fix(studio): honor run settings on initial model load (#7346) by Souravrajvi0 · Pull Request #7351 · unslothai/unsloth

Souravrajvi0 · 2026-07-23T02:44:05Z

Problem

Gear icon, set Context Length or KV Cache Dtype, then click Load model. The first load ignored those values. Context defaulted to 4096 and KV cache defaulted to f16. Users had to reload and click Apply.

Two causes:

NumericValueInput only committed on blur, so Load could run before the typed context reached the parent state.
performLoad read the runtime store instead of the staged selection.config from run settings.

Fix

Add NumericValueInput.commit() to flush valid focused and same-click drafts synchronously.
Preserve Auto context when the field is untouched or contains an invalid draft.
Pass effectiveLoadConfig from model-config-page to onRun.
Recompute the persistence-only Save/Forget path after committing context.
Prefer selection.config for context, KV cache, GPU, and speculative load options.
Keep meta.forceReload on config-page reloads.

Verification

pytest tests/studio/test_model_picker_contracts.py -q
cd studio/frontend && npm run typecheck
cd studio/frontend && npm run build

36 contract tests passed
Typecheck passed
Production frontend build passed
Repository Ruff lint and formatting hooks passed across all files
Real Studio and real GGUF load passed with the Context Length input still focused
/api/inference/load received max_seq_length=4096 on the direct button click
Real Studio flow passed in Chromium, installed Chrome, Firefox, and WebKit
168 isolated component assertions passed across those four engines
Native Linux, macOS, and Windows Actions checks are running

When loading a model from the gear-icon run-settings page, Context Length and KV Cache Dtype were ignored if the user clicked Load before blurring the context field, or before React flushed staged config into the store. - Add NumericValueInput.commit() to flush a focused draft on Load - Pass effectiveLoadConfig from model-config-page to onRun - Prefer selection.config in performLoad for all load knobs - Preserve meta.forceReload from the config-page reload path Fixes unslothai#7346

gemini-code-assist · 2026-07-23T02:44:09Z

Caution

The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased.

Souravrajvi0 · 2026-07-23T02:48:36Z

Manual check done locally (Playwright):

No model loaded → On Device → gear → set KV q8_0 → type context 256 (model native max is 512) → Load model without tabbing out
Intercepted POST /api/inference/load:
- cache_type_kv: q8_0 ✓
- max_seq_length: 256 (not default 512) ✓

Also fixed a follow-up: clicking Load blurs the context field first, so commit() was returning the stale value. Draft is now kept in a ref and flushed even when unfocused.

Clicking Load blurs the context field before handleRun runs, so commit() returned the stale value prop. Keep draft in a ref and parse it even when the input is no longer focused.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 0030839f5f

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

NumericValueInput.commit() now returns null unless the user actually changed the field, so GGUF Load/Save no longer pins the displayed native context into customContextLength when Auto was left untouched.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: e3c776bfcf

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

After a normal blur commit, reset dirtyRef so a later Load cannot replay a stale draftRef when the user changed context via Reset or the slider.

Souravrajvi0 · 2026-07-23T07:31:06Z

Fixed in 78e3f3c (also covered in the consolidated reply on the P1 thread).

P2 — Clear dirty state after blur commits: Valid. onBlur now clears dirtyRef and syncs draftRef after commitDraft(), so a later Reset/slider change is not overwritten by a stale draft when Load calls imperative commit().

Verification: tests/studio/test_model_picker_contracts.py (32 passed, including strengthened dirty/null-commit assertions) + npm run typecheck. Branch is current with main (0 behind).

…i#7346 Lock Codex P1/P2: commit returns null unless dirty, blur clears dirtyRef, and handleRun only promotes a non-null committed context.

Souravrajvi0 · 2026-07-23T07:34:43Z

Ready for review — re-verified locally:

Codex P1 (Auto context) + P2 (dirtyRef after blur) addressed in e3c776bf / 78e3f3c0
Contract tests strengthened in 8bba0f06 (test_initial_load_uses_staged_config_payload)
tests/studio/test_model_picker_contracts.py — 32 passed
npm run typecheck clean
Branch current with main (0 behind)

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 8bba0f06de

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

Blur can commit and clear dirtyRef before Load's onClick; stash that committed value for one imperative commit() so typed context is not lost.

