[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: yesod / libHSrts_thr.a on sparc



Hi,

Am Mittwoch, den 28.09.2011, 11:54 +0300 schrieb Michael Snoyman:
> On Wed, Sep 28, 2011 at 11:48 AM, Joachim Breitner <nomeata@debian.org> wrote:
> > Am Mittwoch, den 28.09.2011, 07:20 +0300 schrieb Michael Snoyman:
> >> On Tue, Sep 27, 2011 at 9:34 PM, Clint Adams <clint@debian.org> wrote:
> >> > On Tue, Sep 27, 2011 at 05:34:17PM +0200, Joachim Breitner wrote:
> >> >> Another approach would be to find out (e.g. ask upstream) if yesod makes
> >> >> sense on a non-threaded runtime. If it turns out that it is hardly
> >> >> useful, then it is better to remove yesod on sparce.
> >> >
> >> > Hi Michael,
> >> >
> >> > Could you comment?
> >> >
> >>
> >> Without knowing the background here, I'd say Yesod is almost always
> >> going to be using the multi-threaded runtime. The only exception I can
> >> think of is some of the backends like CGI and FastCGI, but those are
> >> increasingly not used at all. My guess is you'll be safe.
> >
> > Let me elaborate a bit: There are architectures where the threaded
> > runtime is not available, such as sparc. We have the choice of building
> > without -threaded there, or not at all, and would like to hear your
> > opinion about this choice.
> >
> > Greetings,
> > Joachim
> 
> Yesod will definitely work in a non-threaded environment, it just
> won't be nearly as scalable. For local development and testing,
> however, it should work just fine, and it would even work for moderate
> loads in production. Actually, if someone really wanted to, they could
> always implement a multi-process system and scale pretty far.
> 
> In other words, I think it would still be useful in a non-threaded
> environment, but I wouldn't recommend it as my first choice for
> deployment.

hmm, on second thought – the binary in question is not the user
generated yesod application, but rather the "yesod" binary built from
the yesod cabal package... but as that is only a development tool,
having it non-threaded should hurt even less.

I have submitted a patch that allows us to cleanly disable the flag
using a Cabal flag, following the example set by
http://hackage.haskell.org/packages/archive/hlint/1.8.16/hlint.cabal
https://github.com/yesodweb/yesod/pull/144

Clint: You can use this patch now or wait until it is in a released
version of yesod.


Greetings,
Joachim



-- 
Joachim "nomeata" Breitner
Debian Developer
  nomeata@debian.org | ICQ# 74513189 | GPG-Keyid: 4743206C
  JID: nomeata@joachim-breitner.de | http://people.debian.org/~nomeata

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: