develooper Front page | perl.perl5.porters | Postings from January 2022

Re: Broken stack traces from use statements.

Thread Previous | Thread Next
From:
Ricardo Signes
Date:
January 16, 2022 19:40
Subject:
Re: Broken stack traces from use statements.
Message ID:
85653974-6715-47bc-a8fc-dcf3ccb1c18f@beta.fastmail.com
On Thu, Jan 13, 2022, at 12:04 AM, demerphq wrote:
> This subject has come up multiple times and is a source of deep frustration for me, and deep confusion for the many non-perl developers who work on perl code at Booking.com.
> 
> This was fixed by Dave Mitchell in f2f32cd638746f538da6db804dab6dd54e654f30
> https://github.com/Perl/perl5/commit/f2f32cd638746f538da6db804dab6dd54e654f30
> 
> This patch was then reverted in merge commit c0d05305c156c83e4f9f3a207451b3175fbb7f24.

Hello, I am joining this thread way up at the top, because the bottom looks like it's gotten pretty sticky.

There has been a lot of debate over something that I think, here, is inconsequential:  whether it is accurate to call the current output "correct."  I don't think we are going to get to the universal agreement on that without a lot more arguing, and* I don't think it buys us anything*.

On the other hand, I think everybody basically agrees that the "fixed" behavior is preferable, modulo some downstream inconvenience.  Some test suites will need to be fixed, etc.  In fact, the idea seemed to be that "somebody" would look into this come 5.33.0, but it didn't happen, because sometimes we all get busy.

So, should we have caller output that accurately reflects the relationship of the caller to the source document?  *Yes.*

I have re-opened #15109 <https://github.com/Perl/perl5/issues/15109>.

Should we apply that fix immediately?  *I don't know.*

What other CPAN code is affected by this change?  "Somebody" can run a CPAN smoke, but not me, any time soon.  To get there, we'll want to revert the revert, which at present apply cleanly.  If we smoke CPAN, see that a handful of things are fixed, and can provide fixes or at least point at how they'd be done, I feel okay about trying to ship this change this year.

If it's twenty high-value modules and exceedingly complex to fix, I think we should get it in a branch and flag it to be pushed on throughout 2022 so we can ship the fix in 5.38.0.

Would anybody like to take responsibility for getting a branch and CPAN smoke going?

-- 
rjbs
Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About