[CI] sync config from odoo-elabore-ci

New version : 16.0.1.0.3

_mark_done() only set partner_id/email when it created a *new* res.partner.
If the answer already had a partner_id/email (e.g. inherited from the Odoo
user logged in when the /survey/start link was opened), that value was kept
even though the record creation matched or created a different, correct
contact from the participant's own answers.

Now we always sync partner_id/email to the res.partner actually matched or created
from the participant's answers.

Add a migration to relink partner_id/email on existing done submissions.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

.gitea/scripts/odoo-tests.sh

@@ -0,0 +1,42 @@
+#!/bin/bash
+#
+# Run the Python tests of the given Odoo modules.
+#
+# This script is meant to be executed *inside* an Elabore doodba image
+# (docker.0k.io/mirror/odoo:rc_<version>-ELABORE-LIGHT), with a
+# PostgreSQL service reachable at $PGHOST. The modules under test are
+# read from the current working directory (the checked-out repository),
+# which is prepended to the addons path so that the PR's code takes
+# precedence over the modules already baked in the image.
+#
+# Usage: odoo-tests.sh <module1,module2,...>
+
+set -euo pipefail
+
+modules="${1:?usage: odoo-tests.sh <module1,module2,...>}"
+
+export PGHOST=postgres
+export PGUSER=odoo
+export PGPASSWORD=odoopassword
+export PGDATABASE="test_$(basename "$PWD" | tr '.-' '__')"
+export WITHOUT_DEMO=all
+export DB_FILTER='.*'
+
+# Generate /opt/odoo/auto/odoo.conf from the environment and wait for
+# the database to accept connections.
+/opt/odoo/common/entrypoint
+
+# Odoo test-tags format: "/module1,/module2" (one leading slash).
+tags="/$(echo "$modules" | tr ',' ',/')"
+
+exec odoo \
+    --config=/opt/odoo/auto/odoo.conf \
+    --max-cron-threads=0 \
+    --limit-time-real=10000 \
+    --workers=0 \
+    --test-enable \
+    --stop-after-init \
+    --addons-path="$PWD,/opt/odoo/auto/addons" \
+    -d "$PGDATABASE" \
+    -i "$modules" \
+    --test-tags "$tags"

.gitea/workflows/pre-commit.yml

@@ -2,8 +2,6 @@ name: pre-commit
 
 on:
   pull_request:
-    branches:
-      - "16.0*"
 
 jobs:
   pre-commit:
@@ -15,10 +13,10 @@ jobs:
           python-version: "3.11"
       - name: Get python version
         run: echo "PY=$(python -VV | sha256sum | cut -d' ' -f1)" >> $GITHUB_ENV
-      # - uses: actions/cache@v4
-      #   with:
-      #     path: ~/.cache/pre-commit
-      #     key: pre-commit|${{ env.PY }}|${{ hashFiles('.pre-commit-config.yaml') }}
+      - uses: https://gitea.com/actions/cache@v3
+        with:
+          path: ~/.cache/pre-commit
+          key: pre-commit|${{ env.PY }}|${{ hashFiles('.pre-commit-config.yaml') }}
       - name: Install pre-commit
         run: pip install pre-commit
       - name: Run pre-commit