Souravrajvi0 · 2026-07-23T07:58:22Z

Fixed in 94e49be.

P2 — Preserve same-click draft after blur: Valid. Blur committed via onChange and cleared dirtyRef before Load's onClick, so imperative commit() returned null while parent config was still stale.

Fix: Keep lastBlurCommittedRef for one same-turn commit() after blur. Cleared on focus, typing, Escape, and when the value prop catches up (Reset/slider), so stale drafts still cannot replay later.

Tests: test_model_picker_contracts.py — 32 passed + npm run typecheck.

Souravrajvi0 · 2026-07-23T07:59:49Z

Temporarily closing to force GitHub to resync head SHA (fork tip is 03e68e1; PR was stuck on 8bba0f0).

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 03e68e11c9

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

…-config-4af7

Souravrajvi0 · 2026-07-23T09:16:52Z

Thanks @shimmyshimmer for picking up the latest Codex round on 40dcf714 the invalid-blur-draft guard and the persistence-only handleRun fix look right, and the added contract coverage is helpful.

Synced with your commits on feat/fix-initial-load-config-4af7.

danielhanchen · 2026-07-23T09:58:44Z

@codex review

chatgpt-codex-connector · 2026-07-23T10:01:00Z

Codex Review: Didn't find any major issues. Breezy!

Reviewed commit: a2bf393257

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

danielhanchen · 2026-07-23T18:07:02Z

Reviewed and verified. Run settings were ignored on the initial model load, so the first load did not honor the configured settings. This applies the run settings on the initial load so the first load matches subsequent ones. Confirmed the settings are applied on first load and simulated the load path. Codex review has no open items. The remaining red CI is pre-existing upstream breakage unrelated to this change. Good to merge.

…51head

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 0346dae006

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-07-24T07:45:08Z

+  useEffect(() => {
+    lastBlurCommittedRef.current = null;
+  }, [value]);


Clear cached blur commits when Reset changes configuration

When a user explicitly enters the currently displayed native/loaded context (for example, 32768), blurs the field, then clicks Reset, value remains 32768 even though customContextLength changes back to null. This effect therefore does not clear lastBlurCommittedRef; a later Load/Save returns that stale blur value and recreates the explicit context pin that Reset was meant to remove. Clear this cache on configuration reset/external edits independently of a displayed-value change.

Useful? React with 👍 / 👎.

Good catch. Typing the value already shown (for example the native 32768) commits no onChange because final === value, yet the blur handler still cached it in lastBlurCommittedRef. Since value never changes, the useEffect clear never fires, so a later Reset or external edit that leaves the displayed value unchanged cannot drop it, and the next commit() replays the stale value into an override. Fixed by only caching the blur result when it actually dispatched onChange (final !== value); when final === value the parent is already current and there is nothing to bridge. Added a Playwright regression that re-types the shown context and asserts no override is stored.

…own value NumericValueInput cached every blur commit in lastBlurCommittedRef, even when the draft equalled the current value and no onChange was dispatched. Because the displayed value never changed, the useEffect([value]) clear never fired, so a later Reset or external edit that leaves the shown value unchanged could not drop the cache and the next commit() replayed it into an override that Reset had removed. Only cache the blur result when it actually dispatched onChange (final !== value); when final === value the parent is already current and there is nothing to bridge. Add a Playwright regression that re-types the shown context and asserts no override is stored.

danielhanchen · 2026-07-24T08:39:06Z

@codex review

for more information, see https://pre-commit.ci

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 27a2886bbc

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector · 2026-07-24T08:49:36Z

+    const committedContext = target.isGguf
+      ? contextInputRef.current?.commit()
+      : undefined;


Commit every numeric draft before building load config

When the user edits another NumericValueInput-backed setting and immediately clicks Load/Reload—such as non-GGUF Max Seq Length or GGUF GPU Layers/MoE Layers—the draft is still only flushed by that input's blur handler, but this path synchronously commits only the GGUF context ref before building effectiveRuntimeConfig. In the same-click path the parent config can still contain the old numeric value, so the staged /load request ignores the value the user just typed; wire the imperative commit through the other numeric settings too, or otherwise wait for their blur commits before constructing the load config.

