Considering this should be a drop in replacement for the expr family, a clash wouldn't actually be problematic if we can define which one is preferred, though I'm not sure if that is possible so maybe I'll just stick with the "jit" prefix.

I'll test out -lib later on my linux machine.

On Wed, Feb 21, 2018 at 9:43 AM, Lucas Cordiviola <lucarda27@hotmail.com> wrote:

I'm not sure but -lib might also work on linux.

Mensaje telepatico asistido por maquinas.
On 2/21/2018 2:39 PM, Alex wrote:
[declare -lib jit_expr] worked on my osx machine.

On Wed, Feb 21, 2018 at 9:27 AM, Lucas Cordiviola <lucarda27@hotmail.com> wrote:

Can you try on you osx [declare -lib jit_expr] or [declare -lib full/path/to/jit_expr]?

I'm not sure if -stdlib is currently covering the .../documents/pd/externals/* on all platforms.

see https://puredata.info/docs/faq/how-do-i-install-externals-and-help-files for the paths that -stdlib covers.

there's a discussion here : https://github.com/pure-data/pure-data/pull/205

---

I think there will be no name clashes if you keep your [jit_expr] naming. If you use [expr] it will clash with Pd's [expr].



--

Mensaje telepatico asistido por maquinas.
On 2/21/2018 1:50 PM, Alex wrote:
I have an external I've created that I've put into: /Users/alex/Documents/Pd/externals/jit_expr/
the binary is called "jit_expr.pd_darwin"

xnor-work:~/projects/perfect/center$ ls /Users/alex/Documents/Pd/externals/jit_expr/
LICENSE-parser     LICENSE-pd         LICENSE-xnor       jit_expr-help.pd   jit_expr.pd_darwin

In the binary, the externals themselves are named "jit/expr" "jit/expr~" and "jit/fexpr~"

On my work, osX machine, when I [declare -stdlib jit_expr] I then fail to create [jit/expr]
On my home, linux machine, where my external folder is located at /home/alex/.local/lib/pd/externals/jit_expr and appropriately named jit_expr.pd_linux
I succeed to create [jit/expr] after [declare -stdlib jit_expr]

Is this a bug am I not following some naming convention that I should be?

I could see wanting to put my external in a folder called "jit" and then naming them "expr", "expr~" etc but I could imagine "jit" is more likely to collide and so "jit_expr" seemed more reasonable.
maybe I should just call the whole thing "jit_expr", and just call the objects "expr" etc without the prefix in the code and then be able to create [jit_expr/expr] with no declare because it'll be in the stdlib already?
Am I confusing conventions that exist for single binary libraries with multiple objects and binaries with only one?

Either way, should I file a bug that the behavior isn't the same on osx and linux?

thanks,
Alex


_______________________________________________
Pd-dev mailing list
Pd-dev@lists.iem.at
https://lists.puredata.info/listinfo/pd-dev