| 1 |
On Fri, Mar 28, 2014 at 2:57 PM, Kent Fredric <kentfredric@×××××.com> wrote: |
| 2 |
> |
| 3 |
> On 29 March 2014 06:12, Ciaran McCreesh <ciaran.mccreesh@××××××××××.com> |
| 4 |
> wrote: |
| 5 |
>> |
| 6 |
>> These look a lot like they're just parameters to an eclass... An |
| 7 |
>> alternative approach is to make this explicit, rather than having |
| 8 |
>> zillions of environment variables: |
| 9 |
> |
| 10 |
> inherit perl-module [ |
| 11 |
> runtime.requires = [ "Moose", "HTTP::Tiny" ] |
| 12 |
> ] |
| 13 |
|
| 14 |
At what point are we basically inventing our own high-level |
| 15 |
programming language, mostly designed around the |
| 16 |
limitations/idiosyncrasies of bash? |
| 17 |
|
| 18 |
I'm all for ditching the GLOBALVAR1=foo ; gosub bar paradigm, for all |
| 19 |
the reasons everybody else ditched this in the 60s. It just seems |
| 20 |
painful to try to do it in bash. |
| 21 |
|
| 22 |
For a sane programming language you'd just write: |
| 23 |
reqlist.add("Moose") |
| 24 |
reqlist.add("HTTP::Tiny") |
| 25 |
inherit perl-module [ runtime.requires=reqlist ] |
| 26 |
|
| 27 |
Or better still find some way of doing it without passing the |
| 28 |
parameters on the inherit line. A convenient dynamic syntax would be |
| 29 |
more useful for cache-safe parameters - an RDEPEND isn't necessarily |
| 30 |
the best example. |
| 31 |
|
| 32 |
Mind you, I'm not saying we shouldn't consider the exherbo approach. |
| 33 |
Without a significant overhaul of the EAPI and ditching bash I doubt |
| 34 |
we'd improve on it much. |
| 35 |
|
| 36 |
Rich |
Archives