[clang] [llvm] [LTO] Reduce memory usage for import lists (PR #106772)

[Kazu Hirata via cfe-commits]

kazutakahirata wrote:

> > It looks like this causes a compile-time regression. The clang thin link time increases by 1% (https://llvm-compile-time-tracker.com/compare_clang.php?from=9ccf82543d509bb5a0f5d0551fc4d6c1913b9a9b&to=5c0d61e318a77434487fcec8361d8110fb06e59d&stat=instructions%3Au). There is no regression for thin link on CTMark, so the regression here probably scales with the number of modules involved in the link, so is likely even larger for larger applications.
> 
> Thank you for reporting this! Let me look into this.

To close the loop, https://github.com/llvm/llvm-project/pull/106998 should fix the compile time increase.  Yes, as you've indicate, the regression should scale in a quadratic manner.

https://github.com/llvm/llvm-project/pull/106772