runbot/forwardport/tests/test_updates.py

475 lines
17 KiB
Python
Raw Permalink Normal View History

"""
Test cases for updating PRs during after the forward-porting process after the
initial merge has succeeded (and forward-porting has started)
"""
import pytest
from utils import seen, matches, Commit, make_basic, to_pr
@pytest.mark.parametrize("merge_parent", [False, True])
def test_update_pr(env, config, make_repo, users, merge_parent) -> None:
""" Even for successful cherrypicks, it's possible that e.g. CI doesn't
pass or the reviewer finds out they need to update the code.
In this case, all following forward ports should... be detached? Or maybe
only this one and its dependent should be updated?
"""
prod, _ = make_basic(env, config, make_repo, statuses='ci/runbot,legal/cla')
# create a branch d from c so we can have 3 forward ports PRs, not just 2,
# for additional checks
env['runbot_merge.project'].search([]).write({
'branch_ids': [(0, 0, {'name': 'd', 'sequence': 40})]
})
with prod:
prod.make_commits('c', Commit('1111', tree={'i': 'a'}), ref='heads/d')
[CHG] *: rewrite commands set, rework status management This commit revisits the commands set in order to make it more regular, and limit inconsistent command-sets, although it includes pseudo-command aliases for common tasks now removed from the core set. Hard Errors =========== The previous iteration of the commands set would ignore any non-command term in a command line. This has been changed to hard error (and ignoring the entire thing) if any command is unknown or invalid. This fixes inconsistent / unexpected interpretations where a user sends a command, then writes a novel on the same line some words of which happen to *also* be commands, leading to merge states they did not expect. They should now be told to fuck off. Priority Restructuring ---------------------- The numerical priority system was pretty messy in that it confused "staging priority" (in ways which were not entirely straightforward) with overrides to other concerns. This has now being split along all the axis, with separate command subsets for: - staging prioritisation, now separated between `default`, `priority`, and `alone`, - `default` means PRs are picked by an unspecified order when creating a staging, if nothing better is available - `priority` means PRs are picked first when staging, however if `priority` PRs don't fill the staging the rest will be filled with `default`, this mode did not previously exist - `alone` means the PRs are picked first, before splits, and only `alone` PRs can be part of the staging (which usually matches the modename) - `skipchecks` overrides both statuses and approval checks, for the batch, something previously implied in `p=0`, but now independent. Setting `skipchecks` basically makes the entire batch `ready`. For consistency this also sets the reviewer implicitly: since skipchecks overrides both statuses *and approval*, whoever enables this mode is essentially the reviewer. - `cancel` cancels any ongoing staging when the marked PR becomes ready again, previously this was also implied (in a more restricted form) by setting `p=0` FWBot removal ============= While the "forwardport bot" still exists as an API level (to segregate access rights between tokens) it has been removed as an interaction point, as part of the modules merge plan. As a result, fwbot stops responding ---------------------- Feedback messages are now always sent by the mergebot, the forward-porting bot should not send any message or notification anymore. commands moved to the merge bot ------------------------------- - `ignore`/`up to` simply changes bot - `close` as well - `skipci` is now a choice / flag of an `fw` command, which denotes the forward-port policy, - `fw=default` is the old `ci` and resets the policy to default, that is wait for the PR to be merged to create forward ports, and for the required statuses on each forward port to be received before creating the next - `fw=skipci` is the old `skipci`, it waits for the merge of the base PR but then creates all the forward ports immediately (unless it gets a conflict) - `fw=skipmerge` immediately creates all the forward ports, without even waiting for the PR to be merged This is a completely new mode, and may be rather broken as until now the 'bot has always assumed the source PR had been merged. approval rework --------------- Because of the previous section, there is no distinguishing feature between `mergebot r+` = "merge this PR" and `forwardbot r+` = "merge this PR and all its parent with different access rights". As a result, the two have been merged under a single `mergebot r+` with heuristics attempting to provide the best experience: - if approving a non-forward port, the behavior does not change - else, with review rights on the source, all ancestors are approved - else, as author of the original, approves all ancestors which descend from a merged PR - else, approves all ancestors up to and including the oldest ancestor to which we have review rights Most notably, the source's author is not delegated on the source or any of its descendants anymore. This might need to be revisited if it provides too restrictive. For the very specialized need of approving a forward-port *and none of its ancestors*, `review=` can now take a comma (`,`) separated list of pull request numbers (github numbers, not mergebot ids). Computed State ============== The `state` field of pull requests is now computed. Hopefully this makes the status more consistent and predictable in the long run, and importantly makes status management more reliable (because reference datum get updated naturally flowing to the state). For now however it makes things more complicated as some of the states have to be separately signaled or updated: - `closed` and `error` are now separate flags - `merge_date` is pulled down from forwardport and becomes the transition signal for ready -> merged - `reviewed_by` becomes the transition signal for approval (might be a good idea to rename it...) - `status` is computed from the head's statuses and overrides, and *that* becomes the validation state Ideally, batch-level flags like `skipchecks` should be on, well, the batch, and `state` should have a dependency on the batch. However currently the batch is not a durable / permanent member of the system, so it's a PR-level flag and a messy pile. On notable change is that *forcing* the state to `ready` now does that but also sets the reviewer, `skipchecks`, and overrides to ensure the API-mediated readying does not get rolled back by e.g. the runbot sending a status. This is useful for a few types of automated / programmatic PRs e.g. translation exports, where we set the state programmatically to limit noise. recursive dependency hack ------------------------- Given a sequence of PRs with an override of the source, if one of the PRs is updated its descendants should not have the override anymore. However if the updated PR gets overridden, its descendants should have *that* override. This requires some unholy manipulations via an override of `modified`, as the ORM supports recursive fields but not recursive dependencies (on a different field). unconditional followup scheduling --------------------------------- Previously scheduling forward-port followup was contigent on the FW policy, but it's not actually correct if the new PR is *immediately* validated (which can happen now that the field is computed, if there are no required statuses *or* all of the required statuses are overridden by an ancestor) as nothing will trigger the state change and thus scheduling of the fp followup. The followup function checks all the properties of the batch to port, so this should not result on incorrect ports. Although it's a bit more expensive, and will lead to more spam. Previously this would not happen because on creation of a PR the validation task (commit -> PR) would still have to execute. Misc Changes ============ - If a PR is marked as overriding / canceling stagings, it now does so on retry not just when setting initially. This was not handled at all previously, so a PR in P0 going into error due to e.g. a non-deterministic bug would be retried and still p=0, but a current staging would not get cancelled. Same when a PR in p=0 goes into error because something was failed, then is updated with a fix. - Add tracking to a bunch of relevant PR fields. Post-mortem analysis currently generally requires going through the text logs to see what happened, which is annoying. There is a nondeterminism / inconsistency in the tracking which sometimes leads the admin user to trigger tracking before the bot does, leading to the staging tracking being attributed to them during tests, shove under the carpet by ignoring the user to whom that tracking is attributed. When multiple users update tracked fields in the same transaction all the changes are attributed to the first one having triggered tracking (?), I couldn't find why the admin sometimes takes over. - added and leveraged support for enum-backed selection fields - moved variuous fields from forwardport to runbot_merge - fix a migration which had never worked and which never run (because I forgot to bump the version on the module) - remove some unnecessary intermediate de/serialisation fixes #673, fixes #309, fixes #792, fixes #846 (probably)
2023-10-31 13:42:07 +07:00
with prod:
[p_1] = prod.make_commits(
'a',
Commit('p_0', tree={'x': '0'}),
ref='heads/hugechange'
)
pr = prod.make_pr(target='a', head='hugechange')
pr.post_comment('hansen r+', config['role_reviewer']['token'])
prod.post_status(p_1, 'success', 'legal/cla')
prod.post_status(p_1, 'failure', 'ci/runbot')
env.run_crons()
assert pr.comments == [
(users['reviewer'], 'hansen r+'),
seen(env, pr, users),
(users['user'], "@{user} @{reviewer} 'ci/runbot' failed on this reviewed PR.".format_map(users)),
]
with prod:
prod.post_status(p_1, 'success', 'ci/runbot')
env.run_crons()
with prod:
prod.post_status('staging.a', 'success', 'legal/cla')
prod.post_status('staging.a', 'success', 'ci/runbot')
# should merge the staging then create the FP PR
env.run_crons()
pr0_id, pr1_id = env['runbot_merge.pull_requests'].search([], order='number')
fp_intermediate = (users['user'], '''\
This PR targets b and is part of the forward-port chain. Further PRs will be created up to d.
More info at https://github.com/odoo/odoo/wiki/Mergebot#forward-port
''')
ci_warning = (users['user'], '@%(user)s @%(reviewer)s ci/runbot failed on this forward-port PR' % users)
# oh no CI of the first FP PR failed!
# simulate status being sent multiple times (e.g. on multiple repos) with
# some delivery lag allowing for the cron to run between each delivery
for st, ctx in [('failure', 'ci/runbot'), ('failure', 'ci/runbot'), ('success', 'legal/cla'), ('success', 'legal/cla')]:
with prod:
prod.post_status(pr1_id.head, st, ctx)
env.run_crons()
with prod: # should be ignored because the description doesn't matter
prod.post_status(pr1_id.head, 'failure', 'ci/runbot', description="HAHAHAHAHA")
env.run_crons()
# check that FP did not resume & we have a ping on the PR
assert env['runbot_merge.pull_requests'].search([], order='number') == pr0_id | pr1_id,\
"forward port should not continue on CI failure"
pr1_remote = prod.get_pr(pr1_id.number)
assert pr1_remote.comments == [seen(env, pr1_remote, users), fp_intermediate, ci_warning]
# it was a false positive, rebuild... it fails again!
with prod:
prod.post_status(pr1_id.head, 'failure', 'ci/runbot', target_url='http://example.org/4567890')
env.run_crons()
# check that FP did not resume & we have a ping on the PR
assert env['runbot_merge.pull_requests'].search([], order='number') == pr0_id | pr1_id,\
"ensure it still hasn't restarted"
assert pr1_remote.comments == [seen(env, pr1_remote, users), fp_intermediate, ci_warning, ci_warning]
# nb: updating the head would detach the PR and not put it in the warning
# path anymore
# rebuild again, finally passes
with prod:
prod.post_status(pr1_id.head, 'success', 'ci/runbot')
env.run_crons()
pr0_id, pr1_id, pr2_id = env['runbot_merge.pull_requests'].search([], order='number')
assert pr1_id.parent_id == pr0_id
assert pr2_id.parent_id == pr1_id
pr1_head = pr1_id.head
pr2_head = pr2_id.head
# turns out branch b is syntactically but not semantically compatible! It
# needs x to be 5!
pr_repo, pr_ref = prod.get_pr(pr1_id.number).branch
with pr_repo:
# force-push correct commit to PR's branch
[new_c] = pr_repo.make_commits(
prod.commit(pr1_id.target.name).id,
Commit('whop whop', tree={'x': '5'}),
ref='heads/%s' % pr_ref,
make=False
)
env.run_crons()
assert pr1_id.head == new_c != pr1_head, "the FP PR should be updated"
assert not pr1_id.parent_id, "the FP PR should be detached from the original"
# NOTE: should the followup PR wait for pr1 CI or not?
assert pr2_id.head != pr2_head
assert pr2_id.parent_id == pr1_id, "the followup PR should still be linked"
assert prod.read_tree(prod.commit(pr1_id.head)) == {
'f': 'c',
'g': 'b',
'x': '5'
}, "the FP PR should have the new code"
assert prod.read_tree(prod.commit(pr2_id.head)) == {
'f': 'c',
'g': 'a',
'h': 'a',
'x': '5'
}, "the followup FP should also have the update"
[CHG] *: rewrite commands set, rework status management This commit revisits the commands set in order to make it more regular, and limit inconsistent command-sets, although it includes pseudo-command aliases for common tasks now removed from the core set. Hard Errors =========== The previous iteration of the commands set would ignore any non-command term in a command line. This has been changed to hard error (and ignoring the entire thing) if any command is unknown or invalid. This fixes inconsistent / unexpected interpretations where a user sends a command, then writes a novel on the same line some words of which happen to *also* be commands, leading to merge states they did not expect. They should now be told to fuck off. Priority Restructuring ---------------------- The numerical priority system was pretty messy in that it confused "staging priority" (in ways which were not entirely straightforward) with overrides to other concerns. This has now being split along all the axis, with separate command subsets for: - staging prioritisation, now separated between `default`, `priority`, and `alone`, - `default` means PRs are picked by an unspecified order when creating a staging, if nothing better is available - `priority` means PRs are picked first when staging, however if `priority` PRs don't fill the staging the rest will be filled with `default`, this mode did not previously exist - `alone` means the PRs are picked first, before splits, and only `alone` PRs can be part of the staging (which usually matches the modename) - `skipchecks` overrides both statuses and approval checks, for the batch, something previously implied in `p=0`, but now independent. Setting `skipchecks` basically makes the entire batch `ready`. For consistency this also sets the reviewer implicitly: since skipchecks overrides both statuses *and approval*, whoever enables this mode is essentially the reviewer. - `cancel` cancels any ongoing staging when the marked PR becomes ready again, previously this was also implied (in a more restricted form) by setting `p=0` FWBot removal ============= While the "forwardport bot" still exists as an API level (to segregate access rights between tokens) it has been removed as an interaction point, as part of the modules merge plan. As a result, fwbot stops responding ---------------------- Feedback messages are now always sent by the mergebot, the forward-porting bot should not send any message or notification anymore. commands moved to the merge bot ------------------------------- - `ignore`/`up to` simply changes bot - `close` as well - `skipci` is now a choice / flag of an `fw` command, which denotes the forward-port policy, - `fw=default` is the old `ci` and resets the policy to default, that is wait for the PR to be merged to create forward ports, and for the required statuses on each forward port to be received before creating the next - `fw=skipci` is the old `skipci`, it waits for the merge of the base PR but then creates all the forward ports immediately (unless it gets a conflict) - `fw=skipmerge` immediately creates all the forward ports, without even waiting for the PR to be merged This is a completely new mode, and may be rather broken as until now the 'bot has always assumed the source PR had been merged. approval rework --------------- Because of the previous section, there is no distinguishing feature between `mergebot r+` = "merge this PR" and `forwardbot r+` = "merge this PR and all its parent with different access rights". As a result, the two have been merged under a single `mergebot r+` with heuristics attempting to provide the best experience: - if approving a non-forward port, the behavior does not change - else, with review rights on the source, all ancestors are approved - else, as author of the original, approves all ancestors which descend from a merged PR - else, approves all ancestors up to and including the oldest ancestor to which we have review rights Most notably, the source's author is not delegated on the source or any of its descendants anymore. This might need to be revisited if it provides too restrictive. For the very specialized need of approving a forward-port *and none of its ancestors*, `review=` can now take a comma (`,`) separated list of pull request numbers (github numbers, not mergebot ids). Computed State ============== The `state` field of pull requests is now computed. Hopefully this makes the status more consistent and predictable in the long run, and importantly makes status management more reliable (because reference datum get updated naturally flowing to the state). For now however it makes things more complicated as some of the states have to be separately signaled or updated: - `closed` and `error` are now separate flags - `merge_date` is pulled down from forwardport and becomes the transition signal for ready -> merged - `reviewed_by` becomes the transition signal for approval (might be a good idea to rename it...) - `status` is computed from the head's statuses and overrides, and *that* becomes the validation state Ideally, batch-level flags like `skipchecks` should be on, well, the batch, and `state` should have a dependency on the batch. However currently the batch is not a durable / permanent member of the system, so it's a PR-level flag and a messy pile. On notable change is that *forcing* the state to `ready` now does that but also sets the reviewer, `skipchecks`, and overrides to ensure the API-mediated readying does not get rolled back by e.g. the runbot sending a status. This is useful for a few types of automated / programmatic PRs e.g. translation exports, where we set the state programmatically to limit noise. recursive dependency hack ------------------------- Given a sequence of PRs with an override of the source, if one of the PRs is updated its descendants should not have the override anymore. However if the updated PR gets overridden, its descendants should have *that* override. This requires some unholy manipulations via an override of `modified`, as the ORM supports recursive fields but not recursive dependencies (on a different field). unconditional followup scheduling --------------------------------- Previously scheduling forward-port followup was contigent on the FW policy, but it's not actually correct if the new PR is *immediately* validated (which can happen now that the field is computed, if there are no required statuses *or* all of the required statuses are overridden by an ancestor) as nothing will trigger the state change and thus scheduling of the fp followup. The followup function checks all the properties of the batch to port, so this should not result on incorrect ports. Although it's a bit more expensive, and will lead to more spam. Previously this would not happen because on creation of a PR the validation task (commit -> PR) would still have to execute. Misc Changes ============ - If a PR is marked as overriding / canceling stagings, it now does so on retry not just when setting initially. This was not handled at all previously, so a PR in P0 going into error due to e.g. a non-deterministic bug would be retried and still p=0, but a current staging would not get cancelled. Same when a PR in p=0 goes into error because something was failed, then is updated with a fix. - Add tracking to a bunch of relevant PR fields. Post-mortem analysis currently generally requires going through the text logs to see what happened, which is annoying. There is a nondeterminism / inconsistency in the tracking which sometimes leads the admin user to trigger tracking before the bot does, leading to the staging tracking being attributed to them during tests, shove under the carpet by ignoring the user to whom that tracking is attributed. When multiple users update tracked fields in the same transaction all the changes are attributed to the first one having triggered tracking (?), I couldn't find why the admin sometimes takes over. - added and leveraged support for enum-backed selection fields - moved variuous fields from forwardport to runbot_merge - fix a migration which had never worked and which never run (because I forgot to bump the version on the module) - remove some unnecessary intermediate de/serialisation fixes #673, fixes #309, fixes #792, fixes #846 (probably)