.gitea/workflows/test.yml

@@ -0,0 +1,79 @@
+name: tests
+
+on:
+  pull_request:
+    branches:
+      - "16.0*"
+  workflow_dispatch:
+
+jobs:
+  test:
+    runs-on: odoo-ci
+    container:
+      image: docker.0k.io/mirror/odoo:rc_16.0-ELABORE-LIGHT
+    services:
+      postgres:
+        image: docker.0k.io/postgres:17.2.0-myc
+        env:
+          POSTGRES_USER: odoo
+          POSTGRES_PASSWORD: odoopassword
+          POSTGRES_DB: postgres
+    steps:
+      # The doodba image ships no Node.js, so actions/checkout@v4 cannot
+      # run here. The repository is private, so the job token is required.
+      - name: Checkout
+        env:
+          SECRET_GITEA_TOKEN: ${{ secrets.GITEA_TOKEN }}
+          SECRET_GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          set -eu
+          token="${GITEA_TOKEN:-${GITHUB_TOKEN:-${SECRET_GITEA_TOKEN:-${SECRET_GITHUB_TOKEN:-}}}}"
+          if [ -z "$token" ]; then
+            echo "::error::No job token available to clone the repository"
+            exit 1
+          fi
+          server="${GITHUB_SERVER_URL:-https://git.elabore.coop}"
+          repo="${GITHUB_REPOSITORY:?GITHUB_REPOSITORY is not set}"
+          ref="${GITHUB_REF:-${GITHUB_SHA:?neither GITHUB_REF nor GITHUB_SHA is set}}"
+          auth_url="$(printf '%s' "${server%/}/${repo}.git" |
+            sed -E "s#^https://#https://${token}@#")"
+          git init -q .
+          git config --global --add safe.directory "$PWD"
+          git remote add origin "$auth_url"
+          git fetch -q --depth=1 origin "$ref"
+          git checkout -q FETCH_HEAD
+
+      - name: Detect changed modules
+        id: changed
+        run: |
+          set -eu
+          if [ -n "${GITHUB_BASE_REF:-}" ]; then
+            git fetch -q --depth=1 origin "refs/heads/$GITHUB_BASE_REF"
+            base_rev="FETCH_HEAD"
+            scope="diff against $GITHUB_BASE_REF"
+          else
+            base_rev=""
+            scope="all installable modules (manual run)"
+          fi
+          if [ -n "$base_rev" ]; then
+            modules="$(
+              git diff --name-only "$base_rev" HEAD |
+                awk -F/ '{print $1}' | sort -u |
+                while read -r dir; do
+                  [ -f "$dir/__manifest__.py" ] && echo "$dir"
+                done | paste -sd, -
+            )"
+          else
+            modules="$(
+              for dir in */; do
+                dir="${dir%/}"
+                [ -f "$dir/__manifest__.py" ] && echo "$dir"
+              done | paste -sd, -
+            )"
+          fi
+          echo "modules=$modules" >> "$GITHUB_OUTPUT"
+          echo "Modules to test ($scope): ${modules:-<none>}"
+
+      - name: Run Odoo tests
+        if: steps.changed.outputs.modules != ''
+        run: bash .gitea/scripts/odoo-tests.sh "${{ steps.changed.outputs.modules }}"

.gitignore

@@ -1,40 +1,42 @@
-*.docx
-
-.~*
-
 # Byte-compiled / optimized / DLL files
 __pycache__/
 *.py[cod]
-*$py.class
+/.venv
+/.pytest_cache
+/.ruff_cache
 
 # C extensions
 *.so
 
 # Distribution / packaging
 .Python
+env/
+bin/
 build/
 develop-eggs/
 dist/
-downloads/
 eggs/
-.eggs/
-lib/
 lib64/
 parts/
 sdist/
 var/
-wheels/
-share/python-wheels/
 *.egg-info/
 .installed.cfg
 *.egg
-MANIFEST
+*.eggs
 
-# PyInstaller
-#  Usually these files are written by a python script from a template
-#  before PyInstaller builds the exe, so as to inject date/other infos into it.
-*.manifest
-*.spec
+# Windows installers
+*.msi
+
+# Debian packages
+*.deb
+
+# Redhat packages
+*.rpm
+
+# MacOS packages
+*.dmg
+*.pkg
 
 # Installer logs
 pip-log.txt
@@ -43,123 +45,44 @@ pip-delete-this-directory.txt
 # Unit test / coverage reports
 htmlcov/
 .tox/
-.nox/
 .coverage
-.coverage.*
 .cache
 nosetests.xml
 coverage.xml
-*.cover
-*.py,cover
-.hypothesis/
-.pytest_cache/
-cover/
 
 # Translations
 *.mo
-*.pot
+
+# Pycharm
+.idea
+
+# Eclipse
+.settings
+
+# Visual Studio cache/options directory
+.vs/
+.vscode
+
+# OSX Files
+.DS_Store
 