Useful? React with 👍 / 👎.

Good catch. The imperative same-click commit only covered the GGUF Context Length input, so typing into Max Seq Length (non-GGUF) or GPU Layers / MoE Layers on CPU (GGUF) and clicking Load/Reload in the same gesture staged the load from a stale config. Wired a commit handle through those inputs too and fold every committed draft into the effective config, recomputing the non-GGUF load-time max sequence length from the committed draft. Fixed in 6b0842f.

… config The run-settings Load/Reload button flushed only the GGUF Context Length draft imperatively before building the load config. Max Seq Length (non-GGUF), GPU Layers and MoE Layers on CPU (GGUF) are the same NumericValueInput and stage their typed value only on blur, so editing one and clicking Load in the same gesture staged the load from a still-stale parent config and dropped the value the user just typed. Wire an imperative commit handle through those inputs too and fold every committed draft into the effective config, recomputing the non-GGUF load-time max sequence length from the committed draft.

danielhanchen · 2026-07-24T09:05:43Z

@codex review

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 6b0842fc9c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-07-24T09:09:06Z

+            // Only bridge the still-stale parent value when the blur actually
+            // dispatched onChange (final !== value). When final === value the
+            // parent is already current, so there is nothing to bridge; caching
+            // here would leave a stale pin that a later Reset or external edit
+            // (which doesn't change the displayed value) can never clear, so a
+            // following Load/Save would recreate the override Reset removed.
+            lastBlurCommittedRef.current = final !== value ? final : null;


Clear the blur cache when Reset restores an unchanged value

If a different setting makes Reset enabled while this numeric field is still showing its default (for example, a GGUF with a custom KV cache dtype), edit the field and click Reset without first blurring it. The blur stores the edited value here, but Reset restores the same numeric value it had before, so the [value] effect does not run to clear the cache. The subsequent Load/Save consumes lastBlurCommittedRef and silently reintroduces the numeric override that Reset was meant to remove.

Useful? React with 👍 / 👎.

Confirmed. The prior fix only skipped caching when final === value; a real edit (final !== value) that Reset then reverts to the same shown number nets value back unchanged, so the useEffect([value]) clear never re-ran and the stale pin survived into the next Load/Save. The bridge is only valid across the single synchronous same-click gesture that set it (blur commits on the button mousedown, the button onClick consumes it via commit() before React re-renders), so the clear now runs on every settled render instead of only on a value change. Any settled render means the gesture is over. Added a source-contract regression (test_blur_cache_cleared_on_every_settled_render).

chatgpt-codex-connector · 2026-07-24T09:09:06Z

+    const effectiveRuntimeConfig = hasPending
+      ? { ...runtimeConfig, ...pendingPatch }
+      : runtimeConfig;


Recompute the fixed-layer context pin after committing GPU Layers

When an already auto-fit GGUF has a smaller activeLoadedContext, selecting Manual, typing a positive GPU Layers value, enabling Remember, and immediately clicking Reload builds runtimeConfig before the committed layer value is merged. It therefore saves gpuLayers with customContextLength: null instead of pinning the fitted context; a later fresh load sends the native context with fixed layers, which the surrounding policy explicitly identifies as an OOM risk.

Useful? React with 👍 / 👎.

Good catch. pinFixedLayerContext was computed from the render-time config, before the same-click GPU Layers draft is committed, so typing a positive fixed-layer value and clicking Reload built runtimeConfig with customContextLength: null. handleRun now recomputes the pin from effectiveConfig (effectivePinFixedLayerContext) after folding in the committed drafts, so a committed fixed-layer value still pins activeLoadedContext and a later fresh load keeps the fitted placement instead of sending the native context. Added a source-contract regression (test_fixed_layer_pin_recomputed_after_committing_gpu_layers).