2023-10-31 13:42:07 +07:00
with prod:
prod.post_status(pr2_id.head, 'success', 'ci/runbot')
prod.post_status(pr2_id.head, 'success', 'legal/cla')
env.run_crons()
pr2 = prod.get_pr(pr2_id.number)
if merge_parent:
with prod:
pr2.post_comment('hansen r+', config['role_reviewer']['token'])
env.run_crons()
with prod:
prod.post_status('staging.c', 'success', 'ci/runbot')
prod.post_status('staging.c', 'success', 'legal/cla')
env.run_crons()
assert pr2_id.state == 'merged'
_0, _1, _2, pr3_id = env['runbot_merge.pull_requests'].search([], order='number')
assert pr3_id.parent_id == pr2_id
# don't bother updating heads (?)
pr3_id.write({'parent_id': False, 'detach_reason': "testing"})
# pump feedback messages
env.run_crons()
[CHG] *: rewrite commands set, rework status management This commit revisits the commands set in order to make it more regular, and limit inconsistent command-sets, although it includes pseudo-command aliases for common tasks now removed from the core set. Hard Errors =========== The previous iteration of the commands set would ignore any non-command term in a command line. This has been changed to hard error (and ignoring the entire thing) if any command is unknown or invalid. This fixes inconsistent / unexpected interpretations where a user sends a command, then writes a novel on the same line some words of which happen to *also* be commands, leading to merge states they did not expect. They should now be told to fuck off. Priority Restructuring ---------------------- The numerical priority system was pretty messy in that it confused "staging priority" (in ways which were not entirely straightforward) with overrides to other concerns. This has now being split along all the axis, with separate command subsets for: - staging prioritisation, now separated between `default`, `priority`, and `alone`, - `default` means PRs are picked by an unspecified order when creating a staging, if nothing better is available - `priority` means PRs are picked first when staging, however if `priority` PRs don't fill the staging the rest will be filled with `default`, this mode did not previously exist - `alone` means the PRs are picked first, before splits, and only `alone` PRs can be part of the staging (which usually matches the modename) - `skipchecks` overrides both statuses and approval checks, for the batch, something previously implied in `p=0`, but now independent. Setting `skipchecks` basically makes the entire batch `ready`. For consistency this also sets the reviewer implicitly: since skipchecks overrides both statuses *and approval*, whoever enables this mode is essentially the reviewer. - `cancel` cancels any ongoing staging when the marked PR becomes ready again, previously this was also implied (in a more restricted form) by setting `p=0` FWBot removal ============= While the "forwardport bot" still exists as an API level (to segregate access rights between tokens) it has been removed as an interaction point, as part of the modules merge plan. As a result, fwbot stops responding ---------------------- Feedback messages are now always sent by the mergebot, the forward-porting bot should not send any message or notification anymore. commands moved to the merge bot ------------------------------- - `ignore`/`up to` simply changes bot - `close` as well - `skipci` is now a choice / flag of an `fw` command, which denotes the forward-port policy, - `fw=default` is the old `ci` and resets the policy to default, that is wait for the PR to be merged to create forward ports, and for the required statuses on each forward port to be received before creating the next - `fw=skipci` is the old `skipci`, it waits for the merge of the base PR but then creates all the forward ports immediately (unless it gets a conflict) - `fw=skipmerge` immediately creates all the forward ports, without even waiting for the PR to be merged This is a completely new mode, and may be rather broken as until now the 'bot has always assumed the source PR had been merged. approval rework --------------- Because of the previous section, there is no distinguishing feature between `mergebot r+` = "merge this PR" and `forwardbot r+` = "merge this PR and all its parent with different access rights". As a result, the two have been merged under a single `mergebot r+` with heuristics attempting to provide the best experience: - if approving a non-forward port, the behavior does not change - else, with review rights on the source, all ancestors are approved - else, as author of the original, approves all ancestors which descend from a merged PR - else, approves all ancestors up to and including the oldest ancestor to which we have review rights Most notably, the source's author is not delegated on the source or any of its descendants anymore. This might need to be revisited if it provides too restrictive. For the very specialized need of approving a forward-port *and none of its ancestors*, `review=` can now take a comma (`,`) separated list of pull request numbers (github numbers, not mergebot ids). Computed State ============== The `state` field of pull requests is now computed. Hopefully this makes the status more consistent and predictable in the long run, and importantly makes status management more reliable (because reference datum get updated naturally flowing to the state). For now however it makes things more complicated as some of the states have to be separately signaled or updated: - `closed` and `error` are now separate flags - `merge_date` is pulled down from forwardport and becomes the transition signal for ready -> merged - `reviewed_by` becomes the transition signal for approval (might be a good idea to rename it...) - `status` is computed from the head's statuses and overrides, and *that* becomes the validation state Ideally, batch-level flags like `skipchecks` should be on, well, the batch, and `state` should have a dependency on the batch. However currently the batch is not a durable / permanent member of the system, so it's a PR-level flag and a messy pile. On notable change is that *forcing* the state to `ready` now does that but also sets the reviewer, `skipchecks`, and overrides to ensure the API-mediated readying does not get rolled back by e.g. the runbot sending a status. This is useful for a few types of automated / programmatic PRs e.g. translation exports, where we set the state programmatically to limit noise. recursive dependency hack ------------------------- Given a sequence of PRs with an override of the source, if one of the PRs is updated its descendants should not have the override anymore. However if the updated PR gets overridden, its descendants should have *that* override. This requires some unholy manipulations via an override of `modified`, as the ORM supports recursive fields but not recursive dependencies (on a different field). unconditional followup scheduling --------------------------------- Previously scheduling forward-port followup was contigent on the FW policy, but it's not actually correct if the new PR is *immediately* validated (which can happen now that the field is computed, if there are no required statuses *or* all of the required statuses are overridden by an ancestor) as nothing will trigger the state change and thus scheduling of the fp followup. The followup function checks all the properties of the batch to port, so this should not result on incorrect ports. Although it's a bit more expensive, and will lead to more spam. Previously this would not happen because on creation of a PR the validation task (commit -> PR) would still have to execute. Misc Changes ============ - If a PR is marked as overriding / canceling stagings, it now does so on retry not just when setting initially. This was not handled at all previously, so a PR in P0 going into error due to e.g. a non-deterministic bug would be retried and still p=0, but a current staging would not get cancelled. Same when a PR in p=0 goes into error because something was failed, then is updated with a fix. - Add tracking to a bunch of relevant PR fields. Post-mortem analysis currently generally requires going through the text logs to see what happened, which is annoying. There is a nondeterminism / inconsistency in the tracking which sometimes leads the admin user to trigger tracking before the bot does, leading to the staging tracking being attributed to them during tests, shove under the carpet by ignoring the user to whom that tracking is attributed. When multiple users update tracked fields in the same transaction all the changes are attributed to the first one having triggered tracking (?), I couldn't find why the admin sometimes takes over. - added and leveraged support for enum-backed selection fields - moved variuous fields from forwardport to runbot_merge - fix a migration which had never worked and which never run (because I forgot to bump the version on the module) - remove some unnecessary intermediate de/serialisation fixes #673, fixes #309, fixes #792, fixes #846 (probably)
