Hello! How do I delete a stack version with sensitive data? Despite it being a 1-commit branch, I followed:

We just store the SHA; we don't store code. So if you ask GH to delete the commit containing sensitive data, you are good (edited)

[5:44 PM]

https://docs.github.com/en/authentication/keeping-your-account-and-data-secure/removing-sensitive-data-from-a-repository

There is no commit in github with sensitive data, but the stack version persists. I tried closing the PR, merging the branch onto its parent feature branch, creating new PR with same branch. I have been unable to remove the version.

Is there a way to place a label on a PR that is on the merge queue?

Hello - Can you view github deployments on the graphite pull request review page?

I have a stack of 6 PRs. I manually closed PR number 4 on Github because it contained changes we didn't want. Now PR 5 and 6 are somehow based off of main and no longer in sync with my stack. I suspect this happend because of my manually closing the PR and then doing more work locally on my upstream branches without running

gt sync

. I noticed my that the first of my "lost" branches had it's target branch on my open PR set as

graphite/<number>

which I don't think was correct. I tried manually changing it with the GUI on github to my expected target (PR 3, since 4 was closed manually), and then running

gt sync

, but still it does not track accordingly to what I desire.

I'd like to use claude's /security-review command (blog post), but the source prompt is not compatible with non-base branches in graphite stacks since it uses the following commands to get the diffs to review.

FILES MODIFIED:

!

git diff --name-only origin/HEAD...

COMMITS:

!

git log --no-decorate origin/HEAD...

DIFF CONTENT:

!

