dmd
dmd copied to clipboard
Add option to enable stack unwinding in betterC mode
I wanted to debug my code in my project, but i'm using -betterC, wich removes the ability for the compiler to generate eh_frame, so backtrace wasn't working
This adds an option for -betterC to still get the benefits of nothrow static analysis + being able to debug code with ease, this option is disabled in -release
Documentation perhaps could be improved
void crash_me()
{
int* a;
*a = 42;
}
Before:
-------------------------------------------------------------------+
Received signal 'SIGSEGV' (11)
-------------------------------------------------------------------+
executable: /tmp/dmd_run1UPD7W
backtrace: 1
After:
-------------------------------------------------------------------+
Received signal 'SIGSEGV' (11)
-------------------------------------------------------------------+
executable: /tmp/dmd_runXn753y
backtrace: 7
??:? _start+0x25
??:? __libc_start_main+0x8a
??:? +0x25cd0
/run/media/ryuukk/E0C0C01FC0BFFA3C/dev/kdom/./_.d:4 main+0xe
/run/media/ryuukk/E0C0C01FC0BFFA3C/dev/kdom/./_.d:11 void _.crash_me()+0xe
Test code:
dmd -fPIC -fPIE -gs -debug -g -betterC -run _.d
https://gist.github.com/ryuukk/e2d5260b4c4acec1b972ad9ce9228753
Thanks for your pull request and interest in making D better, @ryuukk! We are looking forward to reviewing it, and you should be hearing from a maintainer soon. Please verify that your PR follows this checklist:
- My PR is fully covered with tests (you can see the coverage diff by visiting the details link of the codecov check)
- My PR is as minimal as possible (smaller, focused PRs are easier to review than big ones)
- I have provided a detailed rationale explaining my changes
- New or modified functions have Ddoc comments (with
Params:andReturns:)
Please see CONTRIBUTING.md for more information.
If you have addressed all reviews or aren't sure how to proceed, don't hesitate to ping us with a simple comment.
Bugzilla references
Your PR doesn't reference any Bugzilla issue.
If your PR contains non-trivial changes, please reference a Bugzilla issue or create a manual changelog.
Testing this PR locally
If you don't have a local development environment setup, you can use Digger to test this PR:
dub run digger -- build "master + dmd#16174"
I'm curious, why was this disabled in the first place?
-betterC does not imply that the entire process does not use exceptions.
https://github.com/dlang/dmd/pull/7305 no mention as to why it was disabled.
I'm curious, why was this disabled in the first place?
-betterCdoes not imply that the entire process does not use exceptions.
I don't know, but when using -betterC, it sets: params.useExceptions = false;
Perhaps the right thing to do is to remove that and let user decide by using -nothrow directly, i think that's how it should be done, but i don't know the rationale behind the default
Agreed, opt-out is better for this I think.
I'm curious, why was this disabled in the first place? -betterC does not imply that the entire process does not use exceptions.
Actually it does. Exceptions are by default allocated with the gc which is not available if you're not using the runtime. The fact that a function is nothrow does not mean that exceptions are not going to be allocated. There is a compiler flag -dip1008 which makes exceptions to be ref counted. This should work with betterC, however, it has a limitation: exceptions that are catched cannot escape the catch block. However, that does not seem to be the case in your situation.
I'm curious, why was this disabled in the first place? -betterC does not imply that the entire process does not use exceptions.
Actually it does. Exceptions are by default allocated with the gc which is not available if you're not using the runtime. The fact that a function is nothrow does not mean that exceptions are not going to be allocated. There is a compiler flag -dip1008 which makes exceptions to be ref counted. This should work with betterC, however, it has a limitation: exceptions that are catched cannot escape the catch block. However, that does not seem to be the case in your situation.
You have misunderstood this.
The unwinding tables are separate from being able to throw and catch.
It might be possible that the implementer made the same term confusion as I did.
cc @WalterBright
Can't you remove -betterC when debugging?
The point is to improve debugging with -betterC (technically it is -nothrow), not to ignore it
Generating unwinding tables in -betterC is fine, but I'm not a fan of the added -unwinding flag. In general, 'flags are bugs', and this one undoes the undoing of another flag.
I also sent this PR after riki comment https://github.com/dlang/dmd/pull/16177 is it better?
I think so