This rewrites semantics towards a more pure instruction model, in pursuit of the simple instruction-style output.
I think I can get this approach to type-check as it goes along, but obviously this change doesn't prove that yet. I'm separating it out because it's a large rewrite of the semantics structure, tossing out a lot of what was there before. But I think it does help towards several requests, like setting up a clear path for consolidating duplicate identifiers and making the node style more standardized.
I expect to need to pass multiple args to function calls, that'd probably be storing vectors of args similar to how I'm showing identifiers and integer literals stored.
This removes the semantics namespace because (a) it was getting annoying writing the `::` everywhere, and (b) I think the leaning with Carbon is to avoid namespaces (@chandlerc asked not to put SemanticsIR/SemanticsFactory in a namespace, which is the crux of the issue). But, it's still necessary to avoid name conflicts so I just prefix everything with "Semantics" (still a lot of typing, but no `::`).
In summary - some of the changes here are focused on producing the same test results for SemanticsIR, but I think the next step will be to change the SemanticsIR structure to reduce how much is added to the traversal stack.
Switching semantics to a postorder traversal is intended to be more efficient. The traversal stack is to eliminate risk of recursion limits within the semantic analysis that could come from layered code structures. However, we need to start considering the implications for type-checking and what the ParseTree looks like, as well as copying of data here.
As we start thinking about type-checking in SemanticsIR, it's helpful for a function to know its own signature in order to perform lookup recursive calls. The challenge in the post-order walk without this change is it doesn't know it's in a function definition (or similar) until it reaches the FunctionDeclaration; this restructures so that either:
1. For a declaration, the signature is a child of FunctionDeclaration(";")
2. For a definition, the signature is a child of FunctionDefinitionStart("{") which pairs with FunctionDefinition("}"), replacing CodeBlock.
This similarly reorients CodeBlock to be CodeBlockStart("{") as the first child of CodeBlock("}"). I'm not doing that with ParameterList here just because it affects a bit more, and felt like it could be delayed.
Overall, my goal is making the postorder traversal more intuitive along scope boundaries. I think we may also not need subtree_size, so I'm avoiding use of that now.
Currently the SemanticsIRFactory implementation is less clean than I might like (there are a couple comments to this point), but I was starting to feel like a more complete rewrite would be appropriate rather than trying to clean it up further: in particular, I think the node structures are off, but changing them is significant and also changes test output; in turn it may also warrant more substantial ParseTree changes. If you prefer from a reviewer POV, I can do a more complete rewrite.
Adds a simple script to merge stdout/stderr and put on labels. This is hidden to the RUN line using lit.cfg.py.
This is my solution to addressing how errors printed by the toolchain break sorting of stdout output; just put stdout first. We could also have toggles for interleaving output or such, which might help test whether we do it properly.
This also moves some previous-distributed replacement logic into lit_autoupdate_base.py: I think having that adjacent to lit.cfg.py is probably the better choice, and it reduces duplication in toolchain scripts. It happens here because I need to change the resulting commands to include the merge.
I've refactored the script in order to make it work in more contexts, which is why the delta is lost. I've actually refactored a significant amount with the intent of making the logic easier to understand, because I was also adjusting bits of it.
Some key notes:
- Removes the multi-pass update that was dealing with unfixed line numbers in explorer (I think the current script should work in one pass)
- Fixed explorer to handle multiple line numbers on the same line (turns out we can rely on local format for line numbers).
- Using execv instead of imports because making Python imports work in a setup like this feels like it's not worth it; only a nuisance.
- Adding __init__.py to satisfy mypy, which otherwise considers the lit_autoupdate.py scripts to be issues.
- Using py because I was thinking sh would be more platform-dependent. py should port better to Windows.
- Getting rid of [[ID#]] capture groups in the semantics-ir tests because with the autoupdate it's kind of moot (also, hard to autogenerate the pairs without relying on the %### value).
Note this does mean tests switch to more of a "make a change, see which tests change" setup. I don't know that that's a _bad_ thing though -- it's pretty much how tests are being written right now, which is why I went down this rabbit hole. It's a nuisance to make a change then _manually_ have to update a bunch of code.
My intent is to use this for to convert parse-tree tests to lit, but I wanted to do this with _existing_ tests first as a proof of concept and to make sure there's agreement.
On #2224 @zygoloid pointed out we needed --implicit-check-not to ensure we were correctly matching output. This is the standard way we're writing explorer tests, so I was looking at unifying our lit approaches.
This is one take on it, making more use of substitutions to bring various testing into alignment, as well as symlinks to avoid config skew (maybe I'll eventually figure out a better solution than symlinks).
Makes a couple small fixes in explorer to remove end-of-line whitespace on output.
This builds on semantics-ir lit support added by #2222
The googletest setup was feeling cumbersome, especially as I'm thinking about how to add more testing: I feel like I'm wrestling with the infrastructure.
The `[[ID1]]` and so on in tests is one advantage of switching: it's easier to do matching of IDs for verification. This is also more agnostic about the numbers than before, something which I'm concerned will be important as I think about builtins.
To explain my builtins thought, I think that needs to be another SemanticsIR with basically names pointing at builtin things. But this (a) creates multiple SemanticsIRs, which would confuse the current singleton approach and (b) starts creating more fluctuation for IDs, potentially impacting the numbers used (also, chandlerc's suggested pointers for some use-cases).
Overall it felt like I was heading towards a situation with googletest where writing the tests would be really difficult, and it was adding to my hesitance to write more code in the toolchain. I'm hoping this acts as a simplification.
Note, the "cp" commit has some incremental changes to googletest that I'd considered for making it easier to add matchers, but ultimately I ended up with this outcome.
Jon Ross-Perkins