 # Django stuff:
 *.log
-local_settings.py
-db.sqlite3
-db.sqlite3-journal
 
-# Flask stuff:
-instance/
-.webassets-cache
+# Mr Developer
+.mr.developer.cfg
+.project
+.pydevproject
 
-# Scrapy stuff:
-.scrapy
+# Rope
+.ropeproject
 
 # Sphinx documentation
 docs/_build/
 
-# PyBuilder
-.pybuilder/
-target/
-
-# Jupyter Notebook
-.ipynb_checkpoints
-
-# IPython
-profile_default/
-ipython_config.py
-
-# pyenv
-#   For a library or package, you might want to ignore these files since the code is
-#   intended to run in multiple environments; otherwise, check them in:
-# .python-version
-
-# pipenv
-#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
-#   However, in case of collaboration, if having platform-specific dependencies or dependencies
-#   having no cross-platform support, pipenv may install dependencies that don't work, or not
-#   install all needed dependencies.
-#Pipfile.lock
-
-# poetry
-#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
-#   This is especially recommended for binary packages to ensure reproducibility, and is more
-#   commonly ignored for libraries.
-#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
-#poetry.lock
-
-# pdm
-#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
-#pdm.lock
-#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
-#   in version control.
-#   https://pdm.fming.dev/#use-with-ide
-.pdm.toml
-
-# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
-__pypackages__/
-
-# Celery stuff
-celerybeat-schedule
-celerybeat.pid
-
-# SageMath parsed files
-*.sage.py
-
-# Environments
-.env
-.venv
-env/
-venv/
-ENV/
-env.bak/
-venv.bak/
-
-# Spyder project settings
-.spyderproject
-.spyproject
-
-# Rope project settings
-.ropeproject
-
-# mkdocs documentation
-/site
-
-# mypy
-.mypy_cache/
-.dmypy.json
-dmypy.json
-
-# Pyre type checker
-.pyre/
-
-# pytype static type analyzer
-.pytype/
-
-# Cython debug symbols
-cython_debug/
-
-# PyCharm
-#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
-#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
-#  and can be added to the global gitignore or merged into this file.  For a more nuclear
-#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
-#.idea/
+# Backup files
+*~
+*.swp
 
+# OCA rules
+!static/lib/

.pre-commit-config.yaml

@@ -49,12 +49,9 @@ repos:
           $(git rev-parse --show-toplevel))"'
       - id: oca-gen-addon-readme
         entry:
-          bash -c 'oca-gen-addon-readme 
-          --addons-dir=. 
-          --branch=$(git symbolic-ref
+          bash -c 'oca-gen-addon-readme --addons-dir=. --branch=$(git symbolic-ref
           refs/remotes/origin/HEAD | sed "s@^refs/remotes/origin/@@")
-          --repo-name=$(basename $(git rev-parse --show-toplevel))
-          --org-name="Elabore"
+          --repo-name=$(basename $(git rev-parse --show-toplevel)) --org-name="Elabore"
           --if-source-changed --keep-source-digest'
 
   - repo: https://github.com/OCA/odoo-pre-commit-hooks

survey_record_generation/__manifest__.py

@@ -11,7 +11,7 @@ Allow to create record of any model when sending the form :
 * Associate question with fields
 * For x2m fields : Associate values to questions
 """, 
-    "version": "16.0.1.0.2",
+    "version": "16.0.1.0.3",
     "license": "AGPL-3",
     "author": "Elabore",
     "website": "https://www.elabore.coop",

