[PATCH] D126172: [clang] Fix comparison of TemplateArgument when they are of template kind

Matheus Izvekov via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Fri Sep 2 12:05:30 PDT 2022


mizvekov added a comment.

In D126172#3767188 <https://reviews.llvm.org/D126172#3767188>, @roberteg16 wrote:

> Hey @mizvekov! Sorry for the late response.
>
> Yes, I am still interested and I am working on this. I'll let you know ASAP if I find something that resembles a valid fix.
>
> I would be very grateful if you could give me some tips for debuging better clang or if you could point me towards documents, links, whatever that could give me nice information about how to reason about the code of the parser/sema of clang.
>
> If you are in a hurry and need to fix it yourself and it happens that you finish the path first, please @ me so I can get to see how an error like this would be fixed, thanks!

Hello! Yeah you were just a bit too late, I have a patch that is accepted and should go in soon: https://reviews.llvm.org/D133072
I had run into this problem while making changes around CheckTemplateArgumentList, and I have this big patch which I am trying to offload into smaller chunks to expedite review.

I don't have any documents to recommend besides the obvious things, like the Clang docs auto-generated from the source code.
If you do want to get started, I can only recommend what worked for me, which is to try fixing bugs first, much like you tried here.
Though perhaps it's not very productive to get stuck in any one for a very long time, instead just ask for help or move on :)

I think that if you want to implement something new or a more substantial change, being able to debug quickly certainly helps.
There is no way you are going to be able to understand what is going on early on, so you will make a big mess, break hundreds of tests and so on, and being able to quickly go over and understand what is going on with them is important.
Invest in isolating bugs before working on them, it always pays off.

When fixing stuff, beware of changing assertions, most of the times the assertion is correct.
Also, if a bug fix is too local and seems easy, it's always worth spending some extra time to be sure.

And lots and lots and lots of free time, patience and being able to work without distractions.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D126172/new/

https://reviews.llvm.org/D126172



More information about the cfe-commits mailing list