2023-10-31 13:42:07 +07:00
pr3 = prod.get_pr(pr3_id.number)
assert pr3.comments == [
seen(env, pr3, users),
(users['user'], f"""\
@{users['user']} @{users['reviewer']} this PR targets d and is the last of the forward-port chain containing:
* {pr2_id.display_name}
To merge the full chain, use
[CHG] *: rewrite commands set, rework status management This commit revisits the commands set in order to make it more regular, and limit inconsistent command-sets, although it includes pseudo-command aliases for common tasks now removed from the core set. Hard Errors =========== The previous iteration of the commands set would ignore any non-command term in a command line. This has been changed to hard error (and ignoring the entire thing) if any command is unknown or invalid. This fixes inconsistent / unexpected interpretations where a user sends a command, then writes a novel on the same line some words of which happen to *also* be commands, leading to merge states they did not expect. They should now be told to fuck off. Priority Restructuring ---------------------- The numerical priority system was pretty messy in that it confused "staging priority" (in ways which were not entirely straightforward) with overrides to other concerns. This has now being split along all the axis, with separate command subsets for: - staging prioritisation, now separated between `default`, `priority`, and `alone`, - `default` means PRs are picked by an unspecified order when creating a staging, if nothing better is available - `priority` means PRs are picked first when staging, however if `priority` PRs don't fill the staging the rest will be filled with `default`, this mode did not previously exist - `alone` means the PRs are picked first, before splits, and only `alone` PRs can be part of the staging (which usually matches the modename) - `skipchecks` overrides both statuses and approval checks, for the batch, something previously implied in `p=0`, but now independent. Setting `skipchecks` basically makes the entire batch `ready`. For consistency this also sets the reviewer implicitly: since skipchecks overrides both statuses *and approval*, whoever enables this mode is essentially the reviewer. - `cancel` cancels any ongoing staging when the marked PR becomes ready again, previously this was also implied (in a more restricted form) by setting `p=0` FWBot removal ============= While the "forwardport bot" still exists as an API level (to segregate access rights between tokens) it has been removed as an interaction point, as part of the modules merge plan. As a result, fwbot stops responding ---------------------- Feedback messages are now always sent by the mergebot, the forward-porting bot should not send any message or notification anymore. commands moved to the merge bot ------------------------------- - `ignore`/`up to` simply changes bot - `close` as well - `skipci` is now a choice / flag of an `fw` command, which denotes the forward-port policy, - `fw=default` is the old `ci` and resets the policy to default, that is wait for the PR to be merged to create forward ports, and for the required statuses on each forward port to be received before creating the next - `fw=skipci` is the old `skipci`, it waits for the merge of the base PR but then creates all the forward ports immediately (unless it gets a conflict) - `fw=skipmerge` immediately creates all the forward ports, without even waiting for the PR to be merged This is a completely new mode, and may be rather broken as until now the 'bot has always assumed the source PR had been merged. approval rework --------------- Because of the previous section, there is no distinguishing feature between `mergebot r+` = "merge this PR" and `forwardbot r+` = "merge this PR and all its parent with different access rights". As a result, the two have been merged under a single `mergebot r+` with heuristics attempting to provide the best experience: - if approving a non-forward port, the behavior does not change - else, with review rights on the source, all ancestors are approved - else, as author of the original, approves all ancestors which descend from a merged PR - else, approves all ancestors up to and including the oldest ancestor to which we have review rights Most notably, the source's author is not delegated on the source or any of its descendants anymore. This might need to be revisited if it provides too restrictive. For the very specialized need of approving a forward-port *and none of its ancestors*, `review=` can now take a comma (`,`) separated list of pull request numbers (github numbers, not mergebot ids). Computed State ============== The `state` field of pull requests is now computed. Hopefully this makes the status more consistent and predictable in the long run, and importantly makes status management more reliable (because reference datum get updated naturally flowing to the state). For now however it makes things more complicated as some of the states have to be separately signaled or updated: - `closed` and `error` are now separate flags - `merge_date` is pulled down from forwardport and becomes the transition signal for ready -> merged - `reviewed_by` becomes the transition signal for approval (might be a good idea to rename it...) - `status` is computed from the head's statuses and overrides, and *that* becomes the validation state Ideally, batch-level flags like `skipchecks` should be on, well, the batch, and `state` should have a dependency on the batch. However currently the batch is not a durable / permanent member of the system, so it's a PR-level flag and a messy pile. On notable change is that *forcing* the state to `ready` now does that but also sets the reviewer, `skipchecks`, and overrides to ensure the API-mediated readying does not get rolled back by e.g. the runbot sending a status. This is useful for a few types of automated / programmatic PRs e.g. translation exports, where we set the state programmatically to limit noise. recursive dependency hack ------------------------- Given a sequence of PRs with an override of the source, if one of the PRs is updated its descendants should not have the override anymore. However if the updated PR gets overridden, its descendants should have *that* override. This requires some unholy manipulations via an override of `modified`, as the ORM supports recursive fields but not recursive dependencies (on a different field). unconditional followup scheduling --------------------------------- Previously scheduling forward-port followup was contigent on the FW policy, but it's not actually correct if the new PR is *immediately* validated (which can happen now that the field is computed, if there are no required statuses *or* all of the required statuses are overridden by an ancestor) as nothing will trigger the state change and thus scheduling of the fp followup. The followup function checks all the properties of the batch to port, so this should not result on incorrect ports. Although it's a bit more expensive, and will lead to more spam. Previously this would not happen because on creation of a PR the validation task (commit -> PR) would still have to execute. Misc Changes ============ - If a PR is marked as overriding / canceling stagings, it now does so on retry not just when setting initially. This was not handled at all previously, so a PR in P0 going into error due to e.g. a non-deterministic bug would be retried and still p=0, but a current staging would not get cancelled. Same when a PR in p=0 goes into error because something was failed, then is updated with a fix. - Add tracking to a bunch of relevant PR fields. Post-mortem analysis currently generally requires going through the text logs to see what happened, which is annoying. There is a nondeterminism / inconsistency in the tracking which sometimes leads the admin user to trigger tracking before the bot does, leading to the staging tracking being attributed to them during tests, shove under the carpet by ignoring the user to whom that tracking is attributed. When multiple users update tracked fields in the same transaction all the changes are attributed to the first one having triggered tracking (?), I couldn't find why the admin sometimes takes over. - added and leveraged support for enum-backed selection fields - moved variuous fields from forwardport to runbot_merge - fix a migration which had never worked and which never run (because I forgot to bump the version on the module) - remove some unnecessary intermediate de/serialisation fixes #673, fixes #309, fixes #792, fixes #846 (probably)
2023-10-31 13:42:07 +07:00
> @hansen r+
More info at https://github.com/odoo/odoo/wiki/Mergebot#forward-port
"""),
(users['user'], f"@{users['user']} @{users['reviewer']} this PR was "
f"modified / updated and has become a normal PR. It "
