OOM errors on checking types when using >= 7.42.0

[alechartung]

With any version newer than 7.41.5, we're running into OOM errors with our fairly large Next.js application:

#13 [7/7] RUN npm run build
#13 0.688 
#13 0.688 > [email protected] build
#13 0.688 > next build
#13 0.688 
#13 1.402 Attention: Next.js now collects completely anonymous telemetry regarding usage.
#13 1.403 This information is used to shape Next.js' roadmap and prioritize features.
#13 1.403 You can learn more, including how to opt-out if you'd not like to participate in this anonymous program, by visiting the following URL:
#13 1.404 https://nextjs.org/telemetry
#13 1.404 
#13 1.621 info  - Linting and checking validity of types...
#13 400.2 
#13 400.2 <--- Last few GCs --->
#13 400.2 
#13 400.2 [29:0x7f0d565d8010]   396102 ms: Mark-sweep (reduce) 2035.5 (2076.9) -> 2035.5 (2076.9) MB, 2344.0 / 0.0 ms  (average mu = 0.302, current mu = 0.000) last resort; GC in old space requested
#13 400.2 [29:0x7f0d565d8010]   398514 ms: Mark-sweep (reduce) 2035.5 (2076.9) -> 2035.5 (2076.9) MB, 2412.0 / 0.0 ms  (average mu = 0.174, current mu = 0.000) last resort; GC in old space requested
#13 400.2 
#13 400.2 
#13 400.2 <--- JS stacktrace --->
#13 400.2 
#13 400.2 FATAL ERROR: CALL_AND_RETRY_LAST Allocation failed - JavaScript heap out of memory
#13 DONE 400.5s

This is happening during the "Linting and checking validity of types..." step.

Has anybody noticed anything similar?

I'll also dig a bit deeper and see if I can determine what's causing it.

[bluebill1049]

Could be related to this: #9873 will release the patch today, hopefully this would resolve the issue?

[alechartung]

Nope, still no luck.
I'll try and bisect v7.42.0 to see if I can find which commit started it.

[alechartung]

Went through all the commits of v7.42.0, and reverting #9540 seems to fix it locally for me.

$ git checkout v7.41.5 -- reports/api-extractor.md src/__tests__/type.test.tsx src/__typetest__/path/eager.test-d.ts src/types/path/eager.ts src/types/utils.ts

[dsthode]

Hey, same here, we were running 7.39.3 and when we upgraded to the latest version in the 7.44 series we noticed an exaggerated amount of memory used and time taken to check types:

Files:              2647
Lines:            247538
Identifiers:      286084
Symbols:         2369561
Types:           4230409
Instantiations: 65048744
Memory used:    3668930K
I/O read:          0.55s
I/O write:         0.00s
Parse time:        2.82s
Bind time:         0.61s
Check time:     1116.21s
Emit time:         0.21s
Total time:     1119.84s

We started checking RHF versions down and the last we could find with acceptable memory usage and check times was 7.41.5 (results below), anything from 7.42.0 and beyond gave results alike to the one above.

Files:             2646
Lines:           247433
Identifiers:     285861
Symbols:         311012
Types:            68722
Instantiations:  325633
Memory used:    427420K
I/O read:         1.22s
I/O write:        0.00s
Parse time:       3.92s
Bind time:        0.63s
Check time:       5.27s
Emit time:        0.06s
Total time:       9.88s

And not only the usage of resources, versions from 7.42.0 and beyond report type errors that are not present in 7.41.5.