git diff --merge-base origin/HEAD

`````` We could of course fork this prompt, but ideally we don't have to do that so we can automatically benefit from any updates that the anthropic team pushes. Has anyone else found a simple workaround for this / other workflows that aren't natively graphite compatible?

Is graphite chat stack-aware? If I ask it for a change and it gives me an apply button, will it restack and modify PRs stacked on top of it?

How can this be?

08-05-08-05-tech-4040-create-new-reports-page-8
Checked out 08-06-tech-4040-create-new-reports-page-10.
This branch has local changes that need to be submitted.
Run gt submit to push your changes.

this branch has local changes that need to be submitted. Run gt submit to push your changes

I run `gt submit`:

🥞 Validating that this Graphite stack is ready to submit...
WARNING: The following branches do not introduce any changes:
▸ 08-05-08-05-tech-4040-create-new-reports-page-8
▸ 08-06-tech-4040-create-new-reports-page-10

Are there local changes or not?

i am running my checks (tests, linters, ...) twice: when creating my PRs and when adding the PRs to the merge queue. Can i deactivate the second run and only run when creating the PR?

any clever solutions to using gitmoji CLI tooling with

gt

? other than manually typing the

:shortcode:

like my caveman ass has been doing

Hey guys - for billing why are we defaulted to annual billing? We want to pay monthly, how do I fix it

how do I use the CLI to auto track all outstanding PRs on top of my current stack that aren't written by me? as in someone created a stack on top of mine and I wouldn't know that someone stacked on top of it without them telling me or checking in the GUI. git fetch pulls the branches but doesn't track in them in graphite

if i fold some PR branches that have been reviewed, will I lose the comments? we have a stack with a bunch of red PRs and then a green one; I want to merge them all at once, without having to fix the CI errors on the lower ones... if there's a better way to do this, then I'm happy to try that instead

Let’s say I have a stack of 3 PRs. How can I instead turn that into 3 separate stacks where they all have the trunk branch as the parent?

Hi folks! I signed up and authenticated Graphite with Github (e.g., I can see my PRs on graphite.dev), but am stuck on the token authentication step (e.g., I can copy-paste in

gt auth --token [token-here]

, but nothing happens in the command line and I can't actually run any of the

gt

commands) I installed graphite with Homebrew -- when I run

gt

, I see the list of possible commands, and it asks me to authenticate before moving forward. However, run I try to run any commands (e.g.,

gt --version

,

gt --help --all

,

gt auth --token \ [my-token-here

), nothing shows up in the console. I've updated my git version (2.50), I've tried uninstalling and reinstalling via brew (and also npm), I've rebooted, and I also applied any XCode + other software updates. I've also tried restarting my terminal and re-sourcing my .zshrc file. Is there something else I'm missing? I've asked a teammate to compare

where gt

and the installation locations on a similar Mac setup, and files look like they're installed in the same place.

Hi Graphite team, Quick questions about the “Time to First Review” metric: • If a reviewer just adds a comment, does that count as the first review, or does it only count when a formal review submission (Approve / Request Changes / Comment) is made? • Does the metric include holidays and weekends, or is it calculated only on business days?

I’ve been digging through the Graphite Diamond docs to better understand how to customize code review hygiene rules, and I could use some clarification. From the docs, I see that Diamond supports: • Exclusions (ignoring generated code, specific paths, etc.) • Custom rules via prompts (UI) and file-based rules • Prebuilt templates (e.g. OWASP, Airbnb JS, PEP) What I haven’t been able to find is: a. What exactly do file-based rules look like? Is there a schema (YAML, JSON, Markdown, etc.) and can these rules live inside the repo under version control? b. Can Diamond directly ingest a repo’s STYLE_GUIDE.md (or similar markdown doc) as rules, or do those need to be re-expressed in custom prompts? c. How much control do we have over severity / enforcement — e.g. marking a rule as blocking vs. just a suggestion? d. Are there examples of repos using custom/file-based rules that I could review? I’ve gone through the Diamond customization and review comments sections, but these specific details weren’t clear. Would love any guidance or examples others here can share 🙏

Is there an easy way to authenticate more complex API calls? I'm working on some tooling for my developer experience, and it appears that I can pass an auth token in the header (e.g.

{ authorization: "token XXXxxxXXX" }

in some of my calls to

<http://api.graphite.dev|api.graphite.dev>

, however more complex requests (e.g.

<http://api.graphite.dev/v1/graphite/mutation/mq-toggle-pause|api.graphite.dev/v1/graphite/mutation/mq-toggle-pause>

) appear to follow a more complex flow that tokens might not be useful for

Is there a way to make

gt create

not super slow when you have a lot of branches locally?

what is the best way to use codex with graphite?

can i somehow remove the "Generate description" button for PR descriptions? i have only ever clicked it by accident and it's annoyed me every time. i think that PR descriptions is a bad use case for AI, since a good PR description doesn't summarize the code, but rather gives context about the PR that isn't available in the diff itself.

I'm using Graphite Chat and just got hit with the usage limit. Is there a way to raise that?

Has anyone else had issues viewing Graphite metrics for the last 7 days? I wanted to look at my PR review request response time. I've submitted several PRs and reviews in that time, but Graphite doesn't display me?

How do I set up the Diamond free plan? Am currently in the trial version and don't want to pay for > 100 PRs / mo

Hey team good morning 😄 We are using the CI optimization in the merge queue (to merge per stack instead of per PRs) but one drawback is that since graphite is creating a new draft PR to group all the PRs of the stack, we cannot leverage our NX cache with the key being commit-based. Is there any best practice recommended for this ?

Hey, did anyone figure out how to use renovates automerge feature with the graphite merge queue? Or do you just merge renovate PRs immediately?

Hi folks. My team is considering moving onto graphite and we have an existing github action process we worried might conflict with graphite. A bot updates versions in code on every push to main. Will that disrupt the graphite merge queue?

Did something change where using GitHub to merge PRs in a stack no longer works? If I do the following: • Merge the first PR in a stack with GitHub • checkout the second PR in the stack locally and run

gt get

Normally it would prompt saying that the first branch was deleted but now it doesn't, version 1.6.8

I tried to improve the merge queue settings (merging with Graphite is currently significantly slower than on Github which the team starts to complain more and more about). Enabling "Fast forward merge" worked but it seems to not change much (mostly talking when merging a stack of multiple accepted PRs). Then I tried to enable "CI settings" but no matter which settings I use, PR merge fail but I can't get any specific details for the error unfortunately beyond "The Graphite merge queue couldn't merge this PR because it failed for an unknown reason." How can I make it work / faster?

Loving graphite so far, definitely seems to speed up our review process. But wondering how to work with efficiently with a feature / epic branch. Right now, I create a branch for a feature, add it as a trunk via repository settings, and start stacking from the feature branch. All going well, except, a co-worker did a few modifications on a stacked PR, and submitted them. This changed the base of the branch to our main trunk branch (develop), and not the feature branch. I think the local repository config I had setup for trunk branches did not sync to his graphite config. Is this a limitation of graphite, or something wrong in our workflow?

When reviewing a PR, is it possible to cmd+click or shift+click to select a range? Currently I'm struggling to select a large code block to comment on