survey_record_generation/migrations/16.0.1.0.3/end-migration.py

@@ -0,0 +1,159 @@
+# License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+"""Relink survey.user_input.partner_id/email to the contact that actually
+matches each participant's own answers.
+
+Before this version, survey.user_input.partner_id/email were only updated
+when _mark_done() created a *new* res.partner. If the answer already had a
+partner_id/email (e.g. inherited from the Odoo user who was logged in when
+the /survey/start link was opened, or from some other unidentified cause),
+_mark_done() silently kept that value even when the survey_record_creation
+config found or created a different, correct contact from the participant's
+own answers. This backfill re-runs that resolution for every already-done
+answer and fixes partner_id/email accordingly.
+
+It never re-triggers the other survey.user_input._mark_done() overrides
+(crm lead / event registration / notification modules, ...): it only calls
+the res.partner-matching helpers directly, so it can't create duplicate
+leads, registrations, etc. for historical submissions.
+
+A res.partner is expected to already exist for every done submission (it
+was necessarily found or created the first time _mark_done() ran), so this
+backfill never creates one. It first looks for the survey.generated.record
+row _mark_done() logged when it created that partner, and falls back to
+re-deriving the match through find_existing_record()/
+find_duplicate_if_there_are_fields_with_unicity_check() when there is no
+such row (existing partner matched instead of created) or it points to a
+partner that has since been deleted (e.g. merged into another contact). If
+neither approach finds anything, the record is left untouched and logged
+for manual review instead.
+"""
+import logging
+
+from odoo import SUPERUSER_ID, api
+
+_logger = logging.getLogger(__name__)
+
+
+def migrate(cr, version):
+    env = api.Environment(cr, SUPERUSER_ID, {})
+    user_input_model = env["survey.user_input"]
+
+    surveys_with_partner_creation = env["survey.record.creation"].search(
+        [("model_id.model", "=", "res.partner")]
+    ).survey_id
+
+    user_inputs = user_input_model.search(
+        [
+            ("survey_id", "in", surveys_with_partner_creation.ids),
+            ("state", "=", "done"),
+        ]
+    )
+
+    _logger.info(
+        "survey_record_generation: relinking partner_id/email on %d done "
+        "survey.user_input records",
+        len(user_inputs),
+    )
+
+    fixed_count = 0
+    not_found_count = 0
+
+    for user_input in user_inputs:
+        record_creations = user_input.survey_id.survey_record_creation_ids.filtered(
+            lambda rc: rc.model_id.model == "res.partner"
+        ).sorted("sequence")
+
+        record_creation = record_creations[:1]
+        if not record_creation:
+            continue
+
+        # 1) Prefer the res.partner this very submission created, if any
+        # (the first one, by id, in the rare case there is more than one):
+        # it's the exact record _mark_done() produced for this answer, no
+        # guessing involved. This is also the only option for surveys whose
+        # record creation has neither update_existing_records nor any
+        # unicity_check field configured, since find_existing_record()/
+        # find_duplicate...() can then never find anything (nothing to
+        # search on).
+        record = False
+        for generated in user_input.generated_record_ids.sorted("id"):
+            if (
+                generated.survey_record_creation_id == record_creation
+                and generated.created_record_id
+                and generated.created_record_id._name == "res.partner"
+                and generated.created_record_id.exists()
+            ):
+                # The referenced partner may have since been deleted (e.g.
+                # merged into another contact): in that case it's not usable
+                # and we fall through to the search-based lookup below.
+                record = generated.created_record_id
+                break
+
+        # 2) Otherwise, this submission matched an already-existing partner
+        # instead of creating one (find_existing_record()/find_duplicate...()
+        # branch of _mark_done()): re-derive it the same way.
+        if not record:
+            # Only compute the fields find_existing_record()/find_duplicate...()
+            # actually read (the search field, and any unicity_check field),
+            # not every field of the record creation: other fields (e.g. a
+            # "record" reference to a model defined in a module that depends
+            # on this one) may not be loadable yet at this point of the
+            # upgrade, and are useless here anyway since this backfill never
+            # writes to res.partner.
+            needed_field_names = set()
+            if (
+                record_creation.update_existing_records
+                and record_creation.field_to_retrieve_existing_records
+            ):
+                needed_field_names.add(
+                    record_creation.field_to_retrieve_existing_records.name
+                )
+            unicity_field_values = record_creation.field_values_ids.filtered(
+                lambda field_value: field_value.unicity_check
+            )
+            needed_field_names.update(unicity_field_values.mapped("field_id.name"))
+
+            vals = {}
+            for field_value in record_creation.field_values_ids:
+                if field_value.field_id.name not in needed_field_names:
+                    continue
+                value, __ = user_input_model.get_value_based_on_value_origin(
+                    field_value=field_value,
+                    user_input=user_input,
+                    created_records={},
+                    model="res.partner",
+                    other_record_fields_to_update=[],
+                )
+                vals[field_value.field_id.name] = value
+
+            existing_record = user_input_model.find_existing_record(
+                record_creation, vals
+            )
+            duplicate = (
+                user_input_model.find_duplicate_if_there_are_fields_with_unicity_check(
+                    "res.partner", record_creation, vals
+                )
+            )
+            record = duplicate or existing_record
+
+        if not record:
+            _logger.warning(
+                "survey_record_generation: no existing res.partner found for "
+                "survey.user_input %s while backfilling partner_id/email, "
+                "leaving it untouched",
+                user_input.id,
+            )
+            not_found_count += 1
+            continue
+
+        if user_input.partner_id != record or user_input.email != record.email:
+            user_input.partner_id = record.id
+            user_input.email = record.email
+            fixed_count += 1
+
+    _logger.info(
+        "survey_record_generation: fixed %d survey.user_input records "
+        "(%d without a matching res.partner)",
+        fixed_count,
+        not_found_count,
+    )

