by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • 17:13
    bors[bot] closed #969
  • 17:13
    bors[bot] closed #968
  • 17:13
    bors[bot] closed #968
  • 14:57
    notriddle edited #969
  • 14:57
    notriddle opened #969
  • 11:43
    zegl opened #968
  • May 27 20:29
    bors[bot] closed #967
  • May 27 20:03
    notriddle closed #111
  • May 27 15:29
    fwitzke opened #111
  • May 27 07:44
    dependabot-preview[bot] closed #951
  • May 27 07:44
    dependabot-preview[bot] labeled #967
  • May 27 07:44
    dependabot-preview[bot] labeled #967
  • May 27 07:44
    dependabot-preview[bot] opened #967
  • May 26 00:24
    dependabot-preview[bot] synchronize #961
  • May 26 00:24
    dependabot-preview[bot] edited #961
  • May 26 00:23
    dependabot-preview[bot] edited #961
  • May 26 00:22
    bors[bot] closed #943
  • May 26 00:04
    notriddle closed #110
  • May 26 00:03
    notriddle closed #194
  • May 25 09:14
    dependabot-preview[bot] closed #960
Michael Howell
@notriddle
If most user names aren't matching, then it's probably something related to conflating the user_id and the user_xref (the former is internal to bors, while the latter is the actual GitHub number).
Which display of the user name is wrong?
The one in the commit? The one in the upper-right corner? Something else?
Jason McFall
@jmcfallargo
The one in the bors web in the upper-right. But I had some users yesterday that tried to bors r+ on the PR and got a "user not permitted" error.
Michael Howell
@notriddle
Weird. Have they tried logging out and logging back in?
notriddle assumes you're using a private instance.
Michael Howell
@notriddle
The only way I know of that could produce that effect would be if you cleared the database without also changing the cookie secret.
Jason McFall
@jmcfallargo
Interesting. Well I definitely did that, but it was after users were complaining that they couldn't bors +. So I wiped the db, restarted bors, and syncd the users. If the cookie needs to be changed, I can try that again.
Michael Howell
@notriddle
Yeah, that's the problem. If you wipe the DB, you should also change the SECRET_KEY_BASE.
When someone logs into the bors dashboard, we give them a signed-and-encrypted cookie that contains their user ID. If you then wipe the database, the user IDs end up getting shuffled around.
Jason McFall
@jmcfallargo
Ah. Thanks - will give that a try!
Jason McFall
@jmcfallargo
Can anybody shed some light on the following crash? I have seen it happening very often lately and don't see a reason why: {{:badmatch, {:error, :push, 422}},
[
{BorsNG.GitHub, :push!, 3,
[file: 'lib/github/github.ex', line: 64]},
{BorsNG.Worker.Batcher, :complete_batch, 3,
[file: 'lib/worker/batcher.ex', line: 381]},
{BorsNG.Worker.Batcher, :maybe_complete_batch, 1,
[file: 'lib/worker/batcher.ex', line: 368]},
{BorsNG.Worker.Batcher, :handle_cast, 2,
[file: 'lib/worker/batcher.ex', line: 77]},
{:gen_server, :try_dispatch, 4,
[file: 'gen_server.erl', line: 637]},
{:gen_server, :handle_msg, 6,
[file: 'gen_server.erl', line: 711]},
{:proc_lib, :init_p_do_apply, 3,
[file: 'proc_lib.erl', line: 249]}
]}
Jason McFall
@jmcfallargo
From what I can tell, this was just failed CI build. Not sure why it logs it in the UI as a crash.
Jason McFall
@jmcfallargo
Just saw this case on a merge conflict - but the PR's get kicked out without reporting such.
Jason McFall
@jmcfallargo
Seeing a lot of these errors in the logs also:
19:33:10.379 [error] #PID<0.11313.15> running BorsNG.Endpoint (cowboy_protocol) terminated
Server: our-internalserver:80 (http)
Request: POST /webhook/github
(exit) an exception was raised: (FunctionClauseError) no function clause matching in BorsNG.WebhookController.do_webhook/3
(bors) lib/web/controllers/webhook_controller.ex:87: BorsNG.WebhookController.do_webhook(%Plug.Conn{adapter: {Plug.Adapters.Cowboy.Conn, :...}, assigns: %{}, before_send: [#Function<1.25166163/1 in Plug.Logger.call/2>], body_params: %{"after" => "0cd65e15aac5eb473f697a51a3b853cbe24271db", "base_ref" => nil, "before" => "53f1f19a34154336da073522cab916fb19d90b67", "commits" => [%{"added" => [
Thandesha
@thanvk
Need some direction. I am using bors running in docker with GiHub Enterprise. The emails links that come from GitHub after merge is done by bors are https://<bors_host>/. But bors docker instance is running in http. How do we run it in https?
Thandesha
@thanvk
Is anybody running BORS on docker with https?
Michael Howell
@notriddle
I think most instances run bors behind a reverse proxy that adds https.
Jason McFall
@jmcfallargo
I am running BORS via docker with HTTPS using a load balancer in front with the certs.
Thandesha
@thanvk
Screen Shot 2018-11-16 at 10.49.07 AM.png
The problem I am facing is that the status link in GitHub Enterprise from Bors is https. I don't know why it is https and not http. How do we tell BORS to do http status URL not the https when bors is running http?
so when I click on that bors Build Succeeded link, it doesn't work as it is pointing to https
Jason McFall
@jmcfallargo
I had that problem, but you can change it to http - Its in the bors configuration somewhere.
I changed mine at some point and had to reinstall bors on github to get the url change to be picked up.
Thandesha
@thanvk

config/prod.secret.exs

config :bors, BorsNG.Endpoint,
http: [port: {:system, "PORT"}],
url: [host: {:system, "PUBLIC_HOST"}, scheme: "https", port: 443],
check_origin: false,
cache_static_manifest: "priv/static/cache_manifest.json",
server: true,
root: ".",
version: Application.spec(:myapp, :vsn),
secret_key_base: {:system, "SECRET_KEY_BASE"}

this is one place I see https
I don't think that is the status return to GitHub details
Thandesha
@thanvk

