Re: [Nix-dev] --ftls-model=global-dynamic for every lib?

2016-02-27 Thread stewart mackenzie
Just to clarify the correct argument passed to gcc or clang should be -fpic . ___ nix-dev mailing list nix-dev@lists.science.uu.nl http://lists.science.uu.nl/mailman/listinfo/nix-dev

[Nix-dev] --ftls-model=global-dynamic for every lib?

2016-02-27 Thread stewart mackenzie
Hello, Seems libraries in nix are not compiled with --ftls-model=global-dynamic which means in the use case of fractadie we quickly run out of TLS variables. Fractalide is a Flow-based programming programming platform tha has a thread per component, a component is a shared object and could be

Re: [Nix-dev] AppImage

2016-02-27 Thread Eelco Dolstra
Hi, On 27/02/16 15:28, Vladimír Čunát wrote: > On 02/27/2016 03:00 PM, Kevin Cox wrote: >> So while I think it is cool that it is simply a binary you can run you >> obviously lose out on the efficiency of Nix. > > The aims seem significantly different. Appimage and many others > evidently focus

Re: [Nix-dev] AppImage

2016-02-27 Thread Vladimír Čunát
On 02/27/2016 10:45 PM, Mateusz Czaplinski wrote: > (though Nix in theory can be used userspace-only > too at the cost of full recompilations, right?) Some reported to successfully use nix on machines without any root access and without recompilation, thanks to the proot tool. smime.p7s

Re: [Nix-dev] AppImage

2016-02-27 Thread Mateusz Czaplinski
Personally, after seeing the articles I also thought about Nix and how they relate, but after some time I too now think they're rather complementary, serving quite different needs and with different pros & cons. Specifically, what I value very highly in Nix is the declarative specification and

Re: [Nix-dev] AppImage

2016-02-27 Thread Anand Patil
Having only read the front page, it looks like it could be complementary to Nix. You could develop and build in Nix for reproducibility, efficiency and completeness. Then, for portable deployment, you could run nix2appimage or else include AppImageKit in a derivation and write the AppImage to

Re: [Nix-dev] AppImage

2016-02-27 Thread Vladimír Čunát
On 02/27/2016 03:00 PM, Kevin Cox wrote: > So while I think it is cool that it is simply a binary you can run you > obviously lose out on the efficiency of Nix. The aims seem significantly different. Appimage and many others evidently focus on creating bundles that "just work" everywhere. That

Re: [Nix-dev] AppImage

2016-02-27 Thread Kevin Cox
Well the key differences I see are: Nix: - Dependencies are shared, less disk space and better cache usage. - Requires software on the system to manage the programs. AppImage: - Dependencies are bundled. - Everything comes in a "standard" ELF executable so it can run anywhere (provided a

Re: [Nix-dev] AppImage

2016-02-27 Thread Vladimír Čunát
On 02/27/2016 02:31 PM, Yacine Hmito wrote: > Tell me what you think. I only had a very brief look now, but the steps described in [3] seem to indicate that the packages aren't as self-contained as one might hope. (Formulations like: add everything that might be missing on your target OS :-) [3]

[Nix-dev] AppImage

2016-02-27 Thread Yacine Hmito
Hi, I just stumbled on the AppImage website[1] on reddit. The github README[2] that roughly explains how it works show some overlap with how Nix organizes the internals of a derivation in the store. I'm sending you this because you might find it interesting and I would be glad to hear your

Re: [Nix-dev] NixOS 16.03 branch-off time in 10 days

2016-02-27 Thread Domen Kožar
Tomorrow is the branch off day, time to merge last sneaky changes. On Thu, 18 Feb 2016, 19:25 Domen Kožar wrote: > Hi all, > > it's that time of the year when it's time to talk about new NixOS release. > According to our somewhat-documented release process at >