survey_record_generation/models/survey_user_input.py

@@ -39,6 +39,7 @@ class SurveyUserInput(models.Model):
         for user_input in self:
             created_records = {}
             other_record_fields_to_update: list[SurveyRecordCreationFieldValues] = []
+            partner_linked_in_this_run = False
 
             record_creation: SurveyRecordCreation
             for (
@@ -85,8 +86,6 @@ class SurveyUserInput(models.Model):
                     try:
                         with self.env.cr.savepoint():
                             record = self.env[model].create(vals)
-                            if model == "res.partner" and not self.partner_id:
-                                self.partner_id = record.id
                     except Exception:
                         # This a broad exception because it could be IntegrityError,
                         # EmptyNamesError in case partner_firstname is installed etc...
@@ -103,6 +102,17 @@ class SurveyUserInput(models.Model):
                         }
                     )
 
+                if model == "res.partner" and not partner_linked_in_this_run:
+                    # Always reflect the partner actually matched/created from
+                    # this participant's own answers, even if partner_id/email
+                    # were already set on the answer (e.g. inherited from the
+                    # logged-in user when the survey link was opened). Only the
+                    # first res.partner record creation of this run wins, in
+                    # case several are configured on the same survey.
+                    user_input.partner_id = record.id
+                    user_input.email = record.email
+                    partner_linked_in_this_run = True
+
                 created_records[record_creation.id] = record
 
             # update linked record
@@ -315,28 +325,44 @@ class SurveyUserInput(models.Model):
                 }
             )
 
-    @staticmethod
-    def get_boolean_value(answer_value_char: str, question_title: str) -> bool:
-        # Below code is a trick to be able to use "simple_choice" question
-        # with values 'yes' and 'no' and transform it to boolean.
-        if boolean_value := answer_value_char in [
-            "1",
-            "True",
-            "true",
-            "Oui",
-            "oui",
-            "Yes",
-            "yes",
-        ]:
-            return boolean_value
-        else:
-            raise UserError(
-                _(
-                    "[Survey record generation] The boolean value %s(value)s "
-                    "is not supported (for question %(question)s)."
-                )
-                % {
-                    "value": answer_value_char,
-                    "question": question_title,
-                }
+
+    def _get_boolean_true_values(self):
+        """Tokens interpreted as true, normalized to lowercase.
+
+        Instance method so that a third-party module or a localization can
+        extend the list without rewriting get_boolean_value.
+        """
+        return {"1", "true", "vrai", "yes", "y", "oui", "o", "on", "x"}
+
+    def _get_boolean_false_values(self):
+        """Tokens interpreted as false, normalized to lowercase."""
+        return {"0", "false", "faux", "no", "n", "non", "off", ""}
+
+    def get_boolean_value(self, answer_value_char: str, question_title: str) -> bool:
+        """Convert the technical value of an answer into a boolean.
+
+        Allows a boolean field to be filled from a "simple_choice" question
+        whose suggested answers carry yes/no values.
+        """
+        if not answer_value_char:
+            # Empty answer: an unset boolean is false, this is not a
+            # configuration error.
+            return False
+
+        token = str(answer_value_char).strip().casefold()
+
+        if token in self._get_boolean_true_values():
+            return True
+        if token in self._get_boolean_false_values():
+            return False
+
+        raise UserError(
+            _(
+                "[Survey record generation] The boolean value %(value)s is not "
+                "supported (for question %(question)s)"
             )
+            % {
+                "value": answer_value_char,
+                "question": question_title,
+            }
+        )

survey_record_generation/tests/__init__.py

@@ -1 +1,2 @@
 from . import test_survey_record_creation
+from . import test_end_migration_16_0_1_0_3

survey_record_generation/tests/test_end_migration_16_0_1_0_3.py

@@ -0,0 +1,156 @@
+# License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+import importlib.util
+import os
+
+from odoo.addons.survey.tests.common import SurveyCase
+
+
+def _load_end_migration():
+    migration_path = os.path.join(
+        os.path.dirname(os.path.dirname(os.path.abspath(__file__))),
+        "migrations",
+        "16.0.1.0.3",
+        "end-migration.py",
+    )
+    spec = importlib.util.spec_from_file_location(
+        "survey_record_generation_end_migration_16_0_1_0_3", migration_path
+    )
+    module = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(module)
+    return module
+
+
+class TestEndMigration160103(SurveyCase):
+    """Covers the use cases the 16.0.1.0.3 end-migration backfill must
+    handle: relinking survey.user_input.partner_id/email to the contact
+    that actually matches each participant's own answers, for submissions
+    whose partner_id/email were left wrong before the fix in this version.
+    """
+
+    def setUp(self):
+        super().setUp()
+        self.migration = _load_end_migration()
+
+        self.survey = self.env["survey.survey"].create({"title": "Test Survey"})
+        self.question_name = self._add_question(
+            page=None,
+            name="Name",
+            qtype="char_box",
+            survey_id=self.survey.id,
+            sequence=1,
+        )
+        self.res_partner_model = self.env["ir.model"]._get("res.partner")
+        self.survey_record_creation = self.env["survey.record.creation"].create(
+            {
+                "name": "Contact",
+                "survey_id": self.survey.id,
+                "model_id": self.res_partner_model.id,
+            }
+        )
+        self.name_field = self.env["ir.model.fields"].search(
+            [("model", "=", "res.partner"), ("name", "=", "name")]
+        )
+        self.env["survey.record.creation.field.values"].create(
+            {
+                "survey_record_creation_id": self.survey_record_creation.id,
+                "survey_id": self.survey.id,
+                "model_id": self.res_partner_model.id,
+                "field_id": self.name_field.id,
+                "value_origin": "question",
+                "question_id": self.question_name.id,
+            }
+        )
+
+    def _submit_answer(self, name):
+        answer = self._add_answer(survey=self.survey, partner=False, email=False)
+        self._add_answer_line(
+            question=self.question_name, answer=answer, answer_value=name
+        )
+        answer._mark_done()
+        return answer
+
+    def _corrupt_with_wrong_partner(self, answer):
+        # Simulate a pre-fix record: partner_id/email point to someone
+        # unrelated to this participant's own answers.
+        wrong_partner = self.env["res.partner"].create({"name": "Wrong Partner"})
+        answer.write({"partner_id": wrong_partner.id, "email": "wrong@test.fr"})
+        return wrong_partner
+
+    def test_migrate_uses_generated_record_when_available(self):
+        # The submission created its own res.partner: the migration must
