Tags: posit-dev/positron
Tags
use correct reh/reh-web package.json in bundle (#6028) This change addresses an issue in which Positron Server doesn't start due to an error like this one: ``` (node:11206) Warning: To load an ES module, set "type": "module" in the package.json or use the .mjs extension. (Use `node --trace-warnings ...` to show where the warning was created) /home/ubuntu/vscode-reh-web-linux-x64/out/server-main.js:26 export function __extends(d, b) { ^^^^^^ SyntaxError: Unexpected token 'export' at wrapSafe (node:internal/modules/cjs/loader:1378:20) at Module._compile ``` The problem was caused by an incompletely merged line in the initial 1.95 merge, which caused a cascade of silent failures that led to `package.json` being omitted from the `reh-web` bundle, which led to the above failure since `package.json` was responsible for setting `"type": "module"` as noted above. The fix is just to update the line to match what it was before the merge: https://github.com/posit-dev/positron/blob/fa567b23598d7f91eab1f2aa182ae8b1e8b50099/build/gulpfile.reh.js#L378-L383 Addresses, coincidentally, #5775, which is where this error was first spotted. After the change, it ought to be possible to run `./positron-server` without error. E.g.: ``` ./positron-server --version Positron: 2025.01.1 build 4 Positron SHA: 9f520e6 Code OSS: 1.96.0 Arch: arm64 ```
Merge 1.96 from upstream into Positron (#6007) This change is identical to #5981, except that it is targeted at `main` rather than the patch branch. ### Release Notes #### New Features - Update to [Version 1.96](https://code.visualstudio.com/updates/v1_96) of upstream Code OSS project #### Bug Fixes - N/A ### QA Notes This merge is much less involved than 1.95 was and mostly adds new features rather than refactoring or changing existing ones (see release notes linked above for a summary). --------- Co-authored-by: positron-bot[bot] <173392469+positron-bot[bot]@users.noreply.github.com> Co-authored-by: sharon wang <sharon-wang@users.noreply.github.com>
Data Explorer: Always read fresh file contents when ingesting a file … …into duckdb (#5890) Addresses #5860. duckdb-wasm has some kind of file caching behavior that is not resilient to file changes on the underlying file system. This patch nukes this issue from orbit by always registering the file contents as a virtual file and reading from that. I think if and when we migrate to the duckdb NodeJS bindings per #5889 that we can avoid this chicanery. This does the same thing for Parquet files, too, since they would also become stale. e2e: @:data-explorer ### QA Notes Here was the reproducible failure from the issue: * Run iris |> readr::write_csv("test.csv") * Click on the file in the File Explorer to open it up; note that we correctly see the iris data * Run mtcars |> readr::write_csv("test.csv") * Again click on the file in the File Explorer to open it up; note that we incorrectly still see the iris data
Refactor file/folder creation modals to be more platform-aware (#5669) - Addresses #5616 - Original PR #5657 ### Implementation - update the New Folder, New Folder from Git, Project Wizard and Save Plot modals with the file path fixes - remove uses of `fsPath` -- the file system path seems to be formatted based on the client file system rather than the server file system - leverage `IPathService` and `ILabelService` to manage paths -- these services are aware of the remote server platform - create new path utils file `src/vs/workbench/browser/utils/path.ts` to convert to and from path strings and path URIs in a platform-aware way - fix up folder checks in the new project initialization tasks - pass folder authority in addition to path and scheme for a more complete URI object ### QA Notes #### Test areas - New Folder modal - New Folder from Git modal - Project Wizard modal - Test with "Initialize project as Git repository" enabled and disabled - Save Plot modal For each modal, both the <kbd>Browse...</kbd> button method of selecting a parent directory and typing in a path should be tested. I have tested on the following platforms: - Mac: Desktop, Server Web, Workbench via Arc browser - Windows: Desktop, Server Web, Workbench via Chrome/Edge browsers I have not yet tested: - Linux: Desktop, Server Web, Workbench via a browser - Remote SSH from any client OS Workbench via a browser on Windows should be fixed with this PR. All other platform configurations should not have regressed. --------- Co-authored-by: sharon wang <25834218+sharon-wang@users.noreply.github.com>
e2e-test: browser in parallel & enable Currents (#5442) ### Summary This PR enables the Currents dashboard integration to evaluate the tool’s usefulness. Additionally, the browser is now configured to run on multiple ports, allowing browser tests to execute in parallel. I also consolidated matrix runs into a single machine and enabled multiple threads/workers. This adjustment seems to align better with the current setup and is likely more cost-efficient. ### QA Notes I performed numerous test runs to monitor the potential for increased flakiness with parallel workers. The results are promising overall. While I observed some minor flakiness, these issues appear manageable and can be addressed with further adjustments.
e2e-test: browser in parallel & enable Currents (#5442) ### Summary This PR enables the Currents dashboard integration to evaluate the tool’s usefulness. Additionally, the browser is now configured to run on multiple ports, allowing browser tests to execute in parallel. I also consolidated matrix runs into a single machine and enabled multiple threads/workers. This adjustment seems to align better with the current setup and is likely more cost-efficient. ### QA Notes I performed numerous test runs to monitor the potential for increased flakiness with parallel workers. The results are promising overall. While I observed some minor flakiness, these issues appear manageable and can be addressed with further adjustments.
e2e-test: add savePlot (#5479) ### Intent The goal of this PR was to improve the test by removing a hardcoded wait that was previously required for the test to work and pass. I was able to make improvements to the test and it runs/passes but occasionally I kept running into a flake of RPC timeout on CI runs for the python plot test. Maybe that was the original issue? (The in-line comment didn't say). Adding the wait back in does seem to remove the flake, so maybe that was it after-all. Regardless, there are still improvements worth merging in. ### Approach A new method, savePlot, was introduced to encapsulate and enhance the logic for saving a plot to a selected file type. Additionally, some before/after hooks were cleaned up for better readability. ### QA Notes * Confirmed plots test [PASS on Windows](https://d38p2avprg8il3.cloudfront.net/playwright-report-12008951342/index.html#?q=plots). * I did not see an open issue for the RPC timeout? If we don't have one, I can add it and tag it to this test.
Fix R restart test with new Kallichore build (#5387) Addresses #5356; fixes restarts on R and re-enables the R console tests. The bug was upstream, in the supervisor; the issue was that during a restart, the channel that broadcast status updates to the client would blink offline. If the 'starting' status got sent during that blink, the client would miss it and not detect startup correctly. The fix is to ensure the channel always broadcasts updates throughout the restart cycle. ### QA Notes As noted by @jonvanausdeln, this is 100% reproducible on Windows, so it should be easy to test the fix there.
PreviousNext