…che on every render Two run-settings edge cases on the model-config page: 1) pinFixedLayerContext was computed from the render-time config, before a same-click GPU Layers draft is committed in handleRun. Typing a positive fixed-layer value on an auto-fit GGUF and clicking Reload therefore built the runtime config with customContextLength: null, so a later fresh load sent the native context with fixed layers (the OOM the pin exists to avoid). Recompute the pin from the committed effectiveConfig. 2) NumericValueInput cleared its blur bridge only on a value change. A real edit (final !== value) that Reset then reverts to the same shown number nets value back unchanged, so the effect never re-ran and the stale pin survived into the next Load/Save, replaying the override Reset removed. The bridge is only valid across the single synchronous same-click gesture that set it, so clear it on every settled render instead. Add source-contract regressions for both.

danielhanchen · 2026-07-24T09:33:42Z

@codex review

for more information, see https://pre-commit.ci

danielhanchen · 2026-07-24T09:36:57Z

@codex review

danielhanchen · 2026-07-24T09:37:09Z

@codex review

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 45e08d1508

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-07-24T09:39:59Z

+              loadCustomContextLength =
+                pendingLoadConfig?.customContextLength ?? null;
+              loadSelectedGpuIds =
+                pendingLoadConfig?.selectedGpuIds !== undefined
+                  ? reconcilePersistedGpuIds(pendingLoadConfig.selectedGpuIds)
+                  : null;
+              loadGpuLayers = pendingLoadConfig?.gpuLayers ?? GPU_LAYERS_AUTO;


Validate staged GGUF context before unloading

When switching from an already-loaded model to a GGUF through Run settings with a staged customContextLength, this branch restores that staged context only after the validation request has been built. The validation path therefore uses validateCustomContextLength = null for a cross-model switch, while the subsequent /load sends the explicit context. In particular during an active training/coexistence check, validation can pass with auto sizing, then the frontend unloads the working model and /load rejects the larger staged context before rolling back. Keep the staged context in the validation inputs when pendingLoadConfig supplies it so both requests evaluate the same load.

Useful? React with 👍 / 👎.

A staged Run settings config can't reach validation with a null context here. onRun stages the picked config into selection.config, which flows to pendingLoadConfig. Any non-null config makes applyModelLoadConfigToRuntime return true, which sets keepSpeculative: true, so resetsPerModelSettings is false. With resetsPerModelSettings false, validateCustomContextLength takes loadCustomContextLength (the staged pendingLoadConfig.customContextLength), not null, so validation and the follow-up load size against the same context. The reset block that only re-applies the staged context later runs solely when keepSpeculative is false, and in that branch selection.config (pendingLoadConfig) is undefined, so there is no staged context to lose. The two requests already evaluate the same load.

danielhanchen · 2026-07-24T10:01:00Z

@codex review

chatgpt-codex-connector · 2026-07-24T10:03:16Z

Codex Review: Didn't find any major issues. Bravo.

Reviewed commit: 45e08d1508

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Souravrajvi0 requested a review from danielhanchen as a code owner July 23, 2026 02:44

fix(studio): flush NumericValueInput draft when Load blurs first

e6bcf10

Clicking Load blurs the context field before handleRun runs, so commit() returned the stale value prop. Keep draft in a ref and parse it even when the input is no longer focused.

chatgpt-codex-connector Bot reviewed Jul 23, 2026

View reviewed changes

Comment thread studio/frontend/src/features/model-picker/components/model-config-page.tsx Outdated

fix(studio): preserve Auto context when Load is clicked without edits

e3c776b

NumericValueInput.commit() now returns null unless the user actually changed the field, so GGUF Load/Save no longer pins the displayed native context into customContextLength when Auto was left untouched.

chatgpt-codex-connector Bot reviewed Jul 23, 2026

View reviewed changes

Comment thread studio/frontend/src/features/model-picker/components/numeric-value-input.tsx Outdated

fix(studio): clear NumericValueInput dirty state after blur commit

78e3f3c

After a normal blur commit, reset dirtyRef so a later Load cannot replay a stale draftRef when the user changed context via Reset or the slider.