+        # relink partner_id/email to that exact contact.
+        answer = self._submit_answer("Jean")
+        jean = self.env["res.partner"].search([("name", "=", "Jean")])
+        self._corrupt_with_wrong_partner(answer)
+
+        self.migration.migrate(self.cr, "16.0.1.0.2")
+        answer.invalidate_recordset()
+
+        self.assertEqual(answer.partner_id, jean)
+        self.assertEqual(answer.email, jean.email)
+
+    def test_migrate_falls_back_to_search_when_no_generated_record(self):
+        # The submission matched an already-existing partner instead of
+        # creating one: no survey.generated.record row exists for it, so the
+        # migration must fall back to re-deriving the match through search.
+        jean = self.env["res.partner"].create(
+            {"name": "Jean", "email": "jean@test.fr"}
+        )
+        self.survey_record_creation.write(
+            {
+                "update_existing_records": True,
+                "field_to_retrieve_existing_records": self.name_field.id,
+            }
+        )
+
+        answer = self._submit_answer("Jean")
+        self.assertFalse(answer.generated_record_ids)
+        self._corrupt_with_wrong_partner(answer)
+
+        self.migration.migrate(self.cr, "16.0.1.0.2")
+        answer.invalidate_recordset()
+
+        self.assertEqual(answer.partner_id, jean)
+        self.assertEqual(answer.email, jean.email)
+
+    def test_migrate_falls_back_to_search_when_generated_partner_was_deleted(self):
+        # The res.partner the submission created has since been deleted
+        # (e.g. merged into another contact): the migration must not crash
+        # on the stale reference and must fall back to search instead.
+        self.survey_record_creation.write(
+            {
+                "update_existing_records": True,
+                "field_to_retrieve_existing_records": self.name_field.id,
+            }
+        )
+
+        answer = self._submit_answer("Jean")
+        first_jean = self.env["res.partner"].search([("name", "=", "Jean")])
+        first_jean.unlink()
+
+        # A new contact with the same name takes over.
+        second_jean = self.env["res.partner"].create(
+            {"name": "Jean", "email": "jean2@test.fr"}
+        )
+        self._corrupt_with_wrong_partner(answer)
+
+        self.migration.migrate(self.cr, "16.0.1.0.2")
+        answer.invalidate_recordset()
+
+        self.assertEqual(answer.partner_id, second_jean)
+        self.assertEqual(answer.email, second_jean.email)
+
+    def test_migrate_leaves_answer_untouched_when_no_partner_can_be_found(self):
+        # No survey.generated.record row (e.g. historical data predating
+        # that log) and no way to search (neither update_existing_records
+        # nor unicity_check configured): the migration must leave
+        # partner_id/email as they are instead of guessing or crashing.
+        answer = self._submit_answer("Jean")
+        answer.generated_record_ids.unlink()
+        wrong_partner = self._corrupt_with_wrong_partner(answer)
+
+        self.migration.migrate(self.cr, "16.0.1.0.2")
+        answer.invalidate_recordset()
+
+        self.assertEqual(answer.partner_id, wrong_partner)
+        self.assertEqual(answer.email, "wrong@test.fr")

