r scripts/requirements.txt',
             'scripts/apply_patch.sh'
         ],
         'agents': {'queue': 'linux'},

scripts/pipeline_master.py

@@ -25,7 +25,6 @@ steps_generators = [
 if __name__ == '__main__':
     scripts_refspec = os.getenv("ph_scripts_refspec", "master")
     no_cache = os.getenv('ph_no_cache') is not None
-    filter_output = '--filter-output' if os.getenv('ph_no_filter_output') is None else ''
     projects = os.getenv('ph_projects', 'clang;clang-tools-extra;libc;libcxx;libcxxabi;lld;libunwind;mlir;openmp;polly')
     log_level = os.getenv('ph_log_level', 'WARNING')
     notify_emails = list(filter(None, os.getenv('ph_notify_emails', '').split(',')))

scripts/pipeline_premerge.py

@@ -16,9 +16,9 @@
 import logging
 import os
 
+from buildkite_utils import annotate, feedback_url
 from choose_projects import ChooseProjects
 import git
-from phabtalk.phabtalk import PhabTalk
 from steps import generic_linux, generic_windows, from_shell_output, checkout_scripts
 import yaml
 
@@ -30,15 +30,15 @@ if __name__ == '__main__':
     scripts_refspec = os.getenv("ph_scripts_refspec", "master")
     diff_id = os.getenv("ph_buildable_diff", "")
     no_cache = os.getenv('ph_no_cache') is not None
-    filter_output = '--filter-output' if os.getenv('ph_no_filter_output') is None else ''
     projects = os.getenv('ph_projects', 'detect')
     log_level = os.getenv('ph_log_level', 'INFO')
     logging.basicConfig(level=log_level, format='%(levelname)-7s %(message)s')
 
     phid = os.getenv('ph_target_phid')
-    # Add link in review to the build.
+    url = f"https://reviews.llvm.org/D{os.getenv('ph_buildable_revision')}?id={diff_id}"
-    if phid is not None:
+    annotate(f"Build for [D{os.getenv('ph_buildable_revision')}#{diff_id}]({url}). "
-        PhabTalk(os.getenv('CONDUIT_TOKEN')).maybe_add_url_artifact(phid, os.getenv('BUILDKITE_BUILD_URL'), 'buildkite')
+             f"[Harbormaster build](https://reviews.llvm.org/harbormaster/build/{os.getenv('ph_build_id')}).\n"
+             f"If there is a build infrastructure issue, please [create a bug]({feedback_url()}).")
 
     # List all affected projects.
     repo = git.Repo('.')
@@ -55,17 +55,19 @@ if __name__ == '__main__':
     steps = []
     projects = cp.add_dependencies(affected_projects)
     logging.info(f'projects with dependencies: {projects}')
+    # Add generic Linux checks.
     excluded_linux = cp.get_excluded('linux')
     logging.info(f'excluded for linux: {excluded_linux}')
     linux_projects = projects - excluded_linux
     if len(linux_projects) > 0:
         steps.extend(generic_linux(';'.join(sorted(linux_projects)), True))
+    # Add generic Windows steps.
     excluded_windows = cp.get_excluded('windows')
     logging.info(f'excluded for windows: {excluded_windows}')
     windows_projects = projects - excluded_windows
     if len(windows_projects) > 0:
         steps.extend(generic_windows(';'.join(sorted(windows_projects))))
-
+    # Add custom checks.
     for gen in steps_generators:
         steps.extend(from_shell_output(gen))
 
@@ -78,7 +80,7 @@ if __name__ == '__main__':
         })
 
         report_step = {
-            'label': ':spiral_note_pad: report',
+            'label': ':phabricator: update build status on Phabricator',
             'commands': [
                 *checkout_scripts('linux', scripts_refspec),
                 '${SRC}/scripts/summary.py',

scripts/premerge_checks.py

@@ -29,53 +29,30 @@ from typing import Callable
 import clang_format_report
 import clang_tidy_report
 import run_cmake
-import test_results_report
+from buildkite_utils import upload_file, annotate, strip_emojis
-from buildkite_utils import upload_file
 from exec_utils import watch_shell, if_not_matches, tee
 from phabtalk.phabtalk import Report, PhabTalk, Step
 