[CHG] *: rewrite commands set, rework status management This commit revisits the commands set in order to make it more regular, and limit inconsistent command-sets, although it includes pseudo-command aliases for common tasks now removed from the core set. Hard Errors =========== The previous iteration of the commands set would ignore any non-command term in a command line. This has been changed to hard error (and ignoring the entire thing) if any command is unknown or invalid. This fixes inconsistent / unexpected interpretations where a user sends a command, then writes a novel on the same line some words of which happen to *also* be commands, leading to merge states they did not expect. They should now be told to fuck off. Priority Restructuring ---------------------- The numerical priority system was pretty messy in that it confused "staging priority" (in ways which were not entirely straightforward) with overrides to other concerns. This has now being split along all the axis, with separate command subsets for: - staging prioritisation, now separated between `default`, `priority`, and `alone`, - `default` means PRs are picked by an unspecified order when creating a staging, if nothing better is available - `priority` means PRs are picked first when staging, however if `priority` PRs don't fill the staging the rest will be filled with `default`, this mode did not previously exist - `alone` means the PRs are picked first, before splits, and only `alone` PRs can be part of the staging (which usually matches the modename) - `skipchecks` overrides both statuses and approval checks, for the batch, something previously implied in `p=0`, but now independent. Setting `skipchecks` basically makes the entire batch `ready`. For consistency this also sets the reviewer implicitly: since skipchecks overrides both statuses *and approval*, whoever enables this mode is essentially the reviewer. - `cancel` cancels any ongoing staging when the marked PR becomes ready again, previously this was also implied (in a more restricted form) by setting `p=0` FWBot removal ============= While the "forwardport bot" still exists as an API level (to segregate access rights between tokens) it has been removed as an interaction point, as part of the modules merge plan. As a result, fwbot stops responding ---------------------- Feedback messages are now always sent by the mergebot, the forward-porting bot should not send any message or notification anymore. commands moved to the merge bot ------------------------------- - `ignore`/`up to` simply changes bot - `close` as well - `skipci` is now a choice / flag of an `fw` command, which denotes the forward-port policy, - `fw=default` is the old `ci` and resets the policy to default, that is wait for the PR to be merged to create forward ports, and for the required statuses on each forward port to be received before creating the next - `fw=skipci` is the old `skipci`, it waits for the merge of the base PR but then creates all the forward ports immediately (unless it gets a conflict) - `fw=skipmerge` immediately creates all the forward ports, without even waiting for the PR to be merged This is a completely new mode, and may be rather broken as until now the 'bot has always assumed the source PR had been merged. approval rework --------------- Because of the previous section, there is no distinguishing feature between `mergebot r+` = "merge this PR" and `forwardbot r+` = "merge this PR and all its parent with different access rights". As a result, the two have been merged under a single `mergebot r+` with heuristics attempting to provide the best experience: - if approving a non-forward port, the behavior does not change - else, with review rights on the source, all ancestors are approved - else, as author of the original, approves all ancestors which descend from a merged PR - else, approves all ancestors up to and including the oldest ancestor to which we have review rights Most notably, the source's author is not delegated on the source or any of its descendants anymore. This might need to be revisited if it provides too restrictive. For the very specialized need of approving a forward-port *and none of its ancestors*, `review=` can now take a comma (`,`) separated list of pull request numbers (github numbers, not mergebot ids). Computed State ============== The `state` field of pull requests is now computed. Hopefully this makes the status more consistent and predictable in the long run, and importantly makes status management more reliable (because reference datum get updated naturally flowing to the state). For now however it makes things more complicated as some of the states have to be separately signaled or updated: - `closed` and `error` are now separate flags - `merge_date` is pulled down from forwardport and becomes the transition signal for ready -> merged - `reviewed_by` becomes the transition signal for approval (might be a good idea to rename it...) - `status` is computed from the head's statuses and overrides, and *that* becomes the validation state Ideally, batch-level flags like `skipchecks` should be on, well, the batch, and `state` should have a dependency on the batch. However currently the batch is not a durable / permanent member of the system, so it's a PR-level flag and a messy pile. On notable change is that *forcing* the state to `ready` now does that but also sets the reviewer, `skipchecks`, and overrides to ensure the API-mediated readying does not get rolled back by e.g. the runbot sending a status. This is useful for a few types of automated / programmatic PRs e.g. translation exports, where we set the state programmatically to limit noise. recursive dependency hack ------------------------- Given a sequence of PRs with an override of the source, if one of the PRs is updated its descendants should not have the override anymore. However if the updated PR gets overridden, its descendants should have *that* override. This requires some unholy manipulations via an override of `modified`, as the ORM supports recursive fields but not recursive dependencies (on a different field). unconditional followup scheduling --------------------------------- Previously scheduling forward-port followup was contigent on the FW policy, but it's not actually correct if the new PR is *immediately* validated (which can happen now that the field is computed, if there are no required statuses *or* all of the required statuses are overridden by an ancestor) as nothing will trigger the state change and thus scheduling of the fp followup. The followup function checks all the properties of the batch to port, so this should not result on incorrect ports. Although it's a bit more expensive, and will lead to more spam. Previously this would not happen because on creation of a PR the validation task (commit -> PR) would still have to execute. Misc Changes ============ - If a PR is marked as overriding / canceling stagings, it now does so on retry not just when setting initially. This was not handled at all previously, so a PR in P0 going into error due to e.g. a non-deterministic bug would be retried and still p=0, but a current staging would not get cancelled. Same when a PR in p=0 goes into error because something was failed, then is updated with a fix. - Add tracking to a bunch of relevant PR fields. Post-mortem analysis currently generally requires going through the text logs to see what happened, which is annoying. There is a nondeterminism / inconsistency in the tracking which sometimes leads the admin user to trigger tracking before the bot does, leading to the staging tracking being attributed to them during tests, shove under the carpet by ignoring the user to whom that tracking is attributed. When multiple users update tracked fields in the same transaction all the changes are attributed to the first one having triggered tracking (?), I couldn't find why the admin sometimes takes over. - added and leveraged support for enum-backed selection fields - moved variuous fields from forwardport to runbot_merge - fix a migration which had never worked and which never run (because I forgot to bump the version on the module) - remove some unnecessary intermediate de/serialisation fixes #673, fixes #309, fixes #792, fixes #846 (probably)
2023-10-31 13:42:07 +07:00
f"must be merged directly."
)
]
assert pr2.comments[:2] == [
seen(env, pr2, users),
(users['user'], """\
This PR targets c and is part of the forward-port chain. Further PRs will be created up to d.
More info at https://github.com/odoo/odoo/wiki/Mergebot#forward-port
"""),
]
if merge_parent:
assert pr2.comments[2:] == [
(users['reviewer'], "hansen r+"),
]
else:
assert pr2.comments[2:] == [
(users['user'], f"@{users['user']} @{users['reviewer']} child PR "
f"{pr3_id.display_name} was modified / updated and has "
f"become a normal PR. This PR (and any of its parents) "
f"will need to be merged independently as approvals "
f"won't cross."),
]
def test_update_merged(env, make_repo, config, users):
""" Strange things happen when an FP gets closed / merged but then its
parent is modified and the forwardport tries to update the (now merged)
child.
Turns out the issue is the followup: given a PR a and forward port targets
B -> C -> D. When a is merged we get b, c and d. If c gets merged *then*
b gets updated, the fwbot will update c in turn, then it will look for the
head of the updated c in order to create d.
However it *will not* find that head, as update events don't get propagated
on closed PRs (this is generally a good thing). As a result, the sanity
check when trying to port c to d will fail.
After checking with nim, the safest behaviour seems to be:
* stop at the update of the first closed or merged PR
* signal on that PR that something fucky happened
* also maybe disable or exponentially backoff the update job after some
number of attempts?
