fix: add environment variable to disable LWC version mismatch loglines #5187
Conversation
| }).not.toLogErrorDev( | ||
| new RegExp( | ||
| `LWC WARNING: current engine is v${process.env.LWC_VERSION}, but template was compiled with v123.456.789` | ||
| ) | ||
| ); |
There was a problem hiding this comment.
| }).not.toLogErrorDev( | |
| new RegExp( | |
| `LWC WARNING: current engine is v${process.env.LWC_VERSION}, but template was compiled with v123.456.789` | |
| ) | |
| ); | |
| }).not.toLogErrorDev(/v123\.456\.789/); |
Checking that we're not logging such a long message feels easy to get an accidental false positive, if the error text ever changes. We should just check for keywords to be more flexible.
There was a problem hiding this comment.
sure I can make it more generic
| @@ -38,6 +38,11 @@ export function checkVersionMismatch( | |||
| ) { | |||
| const versionMatcher = func.toString().match(LWC_VERSION_COMMENT_REGEX); | |||
| if (!isNull(versionMatcher) && !warned) { | |||
| if (process?.env?.SKIP_LWC_VERSION_MISMATCH_CHECK === 'true') { | |||
There was a problem hiding this comment.
process?.env
Are we expecting to encounter a scenario where theprocessvariable is declared but null/undefined?
There was a problem hiding this comment.
was just being extra cautious - I didn't know how things were compiled (and what is shipped to the browser vs whats available in server side engine). Didn't want to unintentionally break something for this one usecase.
scripts/rollup/rollup.config.js
Outdated
| @@ -55,6 +55,9 @@ function sharedPlugins() { | |||
| // This is only used inside @lwc/engine-dom and @lwc/engine-server | |||
| // Elsewhere, it _not_ be replaced, so that it can be replaced later (e.g. in @lwc/engine-core) | |||
| 'process.env.IS_BROWSER': packageName === '@lwc/engine-dom' ? 'true' : 'false', | |||
| ...(packageName === '@lwc/engine-dom' | |||
| ? { 'process.env.SKIP_LWC_VERSION_MISMATCH_CHECK': 'false' } | |||
| : {}), | |||
There was a problem hiding this comment.
@divmain is this right? Or what were you looking for
|
/nucleus test |
Head branch was pushed to by a user without write access
|
/nucleus test |
|
/nucleus test |
|
/nucleus ignore --reason "unrelated __lwc issue in webruntime" |
|
⚠ The test stage has not been run for this PR. Nothing to ignore. |
|
/nucleus test |
|
/nucleus ignore --reason "unrelated __lwc issue in webruntime" |
|
⚠ The test stage has a status of running, which cannot be ignored. |
|
/nucleus ignore --reason "unrelated __lwc issue in webruntime" |
|
/nucleus test |
|
/nucleus ignore --reason "unrelated __lwc issue in webruntime" |
|
⚠ The test stage has a status of cancelled, which cannot be ignored. |
|
/nucleus ignore --reason "unrelated __lwc issue in webruntime" |
Details
When using LWR local dev for experience sites, every SSR request logs an LWC version mismatch error. This is due to the fact that the engine-server is intentionally reloaded on a per page basis and not cached in memory. Therefore the mechanisms in place to only log a the version mismatch once do not work. The logging noise from this makes observing actual useful errors during SSR more difficult.
My change just introduces a new environment variable that can be set on the LWR side to disable this logging during local development.
Would like to backport this also to spring25
Does this pull request introduce a breaking change?
Does this pull request introduce an observable change?
GUS work item
@W-17740011@