 
-def ninja_all_report(step: Step, _: Report, filter_output: bool):
+def ninja_all_report(step: Step, _: Report):
-    print('Full log will be available in Artifacts "ninja-all.log"', flush=True)
     step.reproduce_commands.append('ninja all')
-    with open(f'{artifacts_dir}/ninja-all.log', 'wb') as f:
-        w = sys.stdout.buffer.write
-        if filter_output:
-            r = re.compile(r'^\[.*] (Building|Linking|Linting|Copying|Generating|Creating)')
-            w = partial(if_not_matches, write=sys.stdout.buffer.write, regexp=r)
     rc = watch_shell(
-            partial(tee, write1=w, write2=f.write),
+        sys.stdout.buffer.write,
-            partial(tee, write1=sys.stderr.buffer.write, write2=f.write),
+        sys.stderr.buffer.write,
         'ninja all', cwd=build_dir)
     logging.debug(f'ninja all: returned {rc}')
     step.set_status_from_exit_code(rc)
-        if not step.success:
-            report.add_artifact(artifacts_dir, 'ninja-all.log', 'build failed')
 
 
-def ninja_check_all_report(step: Step, _: Report, filter_output: bool):
+def ninja_check_all_report(step: Step, _: Report):
     print('Full log will be available in Artifacts "ninja-check-all.log"', flush=True)
     step.reproduce_commands.append('ninja check-all')
-    with open(f'{artifacts_dir}/ninja-check-all.log', 'wb') as f:
-        w = sys.stdout.buffer.write
-        if filter_output:
-            r = re.compile(r'^(\[.*] (Building|Linking|Generating)|(PASS|XFAIL|UNSUPPORTED):)')
-            w = partial(if_not_matches, write=sys.stdout.buffer.write, regexp=r)
     rc = watch_shell(
-            partial(tee, write1=w, write2=f.write),
+        sys.stdout.buffer.write,
-            partial(tee, write1=sys.stderr.buffer.write, write2=f.write),
+        sys.stderr.buffer.write,
         'ninja check-all', cwd=build_dir)
     logging.debug(f'ninja check-all: returned {rc}')
     step.set_status_from_exit_code(rc)
-    test_results_report.run(build_dir, 'test-results.xml', step, report)
-    if not step.success:
-        message = 'tests failed'
-        f = report.test_stats['fail']
-        if f == 1:
-            message = '1 test failed'
-        if f > 1:
-            message = f'{f} tests failed'
-        report.add_artifact(artifacts_dir, 'ninja-check-all.log', message)
 
 
 def run_step(name: str, report: Report, thunk: Callable[[Step, Report], None]) -> Step:
@@ -85,9 +62,13 @@ def run_step(name: str, report: Report, thunk: Callable[[Step, Report], None]) -
     step.name = name
     thunk(step, report)
     step.duration = time.time() - start
-    # Expand section if it failed.
+    # Expand section if step has failed.
     if not step.success:
         print('^^^ +++', flush=True)
+    if step.success:
+        annotate(f"{name}: OK")
+    else:
+        annotate(f"{name}: FAILED", style='error')
     report.steps.append(step)
     return step
 