survey_record_generation/tests/test_survey_record_creation.py

@@ -930,6 +930,71 @@ class TestSurveyRecordCreation(SurveyCase):
         partner = self.env["res.partner"].search([("name", "=", "Jean")])
         self.assertEqual(self.answer.partner_id, partner)
 
+    def test_partner_id_and_email_are_corrected_when_creating_a_new_partner(self):
+        # A wrong partner_id/email is already set on the answer (e.g.
+        # inherited from the Odoo user who was logged in when the
+        # /survey/start link was opened). The record creation must still
+        # fill partner_id/email with the contact actually created from the
+        # participant's own answers, overriding that wrong value.
+        wrong_partner = self.env["res.partner"].create({"name": "Wrong Partner"})
+
+        self.answer = self._add_answer(
+            survey=self.survey, partner=wrong_partner, email="wrong@test.fr"
+        )
+        self._add_answer_line(
+            question=self.question_name, answer=self.answer, answer_value="Jean"
+        )
+        self.answer._mark_done()
+
+        partner = self.env["res.partner"].search([("name", "=", "Jean")])
+        self.assertEqual(self.answer.partner_id, partner)
+        self.assertNotEqual(self.answer.partner_id, wrong_partner)
+        self.assertEqual(self.answer.email, partner.email)
+
+    def test_partner_id_and_email_are_corrected_when_matching_an_existing_partner(self):
+        # Same as above, but this time the record creation matches an
+        # already-existing partner (via update_existing_records) instead of
+        # creating a new one: partner_id/email must still be corrected to
+        # that matched partner, not left as the wrong pre-existing value.
+        jean = self.env["res.partner"].create({"name": "Jean", "email": "jean@test.fr"})
+        wrong_partner = self.env["res.partner"].create({"name": "Wrong Partner"})
+
+        self.survey_record_creation.write(
+            {
+                "update_existing_records": True,
+                "field_to_retrieve_existing_records": self.name_field.id,
+            }
+        )
+
+        self.answer = self._add_answer(
+            survey=self.survey, partner=wrong_partner, email="wrong@test.fr"
+        )
+        self._add_answer_line(
+            question=self.question_name, answer=self.answer, answer_value="Jean"
+        )
+        self.answer._mark_done()
+
+        self.assertEqual(self.answer.partner_id, jean)
+        self.assertEqual(self.answer.email, jean.email)
+
+    def test_partner_id_and_email_are_corrected_when_matching_a_duplicate(self):
+        # Same as above, but through the unicity_check/find_duplicate branch
+        # instead of update_existing_records.
+        self.name_survey_record_creation_field_values.unicity_check = True
+        jean = self.env["res.partner"].create({"name": "Jean", "email": "jean@test.fr"})
+        wrong_partner = self.env["res.partner"].create({"name": "Wrong Partner"})
+
+        self.answer = self._add_answer(
+            survey=self.survey, partner=wrong_partner, email="wrong@test.fr"
+        )
+        self._add_answer_line(
+            question=self.question_name, answer=self.answer, answer_value="Jean"
+        )
+        self.answer._mark_done()
+
+        self.assertEqual(self.answer.partner_id, jean)
+        self.assertEqual(self.answer.email, jean.email)
+
     def test_partner_id_in_survey_input_is_filled_up_by_first_contact_record_creation(self):
         # In this test, we verify that when creating several contacts with the same survey,
         # the 1st created contact is used to fill up survey_input.partner_id
@@ -976,3 +1041,58 @@ class TestSurveyRecordCreation(SurveyCase):
 
         partner = self.env["res.partner"].search([("name", "=", "Jean")])
         self.assertEqual(self.answer.partner_id, partner)
+
+    def test_partner_id_is_not_overridden_by_second_contact_record_creation(self):
+        # When several res.partner record creations run on the same survey,
+        # partner_id/email must be corrected to match the first contact
+        # created, and the second contact created must not override that
+        # choice, even though the answer started with a wrong
+        # partner_id/email.
+        self.second_question_name = self._add_question(
+            page=None,
+            name="Name of second person",
+            qtype="char_box",
+            survey_id=self.survey.id,
+            sequence=1,
+        )
+
+        self.second_contact_creation = self.env["survey.record.creation"].create(
+            {
+                "name": "Contact 2",
+                "survey_id": self.survey.id,
+                "model_id": self.res_partner_model.id,
+            }
+        )
+        self.env["survey.record.creation.field.values"].create(
+            {
+                "survey_record_creation_id": self.second_contact_creation.id,
+                "survey_id": self.survey.id,
+                "model_id": self.res_partner_model.id,
+                "field_id": self.name_field.id,
+                "value_origin": "question",
+                "question_id": self.second_question_name.id,
+            }
+        )
+
+        wrong_partner = self.env["res.partner"].create({"name": "Wrong Partner"})
+
+        self.answer = self._add_answer(
+            survey=self.survey, partner=wrong_partner, email="wrong@test.fr"
+        )
+        self._add_answer_line(
+            question=self.question_name,
+            answer=self.answer,
+            answer_value="Jean",
+        )
+        self._add_answer_line(
+            question=self.second_question_name,
+            answer=self.answer,
+            answer_value="Jeanne",
+        )
+        self.answer._mark_done()
+
+        jean = self.env["res.partner"].search([("name", "=", "Jean")])
+        jeanne = self.env["res.partner"].search([("name", "=", "Jeanne")])
+        self.assertEqual(self.answer.partner_id, jean)
+        self.assertNotEqual(self.answer.partner_id, jeanne)
+        self.assertEqual(self.answer.email, jean.email)