"""
prod, _ = make_basic(env, config, make_repo, statuses='default')
# add a 4th branch
with prod:
prod.make_ref('heads/d', prod.commit('c').id)
env['runbot_merge.project'].search([]).write({
'branch_ids': [(0, 0, {'name': 'd', 'sequence': 40})]
})
with prod:
[c] = prod.make_commits('a', Commit('p_0', tree={'0': '0'}), ref='heads/hugechange')
pr = prod.make_pr(target='a', head='hugechange')
prod.post_status(c, 'success')
pr.post_comment('hansen r+', config['role_reviewer']['token'])
env.run_crons()
with prod:
prod.post_status('staging.a', 'success')
env.run_crons()
_, pr1_id = env['runbot_merge.pull_requests'].search([], order='number')
with prod:
prod.post_status(pr1_id.head, 'success')
env.run_crons()
pr0_id, pr1_id, pr2_id = env['runbot_merge.pull_requests'].search([], order='number')
pr2 = prod.get_pr(pr2_id.number)
with prod:
pr2.post_comment('hansen r+', config['role_reviewer']['token'])
prod.post_status(pr2_id.head, 'success')
env.run_crons()
assert pr2_id.staging_id
with prod:
prod.post_status('staging.c', 'success')
env.run_crons()
assert pr2_id.state == 'merged'
assert pr2.state == 'closed'
# now we can try updating pr1 and see what happens
repo, ref = prod.get_pr(pr1_id.number).branch
with repo:
repo.make_commits(
prod.commit(pr1_id.target.name).id,
Commit('2', tree={'0': '0', '1': '1'}),
ref='heads/%s' % ref,
make=False
)
updates = env['forwardport.updates'].search([])
assert updates
assert updates.original_root == pr0_id
assert updates.new_root == pr1_id
env.run_crons()
assert not pr1_id.parent_id
assert not env['forwardport.updates'].search([])
assert pr2.comments == [
seen(env, pr2, users),
(users['user'], '''This PR targets c and is part of the forward-port chain. Further PRs will be created up to d.
More info at https://github.com/odoo/odoo/wiki/Mergebot#forward-port
'''),
(users['reviewer'], 'hansen r+'),
(users['user'], """@%s @%s ancestor PR %s has been updated but this PR is merged and can't be updated to match.
You may want or need to manually update any followup PR.""" % (
users['user'],
users['reviewer'],
pr1_id.display_name,
))
]
def test_duplicate_fw(env, make_repo, setreviewers, config, users):
""" Test for #451
"""
# 0 - 1 - 2 - 3 - 4 master
# \ - 31 v3
# \ - 21 v2
# \ - 11 v1
repo = make_repo('proj')
with repo:
_, c1, c2, c3, _ = repo.make_commits(
None,
Commit('0', tree={'f': 'a'}),
Commit('1', tree={'f': 'b'}),
Commit('2', tree={'f': 'c'}),
Commit('3', tree={'f': 'd'}),
Commit('4', tree={'f': 'e'}),
ref='heads/master'
)
repo.make_commits(c1, Commit('11', tree={'g': 'a'}), ref='heads/v1')
repo.make_commits(c2, Commit('21', tree={'h': 'a'}), ref='heads/v2')
repo.make_commits(c3, Commit('31', tree={'i': 'a'}), ref='heads/v3')
proj = env['runbot_merge.project'].create({
'name': 'a project',
'github_token': config['github']['token'],
'github_prefix': 'hansen',
'github_name': config['github']['name'],
'github_email': "foo@example.org",
'fp_github_token': config['github']['token'],
'fp_github_name': 'herbert',
'branch_ids': [
(0, 0, {'name': 'master', 'sequence': 0}),
(0, 0, {'name': 'v3', 'sequence': 1}),
(0, 0, {'name': 'v2', 'sequence': 2}),
(0, 0, {'name': 'v1', 'sequence': 3}),
],
'repo_ids': [
(0, 0, {
'name': repo.name,
'required_statuses': 'ci',
'fp_remote_target': repo.name,
})
]
})
setreviewers(*proj.repo_ids)
env['runbot_merge.events_sources'].create({'repository': repo.name})
# create a PR in v1, merge it, then create all 3 ports
with repo:
repo.make_commits('v1', Commit('c0', tree={'z': 'a'}), ref='heads/hugechange')
prv1 = repo.make_pr(target='v1', head='hugechange')
repo.post_status('hugechange', 'success', 'ci')
prv1.post_comment('hansen r+', config['role_reviewer']['token'])
env.run_crons()
PRs = env['runbot_merge.pull_requests']
prv1_id = PRs.search([
('repository.name', '=', repo.name),
('number', '=', prv1.number),
])
assert prv1_id.state == 'ready'
with repo:
repo.post_status('staging.v1', 'success', 'ci')
env.run_crons()
assert prv1_id.state == 'merged'
parent = prv1_id
while True:
child = PRs.search([('parent_id', '=', parent.id)])
if not child:
break
assert child.state == 'opened'
with repo:
repo.post_status(child.head, 'success', 'ci')
env.run_crons()
parent = child
pr_ids = _, prv2_id, prv3_id, prmaster_id = PRs.search([], order='number')
_, prv2, _prv3, _prmaster = [repo.get_pr(p.number) for p in pr_ids]
assert pr_ids.mapped('target.name') == ['v1', 'v2', 'v3', 'master']
assert pr_ids.mapped('state') == ['merged', 'validated', 'validated', 'validated']
assert repo.read_tree(repo.commit(prmaster_id.head)) == {'f': 'e', 'z': 'a'}
with repo:
repo.make_commits('v2', Commit('c0', tree={'z': 'b'}), ref=prv2.ref, make=False)
env.run_crons()
[CHG] *: rewrite commands set, rework status management This commit revisits the commands set in order to make it more regular, and limit inconsistent command-sets, although it includes pseudo-command aliases for common tasks now removed from the core set. Hard Errors =========== The previous iteration of the commands set would ignore any non-command term in a command line. This has been changed to hard error (and ignoring the entire thing) if any command is unknown or invalid. This fixes inconsistent / unexpected interpretations where a user sends a command, then writes a novel on the same line some words of which happen to *also* be commands, leading to merge states they did not expect. They should now be told to fuck off. Priority Restructuring ---------------------- The numerical priority system was pretty messy in that it confused "staging priority" (in ways which were not entirely straightforward) with overrides to other concerns. This has now being split along all the axis, with separate command subsets for: - staging prioritisation, now separated between `default`, `priority`, and `alone`, - `default` means PRs are picked by an unspecified order when creating a staging, if nothing better is available - `priority` means PRs are picked first when staging, however if `priority` PRs don't fill the staging the rest will be filled with `default`, this mode did not previously exist - `alone` means the PRs are picked first, before splits, and only `alone` PRs can be part of the staging (which usually matches the modename) - `skipchecks` overrides both statuses and approval checks, for the batch, something previously implied in `p=0`, but now independent. Setting `skipchecks` basically makes the entire batch `ready`. For consistency this also sets the reviewer implicitly: since skipchecks overrides both statuses *and approval*, whoever enables this mode is essentially the reviewer. - `cancel` cancels any ongoing staging when the marked PR becomes ready again, previously this was also implied (in a more restricted form) by setting `p=0` FWBot removal ============= While the "forwardport bot" still exists as an API level (to segregate access rights between tokens) it has been removed as an interaction point, as part of the modules merge plan. As a result, fwbot stops responding ---------------------- Feedback messages are now always sent by the mergebot, the forward-porting bot should not send any message or notification anymore. commands moved to the merge bot ------------------------------- - `ignore`/`up to` simply changes bot - `close` as well - `skipci` is now a choice / flag of an `fw` command, which denotes the forward-port policy, - `fw=default` is the old `ci` and resets the policy to default, that is wait for the PR to be merged to create forward ports, and for the required statuses on each forward port to be received before creating the next - `fw=skipci` is the old `skipci`, it waits for the merge of the base PR but then creates all the forward ports immediately (unless it gets a conflict) - `fw=skipmerge` immediately creates all the forward ports, without even waiting for the PR to be merged This is a completely new mode, and may be rather broken as until now the 'bot has always assumed the source PR had been merged. approval rework --------------- Because of the previous section, there is no distinguishing feature between `mergebot r+` = "merge this PR" and `forwardbot r+` = "merge this PR and all its parent with different access rights". As a result, the two have been merged under a single `mergebot r+` with heuristics attempting to provide the best experience: - if approving a non-forward port, the behavior does not change - else, with review rights on the source, all ancestors are approved - else, as author of the original, approves all ancestors which descend from a merged PR - else, approves all ancestors up to and including the oldest ancestor to which we have review rights Most notably, the source's author is not delegated on the source or any of its descendants anymore. This might need to be revisited if it provides too restrictive. For the very specialized need of approving a forward-port *and none of its ancestors*, `review=` can now take a comma (`,`) separated list of pull request numbers (github numbers, not mergebot ids). Computed State ============== The `state` field of pull requests is now computed. Hopefully this makes the status more consistent and predictable in the long run, and importantly makes status management more reliable (because reference datum get updated naturally flowing to the state). For now however it makes things more complicated as some of the states have to be separately signaled or updated: - `closed` and `error` are now separate flags - `merge_date` is pulled down from forwardport and becomes the transition signal for ready -> merged - `reviewed_by` becomes the transition signal for approval (might be a good idea to rename it...) - `status` is computed from the head's statuses and