✨ Allow to modify source_url property in Loaf script attributions

[thomas-lebeau]

Motivation

We want to be able to modify source_url and invoker properties of Loaf script attributions from the beforeSend function. This allow customer to scrub urls for potential PII.

Changes

Refactor limitModification to allow arrays to be present in the path. (e.g. long_task.scripts[].source_url will allow to modify source_url in long tasks (LoaF) events

Allow modification on long_task.scripts[].source_url and long_task.scripts[].invoker

Testing

• Local
• Staging
• Unit
• End to end

---

I have gone over the contributing documentation.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.96%. Comparing base (0365297) to head (7660dcd).
Report is 5 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3325      +/-   ##
==========================================
+ Coverage   92.94%   92.96%   +0.01%     
==========================================
  Files         297      297              
  Lines        7840     7833       -7     
  Branches     1785     1785              
==========================================
- Hits         7287     7282       -5     
+ Misses        553      551       -2     

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

[cit-pr-commenter]

Bundles Sizes Evolution

📦 Bundle Name	Base Size	Local Size	𝚫	𝚫%	Status
Rum	147.04 KiB	147.06 KiB	18 B	+0.01%	✅
Logs	51.33 KiB	51.33 KiB	0 B	0.00%	✅
Rum Slim	105.77 KiB	105.79 KiB	18 B	+0.02%	✅
Worker	24.50 KiB	24.50 KiB	0 B	0.00%	✅

🚀 CPU Performance

Action Name	Base Average Cpu Time (ms)	Local Average Cpu Time (ms)	𝚫
addglobalcontext	0.002	0.002	0.000
addaction	0.039	0.035	-0.005
addtiming	0.001	0.001	-0.000
adderror	0.049	0.045	-0.003
startstopsessionreplayrecording	0.009	0.012	0.003
startview	0.358	0.357	-0.001
logmessage	0.020	0.020	0.000

🧠 Memory Performance

Action Name	Base Consumption Memory (bytes)	Local Consumption Memory (bytes)	𝚫 (bytes)
addglobalcontext	28.90 KiB	27.31 KiB	-1627 B
addaction	56.42 KiB	53.58 KiB	-2906 B
addtiming	27.15 KiB	26.14 KiB	-1036 B
adderror	58.56 KiB	55.43 KiB	-3210 B
startstopsessionreplayrecording	26.08 KiB	26.07 KiB	-10 B
startview	419.23 KiB	420.73 KiB	1.50 KiB
logmessage	62.94 KiB	61.83 KiB	-1139 B

🔗 RealWorld

[sethfowler-datadog]

I love the approach! The one suggestion I have is to tweak the implementation strategy slightly. If you change this line to:

Suggested change

	const pathSegments = fieldPath.split('.')
	const pathSegments = fieldPath.split(/\.|(?=\[\])/)

Then instead of foo.bar[].baz being split into ['foo', 'bar[]', 'baz'], it will be split into ['foo', 'bar', '[]', 'baz']. This should lead to a cleaner implementation of getValueFromPath(), where you can treat [] as a special field name that maps over all elements of an array. With the appropriate changes to getValueFromPath(), it will also mean that foo.bar[][].baz will work as expected with no special effort, so that we can support modifications of nested arrays.

[sethfowler-datadog]

We don't validate here that the length matches; is that expected? I may be misreading the code, but it seems like this will allow people to remove items from the array while still passing the isValidType() check; the result looks like it would be that you'd get undefined for some elements in the output array. That may be OK, but I just wanted to make sure it was intentional. (Or that it's handled somehow and I'm just missing it.)

[thomas-lebeau]

You're right, I haven't consider this. Actually, it also means that every property of the clone is a valid type. If only some are, none will be copied to the original object.

I've come up with a better algorithm, that should be resilient to such inconsistencies in the clone: traverse both, the object and the clone recursively at the same time, up to the end of the path where I can check the path individually.

At the end, I think the code is simpler and even more efficient as we now do the get and set traversal at once

[sethfowler-datadog]

Yeah, I really like the final result! Nicely done.

[amortemousque]

🥜 nitpick: ‏

if (field === '[]') {
  if (Array.isArray(object) && Array.isArray(clone)) {
    object.forEach((item, i) => setValueAtPath(item, clone[i], restPathSegments, fieldType));
  }
  return;
}

[sethfowler-datadog]

Looks good! I really like the final solution you came up with; very clean!