Core: add minWinningBidCacheTTL so winning bids don't expire (#12987)

[vasujain00]

Summary

Adds minWinningBidCacheTTL config option to prevent winning bids (that have had targeting set) from expiring, fixing "cannot find ad" / expired render errors when using GPT lazy load with scroll-based render.

Problem

Setting minBidCacheTTL causes "cannot find ad" errors in lazy-load scenarios. The ad is fetched from GPT, but the scroll milestone for render takes a long time. Winning bids expire and are dropped from memory before GPT can render them.

Solution

• minWinningBidCacheTTL – When set, overrides minBidCacheTTL for bids that have had targeting set (winning bids sent to the ad server).
• If unset, behavior is unchanged (all bids use minBidCacheTTL).
• Set to Infinity (or a large value) to effectively never expire winning bids.

Usage

pbjs.setConfig({
  minBidCacheTTL: 30,
  minWinningBidCacheTTL: Infinity  // keep winning bids indefinitely for lazy-load GPT
});

Changes

• src/bidTTL.ts – Added minWinningBidCacheTTL config, getMinWinningBidCacheTTL(), and getEffectiveMinBidCacheTTL(bid)
• src/auction.ts – Use effective TTL per bid; refresh bid TTL when targeting is set
• src/auctionManager.js – Use effective TTL when calculating auction TTL

Testing

• gulp lint – passed
• gulp test --file test/spec/auctionmanager_spec.js – 91 tests passed

Labels

feature | minor

---

Fixes #12987

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c06c4650ea

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Refresh auction TTL when targeting status is updated

The auction-level TTL is computed once inside au.end.then(...), so it snapshots each bid’s status at auction end; however bids are marked BID_TARGETING_SET later via setStatusForBids when setTargetingForGPT runs. Because _auctions.refresh() is never triggered on that status transition, auctions can still expire using the pre-targeting TTL (for example minBidCacheTTL/bid TTL) instead of minWinningBidCacheTTL, which reintroduces the lazy-load renderAd/"cannot find ad" failure this change is trying to prevent when all bids have short TTLs.

Useful? React with 👍 / 👎.

[patmmccann]

@vasujain00 thanks for the feature! could you link a docs pull?

[coveralls]

Pull Request Test Coverage Report for Build 22521425597

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

• For more information on this, see Tracking coverage changes with pull request builds.
• To avoid this issue with future PRs, see these Recommended CI Configurations.
• For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

• 57 of 57 (100.0%) changed or added relevant lines in 4 files are covered.
• 46 unchanged lines in 2 files lost coverage.
• Overall coverage increased (+33.2%) to 96.328%

---

Files with Coverage Reduction	New Missed Lines	%
src/adloader.js	9	76.92%
src/utils.js	37	90.51%

Totals
Change from base Build #3:	33.2%
Covered Lines:	217375
Relevant Lines:	225662

---

💛 - Coveralls

[monis0395]

Hi, I am a little caught up in some tasks will review this tom.