Add more Dump methods for debugging
#4747
Conversation
| } | ||
|
|
||
| LLVM_DUMP_METHOD auto Dump(const File& file, ClassId class_id) -> void { | ||
| llvm::errs() << class_id; |
There was a problem hiding this comment.
The idea with the NoNewlines is that all the dumping logic lives in those functions so that they can be composed into other DumpNoNewline classes, and the body of each Dump() is just a DumpNoNewline() + newline. So this body here could move into a DumpNoNewline function.
Same for the other Dump functions below.
There was a problem hiding this comment.
I recognized that was the intent, and that is how I started writing this PR, but in practice it didn't seem to work out that way. The information I wanted to print for a given id (a) benefited from formatting using newlines and indents and (b) didn't decompose into calls to DumpNoNewline functions. To avoid extra boilerplate, I've instead only written DumpNoNewline functions when there was a caller that wanted the functionality of one of the Dump functions without the newline, instead of writing an extra function proactively.
toolchain/sem_ir/dump.cpp
Outdated
|
|
||
| namespace Carbon::SemIR { | ||
|
|
||
| static auto DumpIfValid(const File& file, NameId name_id) -> void { |
|
|
||
| LLVM_DUMP_METHOD auto Dump(const File& file, EntityNameId entity_name_id) | ||
| -> void { | ||
| llvm::errs() << entity_name_id; |
There was a problem hiding this comment.
I would suggest for each of these to write the type being dumped in the output too, see for example https://github.com/carbon-language/carbon-lang/pull/4747/files?diff=unified&w=0#diff-e690edf9337d3bd7e874588bc9d865536d6c69f2b7e8db79a9ef3aa9fd4b6edbL33-L42
So it could be something like EntityNameId(...the stuff in here...)
There was a problem hiding this comment.
I believe the print functions for each of the Id classes already includes its type:
carbon-lang/toolchain/base/index_base.h
Line 52 in 89df777
Follow-on to #4669 . Did some manual testing, but may have not exercised all code paths.