Commits

* 6.3: [HttpKernel] conflict with VarDumper < 6.3 [DependencyInjection] Revert "Use weak references in the container" [AssetMapper] Using a separate logger channel [Console] Fix PHP Doc of InputArgument

* 6.2: [Console] Fix PHP Doc of InputArgument

* 5.4: [Console] Fix PHP Doc of InputArgument

…ontainer" (nicolas-grekas) This PR was merged into the 6.3 branch. Discussion ---------- [DependencyInjection] Revert "Use weak references in the container" | Q | A | ------------- | --- | Branch? | 6.3 | Bug fix? | yes | New feature? | no | Deprecations? | no | Tickets | Fix #50439 | License | MIT | Doc PR | - This is the second time this issue is reported, let's revert PR #48469 and #49684 before it does more harm :) Commits ------- d6ea474 [DependencyInjection] Revert "Use weak references in the container"

This PR was merged into the 6.3 branch. Discussion ---------- [HttpKernel] conflict with VarDumper < 6.3 | Q | A | ------------- | --- | Branch? | 6.3 | Bug fix? | yes | New feature? | no | Deprecations? | no | Tickets | - | License | MIT | Doc PR | - Required to let `DumpDataCollector` use `Data::getContext()`. Commits ------- 09a08cb [HttpKernel] conflict with VarDumper < 6.3

This PR was merged into the 6.3 branch. Discussion ---------- [AssetMapper] Using a separate logger channel | Q | A | ------------- | --- | Branch? | 6.3 | Bug fix? | yes-ish | New feature? | yes-ish | Deprecations? | no | Tickets | None | License | MIT | Doc PR | Still TODO Super minor - putting these in their own logging channel might be useful. For example, we could (later, or in a 3rd party bundle) grab all of these warnings and write them out as JavaScript console warnings to make it easier for the user to see (in dev only). Cheers! Commits ------- 9f31576 [AssetMapper] Using a separate logger channel

This PR was merged into the 5.4 branch. Discussion ---------- [Console] Fix PHP Doc of InputArgument | Q | A | ------------- | --- | Branch? | 5.4 | Bug fix? | no | New feature? | no | Deprecations? | no | Tickets | | License | MIT | Doc PR | Commits ------- f09d08c [Console] Fix PHP Doc of InputArgument

* 6.3: [5.4] Allow PhpUnitBridge v7 [Console] block input stream if needed When the input stream used in the question helper is not blocking, the default value is always used as the stream return false. In order to fix that, we force the stream to be in blocking state and go back to the old state after so other logic is not impacted by this change

* 6.2: [5.4] Allow PhpUnitBridge v7 [Console] block input stream if needed When the input stream used in the question helper is not blocking, the default value is always used as the stream return false. In order to fix that, we force the stream to be in blocking state and go back to the old state after so other logic is not impacted by this change

* 5.4: [5.4] Allow PhpUnitBridge v7 [Console] block input stream if needed When the input stream used in the question helper is not blocking, the default value is always used as the stream return false. In order to fix that, we force the stream to be in blocking state and go back to the old state after so other logic is not impacted by this change

This PR was merged into the 5.4 branch. Discussion ---------- [5.4] Allow PhpUnitBridge v7 | Q | A | ------------- | --- | Branch? | 5.4 | Bug fix? | no | New feature? | no | Deprecations? | no | Tickets | - | License | MIT | Doc PR | - Commits ------- 044bd02 [5.4] Allow PhpUnitBridge v7

This PR was merged into the 5.4 branch. Discussion ---------- [Console] block input stream if needed | Q | A | ------------- | --- | Branch? | 5.4 | Bug fix? | yes | New feature? | no | Deprecations? | n | Tickets | / | License | MIT | Doc PR | / When the input stream used in the question helper is not blocking, the default value is always used as the stream return false instead of waiting. In order to fix that, we force the stream to be in blocking state and go back to the old state after so other logic is not impacted by this change. I don't think making a test for that is possible without writing a lot of codes, as we need to have a stream that is not ready for this test, and make it ready later, given there is no async possible. it could be done with a fork or another process writing to a file **after** the question was asked so we do not block following tests when there is no failure, however i'm not sure if this is wanted ?. If you have a clean way to make a reproducible test for this bug i'm open to write it. Commits ------- 8ae36f3 [Console] block input stream if needed When the input stream used in the question helper is not blocking, the default value is always used as the stream return false. In order to fix that, we force the stream to be in blocking state and go back to the old state after so other logic is not impacted by this change

* 6.3: [VarDumper][HttpKernel] Fix dumping with labels [VarDumper] Fix failures on appveyor Fix: Typo in PHPDoc

This PR was merged into the 6.3 branch. Discussion ---------- [VarDumper] Fix failures on appveyor | Q | A | ------------- | --- | Branch? | 6.3 | Bug fix? | yes | New feature? | no | Deprecations? | no | Tickets | - | License | MIT | Doc PR | - Commits ------- 52f45a4 [VarDumper] Fix failures on appveyor

…rekas) This PR was merged into the 6.3 branch. Discussion ---------- [HttpKernel][VarDumper] Fix dumping with labels | Q | A | ------------- | --- | Branch? | 6.3 | Bug fix? | yes | New feature? | no | Deprecations? | no | Tickets | - | License | MIT | Doc PR | - Should fix #50347 (comment) /cc `@bobthecow` Also improves the display in the WDT: ![image](https://github.com/symfony/symfony/assets/243674/956f7c7d-5569-4ea2-97f3-01d000f34532) ![image](https://github.com/symfony/symfony/assets/243674/6a497d1e-5291-45a7-8af3-1cf11cafded6) Commits ------- e824eb0 [VarDumper][HttpKernel] Fix dumping with labels

