My team has been having trouble using graphite: 1....
# report-bugs
w
My team has been having trouble using graphite: 1. We had a repo
duramata
and renamed it to
duramata-old
. 2. We created a new repo named
duramata
and synced it 3. New stacks fail to load (see image) 4. Everything I've been told over email support has failed We believe there is something wrong on your end with how you're handling this switch. The local graphite state seems to work fine, but when connecting remote it fails. Here is an example of PR #102 failing to retarget a ghost PR #105 that doesn't exist:
Copy code
william@Williams-MacBook-Pro duramata % gt sync --restack
🌲 Fetching branches from remote...
Running git fetch: [========================================] 100% | Done
main fast-forwarded to f4f7ebae7455450f0bea0e908ead71a7afc59ad3.

🔄 Syncing branches...
Branch 05-19-refactor_app_route_sprite_effects_through_react_query is up to date with remote (PR #102 v5).

🥞 Restacking branches...
Restacked 05-19-refactor_app_route_sprite_effects_through_react_query on main.
william@Williams-MacBook-Pro duramata % gt ss
🥞 Validating that this Graphite stack is ready to submit...

📝 Preparing to submit PRs for the following branches...
▸ 05-19-refactor_app_route_sprite_effects_through_react_query (Update code)

ERROR: failed to retarget PR #105: PR not found
william@Williams-MacBook-Pro duramata % gt sync
;🌲 Fetching branches from remote...
^R
main is up to date.

🔄 Syncing branches...

🥞 Restacking branches...
05-19-refactor_app_route_sprite_effects_through_react_query does not need to be restacked on main.
william@Williams-MacBook-Pro duramata % gt restack
05-19-refactor_app_route_sprite_effects_through_react_query does not need to be restacked on main.
william@Williams-MacBook-Pro duramata % gt ss
🥞 Validating that this Graphite stack is ready to submit...

📝 Preparing to submit PRs for the following branches...
▸ 05-19-refactor_app_route_sprite_effects_through_react_query (Update code)

ERROR: failed to retarget PR #105: PR not found
william@Williams-MacBook-Pro duramata %
p
(If applicable) To help us repro, please send the following if you haven't already: - PR links, screenshots, and/or debug output if applicable - Repro steps - Does this happen consistently? Just to you or others on your team also? - Is this a regression? Is the problem new? - If you're having an issue with the CLI, please send us debug logs using
gt feedback --with-debug-context <any message>
w
https://app.graphite.com/github/pr/Mirascope/duramata/102/refactor(app)-route-sprite-effects-through-React-Query?mode=tour Repro steps are to simple create a PR with
gt add -A
and
gt create -m "message"
which results in the issue for us.
Copy code
william@Williams-MacBook-Pro duramata % gt feedback --with-debug-context "consistent retargeting issues on new stacks for duramata"
Uploading log files: [========================================] 100% | Upload complete
Feedback received loud and clear 😊
@Gareth Brickman This is fully blocking our work
g
hey William, we have a lead on the issue that the duplicate name collision is causing in our system and we're working on a solution, will keep you posted
👍 1
w
Ok awesome, thank you for the update!
j
hey william, ran a few scripts on our end. can you re-add your repos for the sync?
👀 1
w
it seemed to be fix for a bit but it's back to not working it seems...
this PR is failing to load the stack
j
had to run one last script, you should be able to see the stack now. lmk if you cant!
w
unfortunately it's still broken
j
can you load a PR and then click
in app feedback
so i can take a look at the logs. from what i can see in https://app.graphite.com/github/pr/Mirascope/duramata/102 the stack is loading
👍 1
here's a new PR I just submitted that isn't loading
Where is in app feedback? I can't seem to find it in any menu
https://app.graphite.com/github/pr/Mirascope/duramata/119 @Jocelyn Beaudoin this one also not loading
And now in the CLI I'm back to getting retargeting issues
Copy code
william@Williams-MacBook-Pro duramata % gt feedback --with-debug-context "(again) consistent retargeting issues on new stacks for duramata"
Uploading log files: [========================================] 100% | Upload complete
Feedback received loud and clear 😊
j
i can see the stack loading for this PR. can you try to fun a
gt sync --force
then
gt track --parent main
for the affected branch and
gt submit --no-restack
?
w
it seems like once they merge something happens to let them render, but before I force merge them with admin privilege I can't see them, problem is consistent @Jocelyn Beaudoin
d
I'm on the same team and my UI is broken and not showing me approved or draft PRs (in addition to often not seeing the stacking) Really hope we can get this fixed b.c. I've come to depend on Graphite but it's really not working anymore.
j
on a PR where youre seeing stacks not loading, can you click the help icon then "Send Feedback" and mark it as a "Bug" so i can take a look at the logs. keyboard short cut would be
?
then
G
then
F
and the modal will open
d
Done, let me know if that made it thru
Our PRs are still broken, getting in this "perpetual loading state" for not displaying the stack
Though the "Could not fetch approved" issue cleared up after I unsynced some stale repos
I'm really sad that graphite is not working for us, do you guys have any ideas on how this can get fixed?
I am also getting errors like this when attempting to submit a stack:
There is no PR 136 (it 404s)
Once I manually published the add_scheduling_to_pocket branch as PR 136,
gt ss
started failing saying it cant find PR 137
@Gareth Brickman / @Jocelyn Beaudoin Continuing to have no stack available on new PRs. Are there any leads on how we can get graphite working again?
g
hey Dandelion, I'm re-investigating this and I suspect we need to do some more surgery in our DB for a sub-set of the new
duramata
repo's PRs to get accurately cached into our system. Will keep you posted
d
ty 🙏🏼
g
okay thanks for hanging in there. I've fixed that PR #139! If you notice the Stack component not loading for any others let me know their number and will whac-a-mole them in our cache too
d
Ok, 139 was created pretty recently though. Were we able to figure out a root cause? When I create new PRs I don't want to need to bother you guys all the time 😛
g
the root cause is there's stale references to the old repo's PRs in our cache for a subset of PRs, and then when ya'll create new PRs they're getting stack associated to the old PRs for some of these. Trying some approaches to try to clean this up broadly, but now I know for sure how to do it per PR at least
d
interesting, the old repo had gotten as far as PR 158 so maybe new prs are hitting entries for the old ones. i'm wondering if you could just pre whack every pr from like 80 to 158
w
I'm totally cool with completely wiping our state with y'all and resyncing the current
duramata
repo.
problem persists, graphite continues to prevent proper work flow
i believe it was PR 143. I used GitHub to manually get it merged.
g
hey all, unfortunately this will impact a subset of PRs as while we wiped the cache and re-associated to the new
duramata
repo our background processes still synced some of the old PRs into the cache in between those operations. Am manually checking and cleaning up #143
#142 and #143 are loading their Stack component again. Am purging the cache manually up to #158. Let me know if any other PRs act up!
👍 1
w
No worries, thank you for helping us resolve this. Will circle back if we run into any more issues. Thanks!
d
I just created PR #159 without any issue! Hopefully we are back up and running now
🤞 1
Been a few days without any issues, seems like this is fixed now. Thank you for your help @Gareth Brickman!
👌 1