@@ -114,13 +95,15 @@ if __name__ == '__main__':
     parser.add_argument('--log-level', type=str, default='WARNING')
     parser.add_argument('--check-clang-format', action='store_true')
     parser.add_argument('--check-clang-tidy', action='store_true')
-    parser.add_argument('--filter-output', action='store_true')
     parser.add_argument('--projects', type=str, default='detect',
                         help="Projects to select, either a list or projects like 'clang;libc', or "
                              "'detect' to automatically infer proejcts from the diff, or "
                              "'default' to add all enabled projects")
     args = parser.parse_args()
     logging.basicConfig(level=args.log_level, format='%(levelname)-7s %(message)s')
+
+    ctx = strip_emojis(os.getenv('BUILDKITE_LABEL', 'default'))
+    annotate(os.getenv('BUILDKITE_LABEL', 'default'), context=ctx)
     build_dir = ''
     step_key = os.getenv("BUILDKITE_STEP_KEY")
     scripts_dir = pathlib.Path(__file__).parent.absolute()
@@ -130,16 +113,13 @@ if __name__ == '__main__':
     report = Report()
     report.os = f'{os.getenv("BUILDKITE_AGENT_META_DATA_OS")}'
     report.name = step_key
-    report.success = False
-    # Create report with failure in case something below fails.
-    with open(report_path, 'w') as f:
-        json.dump(report.__dict__, f, default=as_dict)
     report.success = True
+
     cmake = run_step('cmake', report, lambda s, r: cmake_report(args.projects, s, r))
     if cmake.success:
-        ninja_all = run_step('ninja all', report, partial(ninja_all_report, filter_output=args.filter_output))
+        ninja_all = run_step('ninja all', report, ninja_all_report)
         if ninja_all.success:
-            run_step('ninja check-all', report, partial(ninja_check_all_report, filter_output=args.filter_output))
+            run_step('ninja check-all', report, ninja_check_all_report)
         if args.check_clang_tidy:
             run_step('clang-tidy', report,
                      lambda s, r: clang_tidy_report.run('HEAD~1', os.path.join(scripts_dir, 'clang-tidy.ignore'), s, r))
@@ -147,34 +127,25 @@ if __name__ == '__main__':
         run_step('clang-format', report,
                  lambda s, r: clang_format_report.run('HEAD~1', os.path.join(scripts_dir, 'clang-format.ignore'), s, r))
     logging.debug(report)
-    print('+++ Summary', flush=True)
+    summary = []
-    for s in report.steps:
+    summary.append('''
-        mark = 'OK   '
+<details>
-        if not s.success:
+  <summary>Reproduce build locally</summary>
-            report.success = False
+
-            mark = 'FAIL '
+```''')
-        msg = ''
+    summary.append(f'git clone {os.getenv("BUILDKITE_REPO")} llvm-project')
-        if len(s.messages):
+    summary.append('cd llvm-project')
-            msg = ': ' + '\n  '.join(s.messages)
+    summary.append(f'git checkout {os.getenv("BUILDKITE_COMMIT")}')
-        print(f'{mark} {s.name}{msg}', flush=True)
-    print('--- Reproduce build locally', flush=True)
-    print(f'git clone {os.getenv("BUILDKITE_REPO")} llvm-project')
-    print('cd llvm-project')
-    print(f'git checkout {os.getenv("BUILDKITE_COMMIT")}')
     for s in report.steps:
         if len(s.reproduce_commands) == 0:
             continue
-        print('\n'.join(s.reproduce_commands), flush=True)
+        summary.append('\n'.join(s.reproduce_commands))
-    print('', flush=True)
+    summary.append('```\n</details>')
-    if not report.success:
+    annotate('\n'.join(summary), style='success')
-        print('^^^ +++', flush=True)
-
     ph_target_phid = os.getenv('ph_target_phid')
     if ph_target_phid is not None:
-        phabtalk = PhabTalk(os.getenv('CONDUIT_TOKEN'))
+        phabtalk = PhabTalk(os.getenv('CONDUIT_TOKEN'), dry_run_updates=(os.getenv('ph_dry_run_report') is not None))
-        for u in report.unit:
+        phabtalk.update_build_status(ph_target_phid, True, report.success, report.lint, [])
-            u['engine'] = step_key
-        phabtalk.update_build_status(ph_target_phid, True, report.success, report.lint, report.unit)
         for a in report.artifacts:
             url = upload_file(a['dir'], a['file'])
             if url is not None:

scripts/requirements.txt

@@ -7,3 +7,4 @@ pyaml==20.4.0
 requests==2.24.0
 retrying==1.3.3
 unidiff==0.6.0
+python-benedict==0.22.0

scripts/set_build_status.py

@@ -27,7 +27,7 @@ if __name__ == '__main__':
     args = parser.parse_args()
     logging.basicConfig(level=args.log_level, format='%(levelname)-7s %(message)s')
 
