Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Dec 04 00:54
    Server build 3664 deployed
  • Dec 01 23:58
    pkaminski commented #981
  • Dec 01 22:04
    pkaminski commented #981
  • Dec 01 22:02
    pkaminski assigned #981
  • Dec 01 22:02
    pkaminski labeled #981
  • Dec 01 21:35
    jlinder opened #981
  • Nov 29 19:15
    Client build 5716 deployed
  • Nov 23 05:54
    Client build 5715 deployed
  • Nov 23 04:33
    Client build 5714 deployed
  • Nov 23 04:23
    pkaminski closed #514
  • Nov 22 05:13
    Server build 3663 deployed
  • Nov 22 05:05
    Client build 5707 deployed
  • Nov 17 13:41
    EricCousineau-TRI commented #17
  • Nov 17 13:41
    EricCousineau-TRI commented #17
  • Nov 17 13:38
    EricCousineau-TRI commented #222
  • Nov 17 13:36
    EricCousineau-TRI commented #17
  • Nov 17 13:36
    EricCousineau-TRI commented #17
  • Nov 17 13:34
    EricCousineau-TRI commented #17
  • Nov 17 13:33
    EricCousineau-TRI commented #17
  • Nov 16 04:50
    Client build 5703 deployed
Arran Schlosberg
@arran-fathom
*Thanks
Misha Gusarov
@misha-ridge
@pkaminski Hello Piotr. ridge/tectonic#15318 says: "code-review/reviewable Expected — Waiting for status to be reported ". However the review there was finished 15:01 UTC
So Reviewable shows the big green button "ready to merge", but our bot that watches the checks status does not want to merge the PR.
Misha Gusarov
@misha-ridge
@pkaminski Interesting. This PR has got a conflict, so there is no "merge" button in GH interface, but Reviewable proposes to merge it.
Piotr Kaminski
@pkaminski
Strange, let me look it up.
Here's the proximate culprit, now to dig to root cause: "Dropping mergeability due to last revision commit mismatch for ridge/tectonic#15318"
Piotr Kaminski
@pkaminski
It looks like Reviewable is refusing to match GitHub's commit sequence for this PR. GitHub's ends in 5e1a516 then c45a3e1, while Reviewable has those but then goes back to 5e1a516 as the last commit for some reason.
Misha Gusarov
@misha-ridge
Maybe a clue: the list of commits is not linear. 464d46e is a commit on a branch, then 04e2943 is a merge of 464d46e and master, then e3a1ea0 is a commit on top of 464d4e, then 5e1a516 is a merge of e3a1ea0 and master, and finally c45a3e1 is a merge of 04e2943 and 5e1a516
I know it's a messy branch. Sorry about that :)
Piotr Kaminski
@pkaminski
We sort the commits topologically so I don't understand how 5e1a516 could've ended up after c45a3e1 in Reviewable's ordering. Was there ever a moment when 5e1a516 was the head of the branch after c45a3e1 had already been created?
At this point, Reviewable is stuck because it sees that all commits have been captured, yet the last revision isn't at the current head. However we got there this is a situation we need to be able to recover from -- let me think about it a bit after my upcoming meeting, and I'll endeavor to push out a fix later today.
Misha Gusarov
@misha-ridge
No, these two commits were pushed at the same time.
Thanks
Piotr Kaminski
@pkaminski
Hey @misha-ridge, a proper fix will take some thinking as it gets to an underlying design question we've been struggling with for years. In the meantime, I manually adjusted the review so that r3 is considered obsolete and r2 is the latest revision -- I believe that should get things unstuck.
Piotr Kaminski
@pkaminski
I also still don't understand how the review got into this state. Both r2 and r3 were created in one push, with both 5e1a516 and c45a3e1 already on the branch, so c45a3e1 should have ended up as the last revision no matter what unless the topological sort when haywire.
Piotr Kaminski
@pkaminski
(Actually, I had to delete r3 altogether, otherwise it would just get un-obsoleted on sync. Let me know if this helped or not!)
Misha Gusarov
@misha-ridge
@pkaminski Thanks. The PR is now in a sane state.
Dan Halperin
@dhalperi
Is there a way to get a github link from an arbitrary file? I know reviewable does this for large files, and maybe images, but I'd like it for an arbitrary file. In this case (https://reviewable.io/reviews/batfish/pybatfish/837/jupyter_notebooks/Pandas%20Examples.ipynb#-) I want to review GitHub's Jupyter notebook preview.
Piotr Kaminski
@pkaminski
We don't surface it at the file level yet, but if you create a draft then there's a GitHub link behind the "line N" label in the upper-right corner.
Dan Halperin
@dhalperi
:+1:
Artiom Levinton
@artiomle

Hi Piotr, I have some sttange problem in our Reviewable container, it crushed with the following error message

'Unable to initialize generic GitHub access via "github:24". Unknown encryption prefix: undefined'

I have tried to recreate the app secret token and re authorize my session with github (im user 24)
Piotr Kaminski
@pkaminski
(Moving conversation to private channel.)
Michał Kowalczyk
@mkow
just want to say: I really like this new "pondering" state of discussion, it was a nice idea :)
Piotr Kaminski
@pkaminski
Thanks, that's great to hear! All credit goes to @jwnimmer-tri, though. :)
Misha Gusarov
@misha-ridge

Hey. Is there a way to "see myself out" out of a discussion thread in Reviewable? I commented on a thread in a long-running PR, and every day there is a new comment there, and it pops up in my "things to look at", even though I really don't want to look at it anymore.

I know it's a bad idea to have long-running PRs, but we have them from time to time.

