fix(cli): context length tweaks and terminal output truncation #9467
Conversation
Learn moreAll Green is an AI agent that automatically: ✅ Addresses code review comments ✅ Fixes failing CI checks ✅ Resolves merge conflicts |
1 similar comment
Learn moreAll Green is an AI agent that automatically: ✅ Addresses code review comments ✅ Fixes failing CI checks ✅ Resolves merge conflicts |
|
✅ Review Complete Code Review SummaryTroubleshooting
|
There was a problem hiding this comment.
1 issue found across 11 files (changes from recent commits).
Prompt for AI agents (all issues)
Check if these issues are valid — if so, understand the root cause of each and fix them.
<file name="extensions/cli/src/util/tokenizer.ts">
<violation number="1" location="extensions/cli/src/util/tokenizer.ts:227">
P2: Empty enum array (`enum: []`) would subtract 3 tokens without adding any back, potentially resulting in negative token count. Consider adding a guard for empty arrays or ensuring `tokens` doesn't go negative.</violation>
</file>
Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.
| } | ||
|
|
||
| if (fieldEnum && Array.isArray(fieldEnum)) { | ||
| tokens -= 3; |
There was a problem hiding this comment.
P2: Empty enum array (enum: []) would subtract 3 tokens without adding any back, potentially resulting in negative token count. Consider adding a guard for empty arrays or ensuring tokens doesn't go negative.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At extensions/cli/src/util/tokenizer.ts, line 227:
<comment>Empty enum array (`enum: []`) would subtract 3 tokens without adding any back, potentially resulting in negative token count. Consider adding a guard for empty arrays or ensuring `tokens` doesn't go negative.</comment>
<file context>
@@ -196,16 +197,158 @@ export function calculateContextUsagePercentage(
+ }
+
+ if (fieldEnum && Array.isArray(fieldEnum)) {
+ tokens -= 3;
+ for (const e of fieldEnum) {
+ tokens += 3;
</file context>
| tokens -= 3; | |
| if (fieldEnum.length === 0) { | |
| return tokens; | |
| } | |
| tokens -= 3; |
✅ Addressed in d7b91f8
|
🎉 This PR is included in version 1.38.0 🎉 The release is available on: Your semantic-release bot 📦🚀 |
|
🎉 This PR is included in version 1.37.0 🎉 The release is available on: Your semantic-release bot 📦🚀 |
Description
We already are very conservative about when to auto-compact, since we leave 35% or max tokens and then 80% of that, so compaction usually happens around 50% of actual context length. So it seems the only major cause of the context length exceeded issue is huge tool outputs.
Environment Variables
CONTINUE_CLI_BASH_MAX_OUTPUT_CHARSCONTINUE_CLI_BASH_MAX_OUTPUT_LINESCONTINUE_CLI_READ_FILE_MAX_OUTPUT_CHARSCONTINUE_CLI_READ_FILE_MAX_OUTPUT_LINESCONTINUE_CLI_FETCH_MAX_OUTPUT_LENGTHCONTINUE_CLI_DIFF_MAX_OUTPUT_LENGTHCONTINUE_CLI_SEARCH_CODE_MAX_RESULTSCONTINUE_CLI_SEARCH_CODE_MAX_RESULT_CHARSpartially a remake of #8122
Continue Tasks
Powered by Continue
Summary by cubic
Improves CLI output truncation and context length checks to prevent overflows. Keeps the most relevant output, counts system/tool tokens in auto-compaction, and adds configurable limits per tool.
New Features
Refactors
Written for commit d7b91f8. Summary will update on new commits.