#6014 Spread inventory caching better to not hold shutdown as much

[akleshchev]

• Thread gzip packaging
• Flatten out collection

[Copilot]

Pull request overview

This PR adjusts inventory cache shutdown behavior by moving gzip compression work off the main thread and revising how inventory descendants are collected for caching, aiming to reduce shutdown stalls.

Changes:

• Launch gzip compression for inventory cache output on background threads and add a shutdown-time wait to join pending compression threads.
• Replace collectDescendentsIf() usage in LLInventoryModel::cache() with a manual traversal/collection pass.
• Add profiling scopes for inventory cleanup and caching paths.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

File	Description
indra/newview/llinventorymodel.h	Adds a shutdown wait API and static tracking for async cache compression threads.
indra/newview/llinventorymodel.cpp	Implements async gzip compression, new descendant collection logic, and thread-join shutdown behavior.
indra/newview/llappviewer.cpp	Ensures shutdown waits for pending async inventory cache compression to complete.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

Comments suppressed due to low confidence (1)

indra/newview/llinventorymodel.cpp:2497

• waitForPendingCacheWrites() declares LLTimer wait_timer but never uses it (likely to trigger an unused-variable warning under -Werror). Also, sPendingCacheThreads is never cleared after joining, so the vector can grow across multiple cache() calls and will remain non-empty even after all threads are joined.

        LL_DEBUGS(LOG_INV) << "Waiting for " << sPendingCacheThreads.size()
            << " inventory cache compression thread(s) to complete..." << LL_ENDL;

        LLTimer wait_timer;

        for (auto& thread : sPendingCacheThreads)
        {
            if (thread.joinable())
            {
                thread.join();
            }
        }

        LL_INFOS(LOG_INV) << "Inventory cache compression completed" << LL_ENDL;

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

Comments suppressed due to low confidence (2)

indra/newview/llinventorymodel.cpp:2498

• LLTimer wait_timer; is never used, which can trigger an unused-variable warning (often fatal under -Werror). Either remove it or log the elapsed time.

        LLTimer wait_timer;

indra/newview/llinventorymodel.cpp:2504

• After joining, sPendingCacheThreads is left populated with non-joinable threads. This makes the "Waiting for N" message inaccurate on later calls and allows the vector to grow across multiple cache() operations; clear it after joining.

        for (auto& thread : sPendingCacheThreads)
        {
            if (thread.joinable())
            {
                thread.join();

[github-actions]

Thank you for your submission, we really appreciate it. Like many open-source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution. You can sign the CLA by just posting a Pull Request Comment same as the below format.

---

I have read the CLA Document and I hereby sign the CLA

---

1 out of 2 committers have signed the CLA.
✅ (akleshchev)[https://github.com/akleshchev]
❌ @Copilot
_{You can retrigger this bot by commenting recheck in this Pull Request. Posted by the CLA Assistant Lite bot.}