[erlang-questions] Can't build on new Sun T2?

[Per Hedeland]

Mikael Pettersson <mikpe@REDACTED> wrote:
>
>On Fri, 5 Oct 2007 21:27:38 +0200 (CEST), Per Hedeland wrote:
>> Good point, and I'll raise you one: If the (non-meta) <sigh>, in
>> response to a user that is encountering absurdly weird errors(*),
>> whether he tried to find the answer in the mailing list archives and
>> failed or simply couldn't bother, comes from one of the people working
>> on the component that causes these absurdly weird errors to happen in
>> the first place, it's bordering on being offensive.
>
>There's a bug in Sun's toolchain. Should I as I HiPE developer feel
>guilty about triggering it? H*ll no. I did not _cause_ the problem,
>Sun did. It's not my job to bend over backwards to accomodate every
>**** bug out there, especially when there are viable workaorunds.

I didn't suggest that you should - the problem I referred to was not
that HiPE failed to build, but that the requirements of HiPE, which are
clearly stronger than those of the rest of Erlang/OTP or most other
open-source packages (nothing *wrong* with that, of course), caused the
complete Erlang/OTP build to fail.

And while I guess that you were just venting your frustration over the
Sun problem, for me (and apparently I wasn't alone) your message came
across as reproaching the presumably innocent user that came across it.

>Did per@REDACTED ever contribute code to solve this problem? No.

Well, I occasionally contribute code to various open-source packages,
but I have to admit that I normally need the motivation of a problem
that I'm affected by to do so.

>> OK, to be a bit constructive - wouldn't it be possible to have a
>> configure test for HiPE that identified these kinds of problems, and
>> caused HiPE to be disabled with an appropriate warning message, without
>> failing the build altogether?
>
>Given the amount of grief Sun's crap toolchain is causing us,
>I wouldn't be opposed to not autoenabling HiPE on Solaris/SPARC.

Seems like a good idea to me.

--Per