On 22 March 2010 10:57, Bernie Pope wrote:

Do you have minumum requirements for GHC? I tried to 'cabal install salvia-demo', but with no luck:

Looking at the dependencies listed, they claim that GHC should work with >= 6.10.1.

cabal install salvia-demo Resolving dependencies... cabal: cannot configure syb-with-class-0.6.1. It requires template-haskell ==2.4.* For the dependency on template-haskell ==2.4.* there are these packages: template-haskell-2.4.0.0. However none of them are available. template-haskell-2.4.0.0 was excluded because template-haskell-2.3.0.1 was selected instead template-haskell-2.4.0.0 was excluded because ghc-6.10.4 requires template-haskell ==2.3.0.1

Hmmm.... a quick look through the dependencies couldn't reveal what was pulling in syb-with-class; maybe if you just install syb-with-class-0.6 first and it will use that instead? -- Ivan Lazar Miljenovic Ivan.Miljenovic@gmail.com IvanMiljenovic.wordpress.com

On 22 March 2010 11:05, Carter Schonwald wrote:

apparently sometimes even though cabal can figure out the dependencies for a package you want, it gets confused (or something) when it needs to figure out the "transitive" dependencies (that which needs to be installed for the direct dependencies to also install). So  try doing a cabal install of that version of template  haskell with the explicit version cabal install template-haskell-2.4.0.0 that should solve it

Yes, I tried that, but unfortunately it falls over with: Language/Haskell/TH/Quote.hs:31:12: Not in scope: data constructor `CharConstr' cabal: Error: some packages failed to install: template-haskell-2.4.0.0 failed during the building phase. The exception was: exit: ExitFailure 1 In the version of Data.Data on my machine there does not appear to be a CharConstr. It looks to me like there is an incorrect dependency in Template Haskell. I can see from the hackage build logs that I'm not the only one who has this problem. I will report a bug. Cheers, Bernie.

On 25 March 2010 14:23, Ivan Miljenovic wrote:

On 25 March 2010 12:21, Bernie Pope wrote:

...

Yes, I tried that, but unfortunately it falls over with:

Language/Haskell/TH/Quote.hs:31:12:    Not in scope: data constructor `CharConstr' cabal: Error: some packages failed to install: template-haskell-2.4.0.0 failed during the building phase. The exception was: exit: ExitFailure 1

TH-2.4 comes with/needs GHC 6.12.  The problem here appears to by with syb-with-class: using a lower version of that should work, as 0.6.1 appears to be a compatability release just to get it working on 6.12 (whereas 0.6 works with 6.10).

Thanks Ivan. cabal install syb-with-class-0.6 fixed the problem. I note that you did mention this before, and sorry I missed it the first time. Cheers, Bernie.

Nice! This is certainly worth it. I really liked the simplicity of the threadmanager package, it certainly was better than managing threads manually. But your concurrent-extra package seems to do a way better job, I'm certainly going to apply this patch. Thanks a lot guys! Groet, Sebastiaan On Mar 23, 2010, at 12:36 PM, Bas van Dijk wrote:

On Sun, Mar 21, 2010 at 5:05 PM, Sebastiaan Visser wrote:

...

Straight from Zurihac: I'm very pleased to announce the 1.0.0 release of the Salvia web server.

Hoi Sebastiaan,

(switching to English) I discovered a major space-leak in Network.Salvia.Impl.Server.start due to the use of threadmanager:

When a request handler is forked the threadmanager internally maps the threadId to a status MVar. However when the request handler terminates this association is not deleted from the internal Map. This results in a major O(n) space-leak where n = the number of requests accepted.

This space-leak can be visualized when you heap-profile the salvia-helloworld demo:

salvia-helloworld +RTS -hy

and perform some requests:

$ for ((a=0; a<10000; a++)); do curl -s -o /dev/null localhost:8080; done

You will then get the following heap profile which will clearly highlight the linear space-leak:

http://bifunctor.homelinux.net/~bas/salvia-helloworld_ORIGINAL_10000_request...

I fixed it by replacing threadmanager by Control.Concurrent.Thread[1] from Roel's and mines concurrent-extra. This module has the following advantages over threadmanager:

* Simpler: We don't have the concept of a ThreadManager. So you don't need to make one and pass it to fork and wait.

* More efficient: We don't keep an internal mapping from ThreadIds to status MVars.

* Correct: threadmanager has a bug where it can potentially miss the termination of a thread. It doesn't block asynchronous exceptions before it forks the computation that will install the necessary exception handler before executing the given action. We correctly block asynchronous exceptions.

Besides these advantages we have the ability to get the return value of the forked thread so you don't need a mutable variable.

Besides replacing threadmanager I now manually delete the threadId of the request handler when it terminates.

When I repeat the heap-profile you will now see a nice constant heap-usage:

http://bifunctor.homelinux.net/~bas/salvia-helloworld_NEW_10000_requests.ps

I don't know how to attach my git patch to this email so I just copied my git repository to:

http://bifunctor.homelinux.net/~bas/salvia/

I think you can just pull from there.

Groeten,

Bas

[1] http://hackage.haskell.org/packages/archive/concurrent-extra/0.4/doc/html/Co...

On Tue, Mar 23, 2010 at 9:27 AM, Bas van Dijk wrote:

On Tue, Mar 23, 2010 at 2:13 PM, Sebastiaan Visser wrote:

...

Nice! This is certainly worth it.

I'm glad you like it.

Sebastiaan, I made the same mistake as threadmanager does: I forgot to block before installing the deleteMyPid exception handler in the forked thread. I added a new patch that adds the necessary block and unblock: http://bifunctor.homelinux.net/~bas/salvia/

BTW What's the git equivalent of 'darcs send -o <filename>' which saves the patches to <filename>? I would rather send my patches as email attachements instead of copying my repository to my webserver. (Note this is the first time I used git)

I use ' git format-patch origin'. -- gwern

On Tue, Mar 23, 2010 at 2:30 PM, Gwern Branwen wrote:

I use ' git format-patch origin'.

Thanks! In case you have trouble pulling from my webserver I attached the same two patches using Gwern's method to this email. On Tue, Mar 23, 2010 at 2:34 PM, Chris Eidhof wrote:

The way I like to work is forking the repository on github (you need an account for that, though). Then you can push the changes to your forked repository and the original author will see it.

Mmm interesting, I will check it out when I have time. However I do like the simplicity of just emailing patches. regards, Bas