-    phabtalk = PhabTalk(os.getenv('CONDUIT_TOKEN'))
+    phabtalk = PhabTalk(os.getenv('CONDUIT_TOKEN'), dry_run_updates=(os.getenv('ph_dry_run_report') is not None))
     ph_target_phid = os.getenv('ph_target_phid')
     if ph_target_phid is None:
         logging.warning('ph_target_phid is not specified. Will not update the build status in Phabricator')

scripts/steps.py

@@ -28,7 +28,6 @@ def generic_linux(projects: str, check_diff: bool) -> List:
         return []
     scripts_refspec = os.getenv("ph_scripts_refspec", "master")
     no_cache = os.getenv('ph_no_cache') is not None
-    filter_output = '--filter-output' if os.getenv('ph_no_filter_output') is None else ''
     log_level = os.getenv('ph_log_level', 'WARNING')
     linux_agents = {'queue': 'linux'}
     t = os.getenv('ph_linux_agents')
@@ -48,7 +47,7 @@ def generic_linux(projects: str, check_diff: bool) -> List:
     if check_diff:
         commands.extend([
             '${SRC}/scripts/premerge_checks.py --check-clang-format --check-clang-tidy '
-            f'--projects="{projects}" --log-level={log_level} {filter_output}',
+            f'--projects="{projects}" --log-level={log_level}',
         ])
     else:
         commands.extend([
@@ -63,10 +62,10 @@ def generic_linux(projects: str, check_diff: bool) -> List:
     ])
 
     linux_buld_step = {
-        'label': ':linux: build and test linux',
+        'label': ':linux: x64 debian',
         'key': 'linux',
         'commands': commands,
-        'artifact_paths': ['artifacts/**/*', '*_result.json'],
+        'artifact_paths': ['artifacts/**/*', '*_result.json', 'build/test-results.xml'],
         'agents': linux_agents,
         'timeout_in_minutes': 120,
         'retry': {'automatic': [
@@ -83,7 +82,6 @@ def generic_windows(projects: str) -> List:
     scripts_refspec = os.getenv("ph_scripts_refspec", "master")
     no_cache = os.getenv('ph_no_cache') is not None
     log_level = os.getenv('ph_log_level', 'WARNING')
-    filter_output = '--filter-output' if os.getenv('ph_no_filter_output') is None else ''
     clear_sccache = 'powershell -command "sccache --stop-server; echo \\$env:SCCACHE_DIR; ' \
                     'Remove-Item -Recurse -Force -ErrorAction Ignore \\$env:SCCACHE_DIR; ' \
                     'sccache --start-server"'
@@ -92,7 +90,7 @@ def generic_windows(projects: str) -> List:
     if t is not None:
         win_agents = json.loads(t)
     windows_buld_step = {
-        'label': ':windows: build and test windows',
+        'label': ':windows: x64 windows',
         'key': 'windows',
         'commands': [
             clear_sccache if no_cache else '',
@@ -100,7 +98,7 @@ def generic_windows(projects: str) -> List:
             *checkout_scripts('windows', scripts_refspec),
 
             'powershell -command "'
-            f'%SRC%/scripts/premerge_checks.py --projects=\'{projects}\' --log-level={log_level} {filter_output}; '
+            f'%SRC%/scripts/premerge_checks.py --projects=\'{projects}\' --log-level={log_level}; '
             '\\$exit=\\$?;'
             'sccache --show-stats;'
             'if (\\$exit) {'
@@ -111,7 +109,7 @@ def generic_windows(projects: str) -> List:
             '  exit 1;'
             '}"',
         ],
-        'artifact_paths': ['artifacts/**/*', '*_result.json'],
+        'artifact_paths': ['artifacts/**/*', '*_result.json', 'build/test-results.xml'],
         'agents': win_agents,
         'timeout_in_minutes': 90,
         'retry': {'automatic': [
@@ -164,7 +162,7 @@ def checkout_scripts(target_os: str, scripts_refspec: str) -> []:
             'git checkout x',
             'echo llvm-premerge-checks commit:',
             'git rev-parse HEAD',
-            'pip install -r %SRC%/scripts/requirements.txt',
+            'pip install -q -r %SRC%/scripts/requirements.txt',
             'cd %BUILDKITE_BUILD_CHECKOUT_PATH%',
         ]
     return [
@@ -176,6 +174,6 @@ def checkout_scripts(target_os: str, scripts_refspec: str) -> []:
         'git checkout x',
         'echo "llvm-premerge-checks commit"',
         'git rev-parse HEAD',
-        'pip install -r ${SRC}/scripts/requirements.txt',
+        'pip install -q -r ${SRC}/scripts/requirements.txt',
         'cd "$BUILDKITE_BUILD_CHECKOUT_PATH"',
     ]

scripts/summary.py

@@ -18,44 +18,90 @@ import logging
 import os
 
 from phabtalk.phabtalk import PhabTalk
-from buildkite_utils import format_url, BuildkiteApi
+from buildkite_utils import format_url, BuildkiteApi, strip_emojis
+import test_results_report
+from benedict import benedict
+
+
+def get_failed_jobs(build: benedict) -> []:
+    failed_jobs = []
+    for j in build.get('jobs', []):
+        j = benedict(j)
+        if j.get('state') == 'failed' and j.get('name'):
+            failed_jobs.append(j.get('name'))
+    return failed_jobs
+
+
+def process_unit_test_reports(bk: BuildkiteApi, build: benedict, prefix: str) -> []:
+    failed_tests = []
+    for job in build.get('jobs', []):
+        job = benedict(job)
+        if job.get('state') != 'failed' or job.get('type') != 'script':
+            # Job must run scripts and fail to be considered.
+            # Recursive pipeline triggers are not processed at the moment.
+            continue
+        artifacts_url = job.get('artifacts_url')
+        if artifacts_url is None:
+            continue
+        artifacts = bk.get(artifacts_url).json()
+        for a in artifacts:
+            a = benedict(a)
+            if not a.get('filename').endswith('test-results.xml') or not a.get('download_url'):
+                continue
+            content = bk.get(a.get('download_url')).content
+            ctx = strip_emojis(prefix + ' ' + job.get('name', build.get('pipeline.name')))
+            failed_tests.extend(test_results_report.parse_failures(content, ctx))
+    return failed_tests
+
 
