[llvm-dev] RFC: Building GlobalISel by default

Mehdi Amini via llvm-dev llvm-dev at lists.llvm.org
Sun Jan 15 12:29:57 PST 2017

Hi Renato,

> On Jan 15, 2017, at 9:17 AM, Renato Golin <renato.golin at linaro.org> wrote:
> On 15 January 2017 at 17:08, Mehdi Amini <mehdi.amini at apple.com> wrote:
>> It seems to me that your issue is your bots are “slow”, and clang/LLVM is a large project, and running the tests takes time.
> Hi Mehdi,
> I'd appreciate if you stopped downplaying other people's concerns,
> misrepresenting arguments you don't fully understand.

I’m sorry you feel this way, I’m trying to keep the discussion on point, and technical. I feel your arguments are not relevant to enabling GlobalIsel, and  would apply to any new features. I am not sure why you think it is not appropriate to me to raise this, I believe it is.

It is very possible that I didn’t fully understand what points you’re trying to make, but that’s up to you to clarify and make your point. I think my words were carrying this, as I didn’t say out-right “you’re wrong and you are not making sense”, shutting the door to any discussion, I wrote exactly:  "I personally don’t understand […]”, which should be open enough to let you clarify.

>> I personally don’t understand how anything you’re saying is specific to GlobalISel though, it may deserve a separate thread, and we should focus on the real issue at stand, the *only* relevant question I see in this thread:  "if my (non-GlobalISel) change breaks GlobalISel, how likely is it to be a bug in my code vs. a bug in GlobalISel?”.
> I'd also appreciate if you stopped moving every thread to your own
> personal concerns, as if everything that doesn't concern you shouldn't
> concern anyone.

It is not the first time you’re picking on me personally, often on dubious basis and on moral ground. I already asked you in the past to keep the discussion technical, so I’ll ask again: I’d like you to stop *now* from attacking me personally and characterizing my position as being “personal concerns” when what I’m saying is not going in the direction you’re interested in. 

Especially I don’t think I “downplayed” your concerns by mentioning that they likely deserve their own thread, which seems to be rather promoting than downplaying…



More information about the llvm-dev mailing list