| 1 |
On Sat, Sep 21, 2013 at 03:21:07PM -0400, Ian Stakenvicius wrote: |
| 2 |
> -----BEGIN PGP SIGNED MESSAGE----- |
| 3 |
> IIRC we still don't have an openrc-replacement script in the tree for |
| 4 |
> the /etc/init.d/function.sh symlink to target. Since libeinfo is |
| 5 |
> already public, why not instead of making it private we go the other |
| 6 |
> way -- keep it public, package it out separately in the tree, and make |
| 7 |
> openrc (and others from bug 373219 and elsewhere) depend on it? |
| 8 |
|
| 9 |
Because it is a c library, which means that another program would have |
| 10 |
to be written which provides the einfo/ewarn/etc shell commands and a |
| 11 |
functions.sh wrapper so the shell scripts can use it. |
| 12 |
|
| 13 |
Since the consumers on bug 373219 are shell scripts, why have the |
| 14 |
complexity of a wrapper and not just provide a shell script? |
| 15 |
|
| 16 |
I know I have been slow about it. mostly because I've been doing a lot |
| 17 |
of work on OpenRC lately wrt bug #482396. That should be wrapping up |
| 18 |
soon. |
| 19 |
|
| 20 |
|
| 21 |
> Out of curiosity, what is the reasoning behind making these libs private? |
| 22 |
|
| 23 |
Well, the thought has changed slightly. librc can't be made private |
| 24 |
currently because of openrc-settingsd. libeinfo, on the other hand, does |
| 25 |
not have any known consumers, so there is no reason to keep it as a |
| 26 |
library. |
| 27 |
|
| 28 |
> -----BEGIN PGP SIGNATURE----- |
| 29 |
> Version: GnuPG v2.0.20 (GNU/Linux) |
| 30 |
> |
| 31 |
> iF4EAREIAAYFAlI98aMACgkQ2ugaI38ACPCTcQD9HIqOTlhia/ktPFANAZdJbfEv |
| 32 |
> DqOh7CUCULZw+FqkOpQBAISPbWdsg+flecvnv5OfWGsnLqnYK6GPG4e23KwDyz1e |
| 33 |
> =OCdp |
| 34 |
> -----END PGP SIGNATURE----- |
| 35 |
> |
Archives