| 1 |
On 19 July 2010 21:30, Mike Frysinger <vapier@g.o> wrote: |
| 2 |
> i imagine this might be useful in some scenarios, but i think the more common |
| 3 |
> usage is to enable things inline. otherwise, the exported API would need to |
| 4 |
> be wrapped internally like: |
| 5 |
> get_all_version_components() { |
| 6 |
> eshopts_need _get_all_version_components -s extglob |
| 7 |
> } |
| 8 |
|
| 9 |
It's meant to work to just define get_all_version_components with the |
| 10 |
assumption that extglob will be enabled, and then call eshopts_need in |
| 11 |
global scope right after. |
| 12 |
|
| 13 |
> although, the method we have now also |
| 14 |
> allows for disabling of shopts before calling `die`. not sure if that's |
| 15 |
> important, but i think it's better to disable before all exit/termination |
| 16 |
> points. |
| 17 |
|
| 18 |
That's a reasonable point. |
| 19 |
|
| 20 |
> so unless their is a consumer now we can point to in the tree, i'm inclined to |
| 21 |
> leave this alone. |
| 22 |
|
| 23 |
Fair enough. |
| 24 |
|
| 25 |
> there are already eshopts_push/pop funcs that accomodate more things than the |
| 26 |
> usage here, but i imagine you did it this way just to make testing in your |
| 27 |
> shell easier |
| 28 |
> -mike |
| 29 |
|
| 30 |
It was supposed to support everything the existing functions do (not |
| 31 |
everything demonstrated in the example for brevity), but I might have |
| 32 |
missed something. |
Archives