Piotr Kaminski
@pkaminski
There's no way to do this right now, sorry. I agree it would be useful in longer multi-party reviews, though. Could you open an issue so we can track this? Thanks!
Misha Gusarov
@misha-ridge
sure, done
Deepak Chhikara
@deepak-harness
Hi Everyone
I'm facing issues in accessing reviewable. It keeps on showing the
resume session message But doesn't load the page. Look like issue with the session since i'm able to access it from the incognito. I tried logging out of github so that oauth can reset the session but no luck.
Is there a way to reset the site data so that it behaves in proper way
Piotr Kaminski
@pkaminski
Hey @deepak-harness, sorry about that. I've heard a few reports of this issue but never enough to get a repro. You can wipe out session information by clearing both local storage and IndexedDB for reviewable.io -- let me know if you need more detailed instructions.
Michael Ferris
@Cellule
@pkaminski when you get a chance to take a look at Reviewable/Reviewable#950
it's currently preventing me and some of my colleagues from reviewing some PRs
Jeremy Nimmer
@jwnimmer-tri
The dashboard https://reviewable.io/reviews#- is failing to load for me as of the past couple hours. Failed to fetch a list of your reviews: TypeError: s.$ref.transaction is not a function Try again. FF 103.0 on Ubuntu 20, private window, no extensions. Any suggestions?
3 replies
Ryba
@sepatel
qq, I've been seeing this error for over an hour now. What does it mean? It is preventing me from doing reviews.
Failed to update your list of reviews: Error: GitHub error 403 on POST https://api.github.com/graphql: You have exceeded a secondary rate limit. Please wait a few minutes before you try again. Note that the review list refreshes automatically, so you don't need to reload the page manually. Try again.
Ryba
@sepatel
Also sometimes it is a 502 error instead of 403 error like in this case,
Failed to update your list of reviews: Error: GitHub error 502 on POST https://api.github.com/graphql: Something went wrong while executing your query. This may be the result of a timeout, or it could be a GitHub bug. Please includeBFEA:38C1:139041:4204FB:62F54A60when reporting this issue. Try again.
23 replies
Violet Whitney
@violetwhitney
Hi, I'm trying to track down invoices for my company's reviewable account but don't seem to have the option to see that detail in my account. Can someone please help?
3 replies
Misha Gusarov
@misha-ridge
@pkaminski Hey Piotr. Is there any chance to get back a single-click way to open a full history of a current conversation in Reviewable? This used to be a no-brainer, and now one has to understand which of 6 options is the right one, especially because I don't remember any need for 5 of 6: I've ever only needed "open the whole history of the current discussion".
27 replies
Misha Gusarov
@misha-ridge

@pkaminski I have noticed a weird issue: in some PRs source code does not get displayed (a spinner spins over every file in the PR). This persists after multiple refreshes.

I had a look at network panel, and haven't found any activity to fetch source code. Metadata (repository, labels, users, members etc) are fetched from GH properly.

The PR where I see it right now is https://reviewable.io/reviews/ridge/tectonic/16879

My browser is FF 106.0, and I'm under Linux/x86-64.

50 replies
Michał Kowalczyk
@mkow
is there any know outage happening right now? I'm getting "Unable to sign in: Firebase: A network AuthError (such as timeout, interrupted connection or unreachable host) has occurred. (auth/network-request-failed). " error on Reviewable and it loads the reviews, but without me being logged in
Michał Kowalczyk
@mkow
ok, disabling uMatrix fixes the issue, but can't use Reviewable with it being enabled, despite allowing everything
I've been using Reviewable with uMatrix for a few years already, did something change in Reviewable/GitHub auth process? or it's a change in uMatrix?
Piotr Kaminski
@pkaminski
No recent changes in Reviewable. Can't speak for Firebase, but I haven't heard of any changes on that side either. It's almost certainly a bug in uMatrix -- note that Reviewable's auth component runs in a shared worker process, so it's possible that uMatrix doesn't apply its settings properly there.
Michał Kowalczyk
@mkow
it worked yestarday, so it's probably a recent change in something
anyways, I have a local solution
Piotr Kaminski
@pkaminski
You can test the shared worker hypothesis by loading the page with ?debug=workers (which forces them to be non-shared) and seeing if that works.
Tony Novak
@afn
I just started using Reviewable and I love it already! Quick question: I have a PR that moved a frequently-imported JS file, so there are a lot of files in which the only change is to an import line. Is there a way to ignore changes that, say, match a certain regular expression?
1 reply
Piotr Kaminski
@pkaminski
older_comments.png
Alexey Feldgendler
@feldgendler
Hi! I remember that commit authors (pullRequest.revisions[i].commits[j].author) were added some time ago to the input for the custom completion function. Recently they have disappeared. Can we please have them back?
6 replies
Stefanus
@sdtqb
We're getting errors like Failed to publish: Request queued (action evaluate) but server did not respond using reviewable right now
2 replies
Alexey Feldgendler
@feldgendler
Hello! I'm getting an error message in a specific PR:
Error: Internal error: mergeable block range mismatch, last=442,444, next=358,358
9 replies
Stefanus
@sdtqb

I use stacked pull requests (no special tooling currently, I know there’s stuff like spr out there), i.e. I often have a chain of PRs where the bases are set so that A merges into main, B merges into A, C merges into B, etc.

My understanding[1] is that github is supposed to automatically retarget the dependent PRs, so e.g. if I merge “A”, the merge base for “B” would be set to main. But if I use “merge & delete” in reviewable to merge A into main, github seems to instead close the A<-B pull request. Any idea how to fix this?

[1] https://github.blog/changelog/2020-05-19-pull-request-retargeting/

6 replies