[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gentoo-soc
Subject:    Re: [gentoo-soc] Week 1 Report for Refining ROCm Packages in Gentoo
From:       wuyy <xgreenlandforwyy () gmail ! com>
Date:       2022-06-21 9:14:47
Message-ID: YrGMB2O4BCenWOLO () HEPwuyy
[Download RAW message or body]

On Tue, Jun 21, 2022 at 04:35:20PM +0800, Benda Xu wrote:
> > Benda Xu <heroxbd@gentoo.org> writes:
> 
> You are not answering my question.  My original comment was only on the
> severe issue of hipcc 'isystem'.  Although you mentioned "I came to a
> new hipvars.pm and a patch to hipcc.pl, disabling poisoning `-isystem`
> and correcting many paths", I cannot see any fix pushed into ::gentoo.
> Otherwise we won't be bitten by bug 853184.

The statement "I came to a new hipvars.pm and a patch to hipcc.pl,
disabling poisoning `-isystem`" describes the situation in packaging
dev-util/hip-5.1.3 against vanilla llvm/clang on my own development
branch, and it is independent of ug 853184 (they may share some
similarities, but at that time I didn't realize there can be bug in
hip-5.0.2). Bug 853184 is introduced by https://github.com/gentoo/gentoo/pull/25861. 

I am sorry for pusin dev-util/hip-5.0.2-r1 without test it's functionality.
I'll go back to the 5.0.2 branch and fix up the mess, and setup a CI
system to provide solid testings.

> 
> This is a situation you should avoid at all cost in the future.

-- 

[prev in list] [next in list] [prev in thread] [next in thread]