overrides, and *that* becomes the validation state Ideally, batch-level flags like `skipchecks` should be on, well, the batch, and `state` should have a dependency on the batch. However currently the batch is not a durable / permanent member of the system, so it's a PR-level flag and a messy pile. On notable change is that *forcing* the state to `ready` now does that but also sets the reviewer, `skipchecks`, and overrides to ensure the API-mediated readying does not get rolled back by e.g. the runbot sending a status. This is useful for a few types of automated / programmatic PRs e.g. translation exports, where we set the state programmatically to limit noise. recursive dependency hack ------------------------- Given a sequence of PRs with an override of the source, if one of the PRs is updated its descendants should not have the override anymore. However if the updated PR gets overridden, its descendants should have *that* override. This requires some unholy manipulations via an override of `modified`, as the ORM supports recursive fields but not recursive dependencies (on a different field). unconditional followup scheduling --------------------------------- Previously scheduling forward-port followup was contigent on the FW policy, but it's not actually correct if the new PR is *immediately* validated (which can happen now that the field is computed, if there are no required statuses *or* all of the required statuses are overridden by an ancestor) as nothing will trigger the state change and thus scheduling of the fp followup. The followup function checks all the properties of the batch to port, so this should not result on incorrect ports. Although it's a bit more expensive, and will lead to more spam. Previously this would not happen because on creation of a PR the validation task (commit -> PR) would still have to execute. Misc Changes ============ - If a PR is marked as overriding / canceling stagings, it now does so on retry not just when setting initially. This was not handled at all previously, so a PR in P0 going into error due to e.g. a non-deterministic bug would be retried and still p=0, but a current staging would not get cancelled. Same when a PR in p=0 goes into error because something was failed, then is updated with a fix. - Add tracking to a bunch of relevant PR fields. Post-mortem analysis currently generally requires going through the text logs to see what happened, which is annoying. There is a nondeterminism / inconsistency in the tracking which sometimes leads the admin user to trigger tracking before the bot does, leading to the staging tracking being attributed to them during tests, shove under the carpet by ignoring the user to whom that tracking is attributed. When multiple users update tracked fields in the same transaction all the changes are attributed to the first one having triggered tracking (?), I couldn't find why the admin sometimes takes over. - added and leveraged support for enum-backed selection fields - moved variuous fields from forwardport to runbot_merge - fix a migration which had never worked and which never run (because I forgot to bump the version on the module) - remove some unnecessary intermediate de/serialisation fixes #673, fixes #309, fixes #792, fixes #846 (probably)
2023-10-31 13:42:07 +07:00
assert pr_ids.mapped('state') == ['merged', 'opened', 'opened', 'opened']
assert repo.read_tree(repo.commit(prv2_id.head)) == {'f': 'c', 'h': 'a', 'z': 'b'}
assert repo.read_tree(repo.commit(prv3_id.head)) == {'f': 'd', 'i': 'a', 'z': 'b'}
assert repo.read_tree(repo.commit(prmaster_id.head)) == {'f': 'e', 'z': 'b'}
assert prv2_id.source_id == prv1_id
assert not prv2_id.parent_id
env.run_crons()
assert PRs.search([], order='number') == pr_ids
with repo:
repo.post_status(prv2.head, 'success', 'ci')
prv2.post_comment('hansen r+', config['role_reviewer']['token'])
env.run_crons()
with repo:
repo.post_status('staging.v2', 'success', 'ci')
env.run_crons()
# env.run_crons()
assert PRs.search([], order='number') == pr_ids
def test_subsequent_conflict(env, make_repo, config, users):
""" Test for updating an fw PR in the case where it produces a conflict in
the followup. Cf #467.
"""
repo, fork = make_basic(env, config, make_repo, statuses='default')
# create a PR in branch A which adds a new file
with repo:
repo.make_commits('a', Commit('newfile', tree={'x': '0'}), ref='heads/pr1')
pr_1 = repo.make_pr(target='a', head='pr1')
repo.post_status('pr1', 'success')
pr_1.post_comment('hansen r+', config['role_reviewer']['token'])
env.run_crons()
with repo:
repo.post_status('staging.a', 'success')
env.run_crons()
pr1_id = to_pr(env, pr_1)
assert pr1_id.state == 'merged'
pr2_id = env['runbot_merge.pull_requests'].search([('source_id', '=', pr1_id.id)])
assert pr2_id
with repo:
repo.post_status(pr2_id.head, 'success')
env.run_crons()
pr3_id = env['runbot_merge.pull_requests'].search([('parent_id', '=', pr2_id.id)])
assert pr3_id
assert repo.read_tree(repo.commit(pr3_id.head)) == {
'f': 'c',
'g': 'a',
'h': 'a',
'x': '0',
}
# update pr2: add a file "h"
pr2 = repo.get_pr(pr2_id.number)
t = {**repo.read_tree(repo.commit(pr2_id.head)), 'h': 'conflict!'}
with fork:
fork.make_commits(repo.commit(pr2_id.target.name).id, Commit('newfiles', tree=t), ref=pr2.ref, make=False)
env.run_crons()
assert repo.read_tree(repo.commit(pr3_id.head)) == {
'f': 'c',
'g': 'a',
[CHG] forwardport: perform forward porting without working copies The goal is to reduce maintenance and odd disk interactions & concurrency issues, by not creating concurrent clones, not having to push forks back in the repository, etc... it also removes the need to cleanup "scratch" working copies though that looks not to have been an issue in a while. The work is done on isolated objects without using or mutating refs, so even concurrent work should not be a problem. This turns out to not be any more verbose (less so if anything) than using `cherry-pick`, as that is not really designed for scripted / non-interactive use, or for squashing commits thereafter. Working directly with trees and commits is quite a bit cleaner even without a ton of helpers. Much of the credit goes to Julia Evans for [their investigation of 3-way merges as the underpinnings of cherry-picking][3-way merge], this would have been a lot more difficult if I'd had to rediscover the merge-base trick independently. A few things have been changed by this: - The old trace/stderr from cherrypick has disappeared as it's generated by cherrypick, but for a non-interactive use it's kinda useless anyway so I probably should have looked into removing it earlier (I think the main use was investigation of the inflateinit issue). - Error on emptied commits has to be hand-rolled as `merge-tree` couldn't care less, this is not hard but is a bit annoying. - `merge-tree`'s conflict information only references raw commits, which makes sense, but requires updating a bunch of tests. Then again so does the fact that it *usually* doesn't send anything to stderr, so that's usually disappearing. Conveniently `merge-tree` merges the conflict marker directly in the files / tree so we don't have to mess about moving them back out of the repository and into the working copy as I assume cherry-pick does, which means we don't have to try and commit them back in ether. That is a huge part of the gain over faffing about with the working copy. Fixes #847 [3-way merge]: https://jvns.ca/blog/2023/11/10/how-cherry-pick-and-revert-work/
2024-07-05 18:32:02 +07:00
'h': matches('''<<<\x3c<<< $$
a
[CHG] forwardport: perform forward porting without working copies The goal is to reduce maintenance and odd disk interactions & concurrency issues, by not creating concurrent clones, not having to push forks back in the repository, etc... it also removes the need to cleanup "scratch" working copies though that looks not to have been an issue in a while. The work is done on isolated objects without using or mutating refs, so even concurrent work should not be a problem. This turns out to not be any more verbose (less so if anything) than using `cherry-pick`, as that is not really designed for scripted / non-interactive use, or for squashing commits thereafter. Working directly with trees and commits is quite a bit cleaner even without a ton of helpers. Much of the credit goes to Julia Evans for [their investigation of 3-way merges as the underpinnings of cherry-picking][3-way merge], this would have been a lot more difficult if I'd had to rediscover the merge-base trick independently. A few things have been changed by this: - The old trace/stderr from cherrypick has disappeared as it's generated by cherrypick, but for a non-interactive use it's kinda useless anyway so I probably should have looked into removing it earlier (I think the main use was investigation of the inflateinit issue). - Error on emptied commits has to be hand-rolled as `merge-tree` couldn't care less, this is not hard but is a bit annoying. - `merge-tree`'s conflict information only references raw commits, which makes sense, but requires updating a bunch of tests. Then again so does the fact that it *usually* doesn't send anything to stderr, so that's usually disappearing. Conveniently `merge-tree` merges the conflict marker directly in the files / tree so we don't have to mess about moving them back out of the repository and into the working copy as I assume cherry-pick does, which means we don't have to try and commit them back in ether. That is a huge part of the gain over faffing about with the working copy. Fixes #847 [3-way merge]: https://jvns.ca/blog/2023/11/10/how-cherry-pick-and-revert-work/
2024-07-05 18:32:02 +07:00
||||||| $$
=======
conflict!
