refactor(npm): Remove unused parallelization constructs

[fviernau]

The code looks like it would run things in parallel, but it executes getRemotePackageDetails() strictly sequentially. The pull request which had introduced these constructs 1 also introduced the caching. So, the speed-up observed back then was probably solely due to the added cache. Remove the parallelization constructs to not mislead the reader and to simplify.

Part of #9261.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 67.49%. Comparing base (8885a75) to head (697903b).
Report is 2 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #9316      +/-   ##
============================================
- Coverage     67.49%   67.49%   -0.01%     
- Complexity     1198     1200       +2     
============================================
  Files           241      241              
  Lines          8495     8493       -2     
  Branches        900      899       -1     
============================================
- Hits           5734     5732       -2     
  Misses         2399     2399              
  Partials        362      362              

Flag	Coverage Δ
funTest-docker	60.54% <100.00%> (-0.05%)	⬇️
funTest-non-docker	33.63% <ø> (ø)
test	37.47% <0.00%> (-0.02%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[sschuberth]

Mind having a look, @oheger-bosch?

[oheger-bosch]

The changes look good to me. Nice finding, that the async logic was rather useless.

[oheger-bosch]

Commit message:
Nit: "...strictly sequentially."
So, if I understand correctly, the whole fiddling with async had no real effect, since only a single package was processed, and the parsePackage function is only called sequentially? Maybe make this a bit more explicit in the commit message?

[fviernau]

since only a single package was processed

hm, I made the observation when running the fun tests, so not a single package but multiple package were involved. (A single package cannot be processed in parallel, so I guess this is a mis-understanding)

and the parsePackage function is only called sequentially

I in fact verified that getRemotePackageDetails() executes only sequentally. But, parsePackage() is the only caller of that one, so I guess the same applied to parsePackage().

What change to the commit message do you suggest?

[oheger-bosch]

From what I understand, there are multiple aspects:

• The withContext construct in getRemotePackageDetailsAsync is only exited after all coroutines started within it are finished. Therefore, the async invocations are useless, since the function waits until the asynchronous processing is complete.
• But even on a higher level, no parallelization is applied; parsePackage is invoked sequentially for each package and waits for the result. So, even if asynchronous processing worked, it would not help here.

Do you agree with this?

Regarding the commit message, it is probably not necessary to describe this in detail. So, your text is probably fine.

[fviernau]

Do you agree with this?

Yes, makes sense to me.

[oheger-bosch]

Commit message nit: "...would have searched..."