@jmcfallargo, does your instance of bors has this entry as

url: [host: {:system, "PUBLIC_HOST"}, scheme: "http", port: 80],

?
Michael Moravec
@Majkl578
hi, I'm looking at merge workflow automation tools and came around bors - overall it looks nice, but I have a problem with commenting r+ to indicate approval - why does bors not leverage native GitHub approvals?
Michael Howell
@notriddle
Partially because bors predates native GitHub Reviews.
Partially because GitHub Reviews don't normally trigger your pull request to be merged immediately, so "migrating" to GitHub Reviews doesn't seem to actually simplify anything.
Michael Moravec
@Majkl578
there is also a bit different workflow I would prefer: using GH approvals and labels to indicate PR should be auto-merged, I guess it's not configurable like this?
Michael Howell
@notriddle
No, it doesn't support that.
Thandesha
@thanvk

@notriddle , Since the BORS is running as http, why keep https as default in config/prod.secret.exs for url?
config/prod.secret.exs

config :bors, BorsNG.Endpoint,
http: [port: {:system, "PORT"}],
url: [host: {:system, "PUBLIC_HOST"}, scheme: "https", port: 443],
check_origin: false,
cache_static_manifest: "priv/static/cache_manifest.json",
server: true,
root: ".",
version: Application.spec(:myapp, :vsn),
secret_key_base: {:system, "SECRET_KEY_BASE"}

can default be http and port 80 ?
Jason McFall
@jmcfallargo
Any ideas on this one:

20:07:06.236 [error] GenServer #PID<0.3613.0> terminating
** (Ecto.ConstraintError) constraint error when attempting to insert struct:

* unique: statuses_identifier_batch_id_index

If you would like to convert this constraint into an error, please
call unique_constraint/3 in your changeset and define the proper
constraint name. The changeset has not defined any constraint.

(ecto) lib/ecto/repo/schema.ex:574: anonymous fn/4 in Ecto.Repo.Schema.constraints_to_errors/3
(elixir) lib/enum.ex:1314: Enum."-map/2-lists^map/1-0-"/2
(ecto) lib/ecto/repo/schema.ex:559: Ecto.Repo.Schema.constraints_to_errors/3
(ecto) lib/ecto/repo/schema.ex:222: anonymous fn/14 in Ecto.Repo.Schema.do_insert/4
(ecto) lib/ecto/repo/schema.ex:774: anonymous fn/3 in Ecto.Repo.Schema.wrap_in_transaction/6
(ecto) lib/ecto/adapters/sql.ex:576: anonymous fn/3 in Ecto.Adapters.SQL.do_transaction/3
(db_connection) lib/db_connection.ex:1283: DBConnection.transaction_run/4
(db_connection) lib/db_connection.ex:1207: DBConnection.run_begin/3

Last message: {:poll, :repeat}

We have also had to use the newly introduced timeout - we were seeing many calls take longer than 5 seconds - we have been trying to debug. When it times out, it crashes and removes all of the PRs from the batch
Jason McFall
@jmcfallargo
Any way to mitigate that situation
Thandesha
@thanvk
if the bors is doing the merge build and in meantime someone directly commits changes to the target branch, what will happen?
Jason McFall
@jmcfallargo
You will see cancelled on the PR's and it should re-queue.
Jan-Oliver Kaiser
@Janno
Getting Started mentions that I can optionally protect master, including a restriction to commits with the "bors" status. But Github doesn't offer me that status a s a choice. I can only select the travis status. Do I need to do anything special to do this?
Ah, this forum post contains the same question and it seems that I need to go through one PR before Github knows about the "bors" status.
Jason McFall
@jmcfallargo
We have been hitting this issues - is this known or any idea on why this is happening?
batch
{%Postgrex.Error{
connection_id: 36,
message: nil,
postgres: %{
code: :string_data_right_truncation,
file: "varchar.c",
line: "624",
message: "value too long for type character varying(255)",
pg_code: "22001",
routine: "varchar",
severity: "ERROR",
unknown: "ERROR"
}
},
[
{Ecto.Adapters.SQL, :execute_or_reset, 7,
[file: 'lib/ecto/adapters/sql.ex', line: 440]},
{BorsNG.Worker.Batcher, :do_handle_cast, 2,
[file: 'lib/worker/batcher.ex', line: 148]},
{BorsNG.Worker.Batcher, :handle_cast, 2,
[file: 'lib/worker/batcher.ex', line: 81]},
{:gen_server, :try_dispatch, 4,
[file: 'gen_server.erl', line: 637]},
{:gen_server, :handle_msg, 6,
[file: 'gen_server.erl', line: 711]},
{:proc_lib, :init_p_do_apply, 3,
[file: 'proc_lib.erl', line: 249]}
]}
Gabriel Ferraz
@sudoFerraz
Hey guys, is there a way to wait for internal releases on master branch to be released from circleCI before merging a new commit in the queue generated by bors
?
Jay
@jayzalowitz
Anyone know how to get bors to include more of the commit information in the actual merge