[CHG] forwardport: perform forward porting without working copies The goal is to reduce maintenance and odd disk interactions & concurrency issues, by not creating concurrent clones, not having to push forks back in the repository, etc... it also removes the need to cleanup "scratch" working copies though that looks not to have been an issue in a while. The work is done on isolated objects without using or mutating refs, so even concurrent work should not be a problem. This turns out to not be any more verbose (less so if anything) than using `cherry-pick`, as that is not really designed for scripted / non-interactive use, or for squashing commits thereafter. Working directly with trees and commits is quite a bit cleaner even without a ton of helpers. Much of the credit goes to Julia Evans for [their investigation of 3-way merges as the underpinnings of cherry-picking][3-way merge], this would have been a lot more difficult if I'd had to rediscover the merge-base trick independently. A few things have been changed by this: - The old trace/stderr from cherrypick has disappeared as it's generated by cherrypick, but for a non-interactive use it's kinda useless anyway so I probably should have looked into removing it earlier (I think the main use was investigation of the inflateinit issue). - Error on emptied commits has to be hand-rolled as `merge-tree` couldn't care less, this is not hard but is a bit annoying. - `merge-tree`'s conflict information only references raw commits, which makes sense, but requires updating a bunch of tests. Then again so does the fact that it *usually* doesn't send anything to stderr, so that's usually disappearing. Conveniently `merge-tree` merges the conflict marker directly in the files / tree so we don't have to mess about moving them back out of the repository and into the working copy as I assume cherry-pick does, which means we don't have to try and commit them back in ether. That is a huge part of the gain over faffing about with the working copy. Fixes #847 [3-way merge]: https://jvns.ca/blog/2023/11/10/how-cherry-pick-and-revert-work/
2024-07-05 18:32:02 +07:00
>>>\x3e>>> $$
'''),
'x': '0',
}
# skip comments:
# 1. link to mergebot status page
# 2. "forward port chain" bit
# 3. updated / modified & got detached
assert pr2.comments[3:] == [
(users['user'], f"@{users['user']} @{users['reviewer']} WARNING: the latest change ({pr2_id.head}) triggered "
f"a conflict when updating the next forward-port "
f"({pr3_id.display_name}), and has been ignored.\n\n"
f"You will need to update this pull request "
f"differently, or fix the issue by hand on "
f"{pr3_id.display_name}.")
]
# skip comments:
# 1. link to status page
# 2. forward-port chain thing
assert repo.get_pr(pr3_id.number).comments[2:] == [
[CHG] forwardport: perform forward porting without working copies The goal is to reduce maintenance and odd disk interactions & concurrency issues, by not creating concurrent clones, not having to push forks back in the repository, etc... it also removes the need to cleanup "scratch" working copies though that looks not to have been an issue in a while. The work is done on isolated objects without using or mutating refs, so even concurrent work should not be a problem. This turns out to not be any more verbose (less so if anything) than using `cherry-pick`, as that is not really designed for scripted / non-interactive use, or for squashing commits thereafter. Working directly with trees and commits is quite a bit cleaner even without a ton of helpers. Much of the credit goes to Julia Evans for [their investigation of 3-way merges as the underpinnings of cherry-picking][3-way merge], this would have been a lot more difficult if I'd had to rediscover the merge-base trick independently. A few things have been changed by this: - The old trace/stderr from cherrypick has disappeared as it's generated by cherrypick, but for a non-interactive use it's kinda useless anyway so I probably should have looked into removing it earlier (I think the main use was investigation of the inflateinit issue). - Error on emptied commits has to be hand-rolled as `merge-tree` couldn't care less, this is not hard but is a bit annoying. - `merge-tree`'s conflict information only references raw commits, which makes sense, but requires updating a bunch of tests. Then again so does the fact that it *usually* doesn't send anything to stderr, so that's usually disappearing. Conveniently `merge-tree` merges the conflict marker directly in the files / tree so we don't have to mess about moving them back out of the repository and into the working copy as I assume cherry-pick does, which means we don't have to try and commit them back in ether. That is a huge part of the gain over faffing about with the working copy. Fixes #847 [3-way merge]: https://jvns.ca/blog/2023/11/10/how-cherry-pick-and-revert-work/
2024-07-05 18:32:02 +07:00
(users['user'], f'''\
@{users['user']} @{users['reviewer']} WARNING: the update of {pr2_id.display_name} to {pr2_id.head} has caused a \
conflict in this pull request, data may have been lost.
stdout:
```
Auto-merging h
CONFLICT (add/add): Merge conflict in h
[CHG] forwardport: perform forward porting without working copies The goal is to reduce maintenance and odd disk interactions & concurrency issues, by not creating concurrent clones, not having to push forks back in the repository, etc... it also removes the need to cleanup "scratch" working copies though that looks not to have been an issue in a while. The work is done on isolated objects without using or mutating refs, so even concurrent work should not be a problem. This turns out to not be any more verbose (less so if anything) than using `cherry-pick`, as that is not really designed for scripted / non-interactive use, or for squashing commits thereafter. Working directly with trees and commits is quite a bit cleaner even without a ton of helpers. Much of the credit goes to Julia Evans for [their investigation of 3-way merges as the underpinnings of cherry-picking][3-way merge], this would have been a lot more difficult if I'd had to rediscover the merge-base trick independently. A few things have been changed by this: - The old trace/stderr from cherrypick has disappeared as it's generated by cherrypick, but for a non-interactive use it's kinda useless anyway so I probably should have looked into removing it earlier (I think the main use was investigation of the inflateinit issue). - Error on emptied commits has to be hand-rolled as `merge-tree` couldn't care less, this is not hard but is a bit annoying. - `merge-tree`'s conflict information only references raw commits, which makes sense, but requires updating a bunch of tests. Then again so does the fact that it *usually* doesn't send anything to stderr, so that's usually disappearing. Conveniently `merge-tree` merges the conflict marker directly in the files / tree so we don't have to mess about moving them back out of the repository and into the working copy as I assume cherry-pick does, which means we don't have to try and commit them back in ether. That is a huge part of the gain over faffing about with the working copy. Fixes #847 [3-way merge]: https://jvns.ca/blog/2023/11/10/how-cherry-pick-and-revert-work/
2024-07-05 18:32:02 +07:00
```'''),
]