commit e4d9b34a90b5b15bee9607bc3bf81c6044ccc655
parent 210b064e7f56974a6da4418df8b0037a5631ed30
Author: Luís Ferreira <[email protected]>
Date: Thu, 25 Nov 2021 23:36:53 +0000
posts: Add 'SAOC LLDB D integration: 10th Weekly Update'
Signed-off-by: Luís Ferreira <[email protected]>
Diffstat:
2 files changed, 63 insertions(+), 1 deletion(-)
diff --git a/content/posts/d-saoc-2021-09.md b/content/posts/d-saoc-2021-09.md
@@ -74,4 +74,5 @@ You can also read this on the D programming language forum,
[here](https://forum.dlang.org/thread/mailman.891.16372[email protected]),
and discuss there.
-Read about the [previous week](../d-saoc-2021-08/).
+Read about the [previous week](../d-saoc-2021-08/) and [next
+week](../d-saoc-2021-10/).
diff --git a/content/posts/d-saoc-2021-10.md b/content/posts/d-saoc-2021-10.md
@@ -0,0 +1,61 @@
+---
+title: 'SAOC LLDB D integration: 10th Weekly Update'
+date: '2021-11-25T21:26:00+01:00'
+tags: ['saoc', 'saoc2021', 'dlang', 'llvm', 'lldb', 'debug', 'debugging', 'dwarf']
+description: "This post describes what I've done on the 10th week of the
+Symmetry Autumn of Code 2021, including follow up updates on the LLVM patches
+and implementation of a minimal TypeSystem for D."
+---
+
+Hi D community!
+
+I'm here again, to describe what I've done during the tenth week of Symmetry
+Autumn of Code.
+
+## LLVM upstream follow up update
+
+During the discussion for the next patch in the train, 4 more patches have been
+raised, due to request to split it even more:
+
+- https://reviews.llvm.org/D114305
+- https://reviews.llvm.org/D114307
+- https://reviews.llvm.org/D114308
+- https://reviews.llvm.org/D114309
+
+Some more changes were made on other existing patches due to further discussion
+on implementation details.
+
+I also proposed a minor change on the specification to accommodate artificial
+local parent symbols when mangling ambiguities happen, see
+[here](https://github.com/dlang/dlang.org/pull/3124).
+
+## Minimal TypeSystem for D
+
+This week I finally focused on getting a minimal TypeSystem working for D on
+the LLDB. This means nothing to the end user and huge change need to be made to
+make it functional exactly the same as is right now, the main difference is
+that, instead of relying on Clang type system, we have our own.
+
+For now it has the Clang type system disabled and detects D as usual, but does
+nothing, when requesting any type information.
+
+The current changes for the implementation of D TypeSystem will be in
+[this](https://github.com/ljmf00/lldb-d/pull/1) PR.
+
+I decided to create a seperate repository for this until I have a decent plugin
+to upstream. As this might take a while to implement, I would like to ask the
+community if it is worth to do some builds to several platforms with the
+demangling and language formatters enabled? I'm particularly asking this since
+building LLVM is a beefy task and some may not be able to do it easily with
+their machine. What is your thoughts?
+
+## What is next?
+
+I will continue implementing TypeSystem for D as well as a new type wrapper to
+represent a D type. I'm going to start adding support for very basic types.
+
+You can also read this on the D programming language forum,
+[here](https://forum.dlang.org/thread/[email protected]),
+and discuss there, including your thoughts about the asked questions!
+
+Read about the [previous week](../d-saoc-2021-09/).