test(studio): pin NumericValueInput Auto/dirty contracts for unslotha…

8bba0f0

…i#7346 Lock Codex P1/P2: commit returns null unless dirty, blur clears dirtyRef, and handleRun only promotes a non-null committed context.

chatgpt-codex-connector Bot reviewed Jul 23, 2026

View reviewed changes

Comment thread studio/frontend/src/features/model-picker/components/numeric-value-input.tsx Outdated

fix(studio): keep same-click context draft after blur (unslothai#7346)

94e49be

Blur can commit and clear dirtyRef before Load's onClick; stash that committed value for one imperative commit() so typed context is not lost.

chore: refresh PR head for unslothai#7351

03e68e1

Souravrajvi0 closed this Jul 23, 2026

Souravrajvi0 reopened this Jul 23, 2026

chatgpt-codex-connector Bot reviewed Jul 23, 2026

View reviewed changes

Comment thread studio/frontend/src/features/model-picker/components/numeric-value-input.tsx Outdated

Comment thread studio/frontend/src/features/model-picker/components/model-config-page.tsx Outdated

Unsloth and others added 2 commits July 23, 2026 01:34

fix(studio): handle context commit edge cases

e79d414

chore: refresh PR head

2378e47

shimmyshimmer force-pushed the feat/fix-initial-load-config-4af7 branch from 23462f3 to 2378e47 Compare July 23, 2026 08:39

shimmyshimmer added 2 commits July 23, 2026 01:41

test(studio): guard invalid context drafts

02cfb4a

style(studio): format context draft guard

6d1a597

shimmyshimmer force-pushed the feat/fix-initial-load-config-4af7 branch from e79d414 to 3f0453b Compare July 23, 2026 08:55

shimmyshimmer added 2 commits July 23, 2026 02:13

test(studio): exercise same-click model config loads

40dcf71

Merge remote-tracking branch 'origin/main' into feat/fix-initial-load…

a2bf393

…-config-4af7

Merge branch 'main' of https://github.com/unslothai/unsloth into pr73…

0346dae

…51head

chatgpt-codex-connector Bot reviewed Jul 24, 2026

View reviewed changes

[pre-commit.ci] auto fixes from pre-commit.com hooks

27a2886

for more information, see https://pre-commit.ci

chatgpt-codex-connector Bot reviewed Jul 24, 2026

View reviewed changes

[pre-commit.ci] auto fixes from pre-commit.com hooks

45e08d1

for more information, see https://pre-commit.ci

chatgpt-codex-connector Bot reviewed Jul 24, 2026

View reviewed changes

Uh oh!

Uh oh!

Conversation

Souravrajvi0 commented Jul 23, 2026 • edited by shimmyshimmer Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Problem

Fix

Verification

Uh oh!

gemini-code-assist Bot commented Jul 23, 2026

Uh oh!

Souravrajvi0 commented Jul 23, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Souravrajvi0 commented Jul 23, 2026

Uh oh!

Souravrajvi0 commented Jul 23, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Souravrajvi0 commented Jul 23, 2026

Uh oh!

Souravrajvi0 commented Jul 23, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

Souravrajvi0 commented Jul 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

danielhanchen commented Jul 23, 2026

Uh oh!

chatgpt-codex-connector Bot commented Jul 23, 2026

Uh oh!

danielhanchen commented Jul 23, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Jul 24, 2026

Choose a reason for hiding this comment

Uh oh!

danielhanchen Jul 24, 2026

Choose a reason for hiding this comment

Uh oh!

danielhanchen commented Jul 24, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Jul 24, 2026

Choose a reason for hiding this comment

Uh oh!

danielhanchen Jul 24, 2026

Choose a reason for hiding this comment

Uh oh!

danielhanchen commented Jul 24, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Jul 24, 2026

Souravrajvi0 commented Jul 23, 2026 •

edited by shimmyshimmer

Loading

Souravrajvi0 commented Jul 23, 2026 •

edited

Loading