Repository navigation
Conversation
cgivre
force-pushed
the
feature/sqllab-react-ui
branch
from
March 9, 2026 13:45
080ba0b to
36af30f
Compare
cgivre
force-pushed
the
feature/sqllab-react-ui
branch
2 times, most recently
from
March 25, 2026 13:50
33fb68e to
34bfc23
Compare
cgivre
force-pushed
the
feature/sqllab-react-ui
branch
from
June 16, 2026 14:20
1b68007 to
bc4e985
Compare
cgivre
force-pushed
the
feature/sqllab-react-ui
branch
2 times, most recently
from
June 25, 2026 17:38
ff3d719 to
dc96869
Compare
cgivre
force-pushed
the
feature/sqllab-react-ui
branch
from
July 30, 2026 18:12
dfde1d3 to
5d16fd0
Compare
cgivre
force-pushed
the
feature/sqllab-react-ui
branch
from
August 10, 2026 20:26
1cdcb53 to
09690cf
Compare
cgivre
force-pushed
the
feature/sqllab-react-ui
branch
from
August 20, 2026 11:00
3aac985 to
65f6b74
Compare
cgivre
force-pushed
the
feature/sqllab-react-ui
branch
from
August 27, 2026 16:24
1ffabfe to
fbcc239
Compare
…lti-file drop once
Return a queryable table path from the upload endpoint, reuse the scanner's quoting for it, and stop the two consumers interpolating a folder path that is not valid Drill SQL. - Add UploadResponse.tablePath, built by ProjectSchemaCache.quoteTablePath. A project id is a UUID, so the "schema" field it replaces does not parse. Use it in the Prospector message and in the upload toast, type the response properly in api/uploads.ts, and replace the mocked response shapes in the tests with the one the server sends. - Match a folder dataset reference on type, schema and table, so a repointed upload target does not leave every project orphaned. - Report refusals through getErrorMessage, so the 403 and the 413 reach the user instead of "Request failed with status code 403". - Append the download anchor to the document before clicking it; Firefox will not follow a click on a detached one. - Give fileScan a 60 second wall-clock budget and collapse a herd of stale readers to one sweep, so a slow folder cannot hold the schema cache's monitor for minutes. - Re-check that the configured workspace is writable at upload time; "writable" is a planner concept, not a filesystem permission. - Run the project id through the containment guard every file name passes: a scratch project's id carries the owner's principal name. - Document who can read an uploaded file: the tree filter is cosmetic, and real isolation needs impersonation plus storage permissions.
The project id went through the containment guard in the previous commit, and that guard appended a separator to a folder URI that already ended in one. A workspace located at "/" therefore refused every id, so a bucket given over to uploads could not be used at all. - Append the separator only when the folder does not already end in one, and require the resolved path to be strictly longer than the folder, so a name resolving back onto a root folder is still refused. - Start the file scan's budget once per sweep in combinedScan and share the deadline across every folder, rather than once per folder; a project now accumulates a folder reference per target it has been pointed at. - Document the project id guard, including that a principal carrying a backslash now refuses a scratch project's uploads, and that a workspace at the storage root is supported.
Refuse a model file whose bytes declare a DOCTYPE before pmml4s parses it. pmml4s builds its own StAX factory with default settings and exposes no hook on it; on Drill's classpath that factory resolves external entities and external DTD subsets, so model content could read local files, probe absolute paths through error messages, and make the Drillbit fetch a URL. File uploads land arbitrary user bytes in a workspace by design, so that content is untrusted. Setting the StAX hardening properties is not enough, because Woodstox ignores a disabled DOCTYPE silently and pmml4s re-parses the bytes with its own factory anyway. Read the model as the query user when impersonation is enabled, so predict() cannot open a file the user could not select. Treat an unset feature reader as a missing value and pass null to pmml4s, which applies the model's own missing-value handling. Previously a NULL in any column other than FLOAT8 failed the query with a NullPointerException. Also refuse a climbing segment left in place by a workspace at the storage root, and compare a URI-form workspace location against the form that resolving a name against it produces.
Pass a real UsageObserver to the summarize call and write one AiEvent under the feature workflow_ai_summary, against the workflow's owner. It was the only provider call site in the codebase that recorded nothing, so a scheduled run's tokens were invisible to /ai-analytics and to cost accounting. Name the ObjectMapper on every AI config store call. Falling back to a new ObjectMapper let whichever caller warmed the cache fix the mapper for the whole process.
Add ai_summary to the create_workflow tool's action-type enum and description, and describe it in the activation dialog. Its line says the run's columns and rows go to the AI provider, since that is what an owner needs to know before switching the workflow on. Label the new workflow_ai_summary feature slug so the analytics dashboard can select it.
Rewrite the security section of the machine learning doc: paths that resolve outside the workspace are refused rather than all absolute paths, confinement is lexical with no symlink resolution, model reads happen as the query user under impersonation, and model content is screened for a DOCTYPE. Add the supported feature types and what mapFields skips. Describe ai_summary in the workflows guide, including that AI data sharing defaults to on, that up to 20 rows and every column name go to the provider while it is on, the 60 second action timeout, and that row values entering the prompt are untrusted.
cgivre
force-pushed
the
feature/sqllab-react-ui
branch
from
November 13, 2026 04:19
77bb264 to
7dfc440
Compare
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
DRILL-XXXX: Refactor User Interface
Description
This PR refactors Drill's UI and refactors the Query view, adds visualizations and dashboards and in general makes Drill much more user friendly.
Documentation
(Please describe user-visible changes similar to what should appear in the Drill documentation.)
Testing
(Please describe how this PR has been tested.)