Rebase from master.

@nicolasvasilache, it's currently blocked on D73609 and D73726. Then I will add tests and send it to @dblaikie for review.

Revert back to llvm-prettyprinters directory.

Maybe I should just leave the directory as is and only rename the files?

Rebase from master.

In D72589#1845519, @dblaikie wrote:

Thanks for the details! I'd be /OK/ with this, but it doesn't seem ideal to be printing the next/prev when printing the container? Do you think it's worth/there's any nice way to avoid that significant redundancy/verbosity in the printing & just print the payload?

Add detailed CHECKs.

No change, but I updated the revision summary.

GDB has a python API to customize how values are printed. See here for an introduction. The API documentation is here.

Switch from factory method to printer class.

Switch from factory method to printer class.

Remove capture from lambda.

Reworking things and improving after PointerIntPair printer has landed.

Getting rid of for loop to handle optional return value.

Revert merge accident.

Hi David. Thanks a lot for you review. I did another pass to switch to factory methods that catch exceptions without spewing common errors on the GDB console. Would you mind taking another quick peek? Thanks a lot for your patience!

Fix base. I'm so terrible at this.

Thanks a lot for your help. Sorry I've not been keeping up.

Use factory methods for printeres to catch exceptions.

clang-format.

Do not wrap temporaries in ValueRange.

Remove debug hack.

Switch from lookup_symbol of PtrTraits's 'NumBitsAvailable' to lookup_type of Info::MaskAndShiftConstants.
The latter is not (or less) part of the public interface of PointerIntPair, and requires less code changes.

clang-format

Applying dblaikie's review comments.

I was not able to get to NumLowBitsAvailable through Python GDB if it's an enumeration inside an anonymous struct.

Apply clang-format.

Fix build error after Value* -> Value change.

Rebase.

Fix formatting.

In D72129#1815174, @nicolasvasilache wrote:

What is the status re EDSC discussion?
Were the pointers I sent offline enough to give a good picture / do you see how to followup on this to use (and maybe extend) EDSCs?

Revert test_debuginfo.pl.

Move to different directory, split file.

The approach you took for building code is very similar to our motivation for declarative builders (aka EDSC), maybe it's time to reconsider those again.

Address ftynse's review comments.

1. What do you think about putting this into an llvm subdirectory, to make clear that this is a test of llvm/utils rather than generic debug info as all the other tests?

Do you mean something like debuginfo-tests/llvm-prettyprinters/? Or 'llvm/tests/prettyprinter'?

O0 > optnone

Updating again, hopefully with all changes this time.

Addressing rriddle's review comments.

Fix variable name casing.

Thanks a lot for the review, Alex!

Update reflecting review comments from ftynse.