This PR was merged into the 6.3 branch. Discussion ---------- [Messenger] Fix: Typo in PHPDoc | Q | A | ------------- | --- | Branch? | 6.3 | Bug fix? | no | New feature? | no | Deprecations? | no | Tickets | - | License | MIT | Doc PR | - Commits ------- 743aff9 Fix: Typo in PHPDoc

When the input stream used in the question helper is not blocking, the default value is always used as the stream return false. In order to fix that, we force the stream to be in blocking state and go back to the old state after so other logic is not impacted by this change

* 6.3: [AssetMapper] Fix wiring resolvers, send requests in parallel and use readonly properties in MappedAsset [AssetMapper] Change default importmap "provider" to JsDelivr+esm

…ivr+esm (weaverryan, nicolas-grekas) This PR was merged into the 6.3 branch. Discussion ---------- [AssetMapper] Change default importmap "provider" to JsDelivr+esm | Q | A | ------------- | --- | Branch? | 6.3 | Bug fix? | yes | New feature? | yes | Deprecations? | no | Tickets | None | License | MIT | Doc PR |Still TODO We currently use the https://jspm.org/ API in `importmap:require` to find a CDN URL for an npm package - just like Rails. Unfortunately, this is NOT as robust as we had thought. For me, it's broken. 3 big issues: A) **Not Combined** Some packages are not packed/combined. Example: [chart.js/auto](https://ga.jspm.io/npm:chart.js@4.3.0/auto/auto.js) imports other packages and results in 3 requests instead of 1. Not TERRIBLE... so here IS a terrible example: [`@popperjs`/core](https://ga.jspm.io/npm:`@popperjs`/core@2.11.7/lib/index.js) (needed by `bootstrap`) results in nearly 50 requests ❗ B) **Downloading Broken** For some packages, downloading simply doesn't work - rails/importmap-rails#65. ``@popperjs`/core` is another good example. Many of its imports have the form `import"./utils/getOppositeVariationPlacement.js`. If we download the main file, it looks locally for that `utils/` file, which won't be there. [`@chart`.js/auto](https://ga.jspm.io/npm:chart.js@4.3.0/auto/auto.js) is another example. C) **process.env.NODE_ENV included** Some packages (yes, again ``@popperjs`/core` is a great example!) contain `process.env.NODE_ENV` inside - rails/importmap-rails#65 (comment) I believe that some package advertise an "esm" package... but just don't do a good job of creating it... or create it without the browser context in mind (or at least in a way that's inconvenient for downloading). ### jsDelivr to the rescue THANKFULLY, jsDelivr seems to have a fantastic API/hosting that is *almost* exactly what we want: https://www.jsdelivr.com/?docs=esm They deliver fully "packaged" modules, where the only import is for peer dependencies - e.g. https://cdn.jsdelivr.net/npm/bootstrap@5.2.3/+esm There IS still an issue when downloading. "Peer imports" are relative -e.g. `import*as t from"/npm/`@popperjs`/core@2.11.7/+esm";` However, these imports follow a VERY strict pattern. So, when `--download` is passed, we parse these, download the peer dependency and update the import contents to ``@popperjs`/core`, which works with the importmap. It's not ideal that we need to do that, but it's straightforward and works great. Sorry again for this late PR - I had assumed that jspm was robust because Rails is using it. It turns out it's robust... unless you hit a "bad" package, then it's terrible. And they're not that rare: on ux.symfony.com, I have hit several. Thanks! Commits ------- b530dc3 [AssetMapper] Fix wiring resolvers, send requests in parallel and use readonly properties in MappedAsset de44614 [AssetMapper] Change default importmap "provider" to JsDelivr+esm

… readonly properties in MappedAsset

* 6.3: Update github/workflows/scorecard [AssetMapper] Avoid loading potentially ALL assets in dev server Fix CI for experimental mode [VarDumper] Fix `dd()` showing line with `null` [AssetMapper] Sometimes asset contents are built from non-asset files

* 6.2: Update github/workflows/scorecard Fix CI for experimental mode

* 5.4: Fix CI for experimental mode

…server (weaverryan) This PR was squashed before being merged into the 6.3 branch. Discussion ---------- [AssetMapper] Avoid loading potentially ALL assets in dev server | Q | A | ------------- | --- | Branch? | 6.3 | Bug fix? | yes | New feature? | no | Deprecations? | no | Tickets | None | License | MIT | Doc PR | Still TODO Hi! One other rough edge found when using on a real project. The dev server needs to quickly go from "public path" -> "logical path", so it can then look up the `MappedAsset`. Previously, for every served asset, it would loop over ALL assets until it found a match. We have a `CachedMappedAssetFactory`, which makes this happen quickly, but it still loads many assets into memory, including their contents. I was seeing a 13mb jump in the memory on those requests in a modest-sized app. If someone was serving a lot of images, it could get huge. No reason to do that work multiple times. Btw, none of this happens on production - caching is here just for dev experience. Thanks! Commits ------- bb97591 [AssetMapper] Avoid loading potentially ALL assets in dev server

Commits on May 26, 2023

Commits on May 25, 2023

Commits on May 24, 2023

Commits on May 23, 2023