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     parser.add_argument('--log-level', type=str, default='INFO')
     args = parser.parse_args()
     logging.basicConfig(level=args.log_level, format='%(levelname)-7s %(message)s')
-
-    print(f'Branch {os.getenv("BUILDKITE_BRANCH")} at {os.getenv("BUILDKITE_REPO")}', flush=True)
     ph_buildable_diff = os.getenv('ph_buildable_diff')
-    if ph_buildable_diff is not None:
-        url = f'https://reviews.llvm.org/D{os.getenv("ph_buildable_revision")}?id={ph_buildable_diff}'
-        print(f'Review: {format_url(url)}', flush=True)
-    if os.getenv('BUILDKITE_TRIGGERED_FROM_BUILD_NUMBER') is not None:
-        url = f'https://buildkite.com/llvm-project/' \
-              f'{os.getenv("BUILDKITE_TRIGGERED_FROM_BUILD_PIPELINE_SLUG")}/' \
-              f'builds/{os.getenv("BUILDKITE_TRIGGERED_FROM_BUILD_NUMBER")}'
-        print(f'Triggered from build {format_url(url)}', flush=True)
     ph_target_phid = os.getenv('ph_target_phid')
     if ph_target_phid is None:
         logging.warning('ph_target_phid is not specified. Will not update the build status in Phabricator')
         exit(0)
-
+    phabtalk = PhabTalk(os.getenv('CONDUIT_TOKEN'), dry_run_updates=(os.getenv('ph_dry_run_report') is not None))
+    report_success = False  # for try block
+    failed_tests = []
+    try:
         bk = BuildkiteApi(os.getenv("BUILDKITE_API_TOKEN"), os.getenv("BUILDKITE_ORGANIZATION_SLUG"))
         build = bk.get_build(os.getenv("BUILDKITE_PIPELINE_SLUG"), os.getenv("BUILDKITE_BUILD_NUMBER"))
         success = True
-    build.setdefault('jobs', [])
+        failed_tests = process_unit_test_reports(bk, build, '')
-    for j in build['jobs']:
+        for i, job in enumerate(build.get('jobs', [])):
-        j.setdefault('state', '')
+            job = benedict(job)
-        j.setdefault('id', '')
+            job_web_url = job.get('web_url', os.getenv('BUILDKITE_BUILD_URL', ''))
-        logging.info(f'{j["id"]} state {j["state"]}')
+            logging.info(f'{job.get("id")} state {job.get("state")}')
-        success = success and (j['state'] != 'failed')
+            job_state = job.get('state')
-
+            if job.get('type') == 'waiter':
-    phabtalk = PhabTalk(os.getenv('CONDUIT_TOKEN'))
+                continue
+            if job_state != 'passed' and job_state != 'failed':
+                # Current and irrelevant steps.
+                continue
+            if job_state == 'passed' and i == 0:
+                # Skip successful first step as we assume it to be a pipeline setup
+                continue
+            name = job.get('name')
+            if job.get('type') == 'trigger':
+                job_web_url = job.get('triggered_build.web_url', job_web_url)
+                triggered_url = job.get('triggered_build.url')
+                if triggered_url != '':
+                    sub_build = benedict(bk.get(triggered_url).json())
+                    name = name or sub_build.get('pipeline.name')
+                    failed_steps = get_failed_jobs(sub_build)
+                    failed_tests.extend(process_unit_test_reports(bk, sub_build, name))
+                    if job_state == 'failed' and failed_steps:
+                        name = f"{name} ({', '.join(failed_steps[:2])}{', ...' if len(failed_steps) > 2 else ''})"
+            name = strip_emojis(name) or 'unknown'
+            phabtalk.maybe_add_url_artifact(ph_target_phid, job_web_url, f"{name} {job_state}")
+            if job_state == 'failed':
+                success = False
+        report_success = success  # Must be last before finally: block to report errors in this script.
+    finally:
         build_url = f'https://reviews.llvm.org/harbormaster/build/{os.getenv("ph_build_id")}'
         print(f'Reporting results to Phabricator build {format_url(build_url)}', flush=True)
-    phabtalk.update_build_status(ph_target_phid, False, success, {}, [])
+        phabtalk.update_build_status(ph_target_phid, False, report_success, {}, failed_tests)
-    bug_url = f'https://github.com/google/llvm-premerge-checks/issues/new?assignees=&labels=bug' \
-              f'&template=bug_report.md&title=buildkite build {os.getenv("BUILDKITE_PIPELINE_SLUG")} ' \
-              f'{os.getenv("BUILDKITE_BUILD_NUMBER")}'
-    print(f'{format_url(bug_url, "report issue")}', flush=True)

scripts/test_results_report.py

@@ -28,9 +28,8 @@ def run(working_dir: str, test_results: str, step: Optional[Step], report: Optio
         step = Step()
     path = os.path.join(working_dir, test_results)
     if not os.path.exists(path):
-        logging.warning(f'{path} is not found')
+        logging.error(f'{path} is not found')
         step.success = False
-        step.messages.append(f'test report "{path}" is not found')
         return
     try:
         success = True
@@ -63,13 +62,30 @@ def run(working_dir: str, test_results: str, step: Optional[Step], report: Optio
                     msg += f'{test_case["namespace"]}/{test_case["name"]}\n'
     except Exception as e:
         logging.error(e)
-        step.messages.append('Parsing of test results failed')
         step.success = False
 
     logging.debug(f'report: {report}')
     logging.debug(f'step: {step}')
 
 
+def parse_failures(test_xml: bytes, context: str) -> []:
+    failed_cases = []
+    root_node = etree.fromstring(test_xml)
+    for test_case in root_node.xpath('//testcase'):
+        failure = test_case.find('failure')
+        if failure is None:
+            continue
+        failed_cases.append({
+            'engine': context,
+            'name': test_case.attrib['name'],
+            'namespace': test_case.attrib['classname'],
+            'result': 'fail',
+            'duration': float(test_case.attrib['time']),
+            'details': failure.text
+        })
+    return failed_cases
+
+
 if __name__ == '__main__':
     parser = argparse.ArgumentParser(description='Processes results from xml report')
     parser.add_argument('test-report', default='build/test-results.xml')