| 1 |
Hi anonymous reviewer, |
| 2 |
|
| 3 |
R0b0t1 <r030t1@×××××.com> writes: |
| 4 |
|
| 5 |
> What is it intended to solve? |
| 6 |
|
| 7 |
To simplify ebuilds that need to call eprefixify. |
| 8 |
|
| 9 |
> The current behavior seems to make more sense. Hiding defaults causes |
| 10 |
> problems. |
| 11 |
|
| 12 |
I am not sure what you mean by "Hiding defaults". It is documented, not |
| 13 |
hidden. |
| 14 |
|
| 15 |
The regular expression |
| 16 |
|
| 17 |
"s,([^[:alnum:]}])/(usr|etc|bin|sbin|var|opt)/,\1${EPREFIX}/\2/,g" |
| 18 |
|
| 19 |
is conesrvative. And it will be scrutinized by the community. |
| 20 |
|
| 21 |
Most files can be trivially prefixified by this regular expression. |
| 22 |
|
| 23 |
Traditionally, we need generate a patch with @GENTOO_PORTAGE_EPREFIX@, |
| 24 |
apply the patch and then eprefixify the source (which was |
| 25 |
"s|@GENTOO_PORTAGE_EPREFIX@|${EPREFIX}|g"). We need a lot of such |
| 26 |
trivial patches and it is not version-bump-proof. |
| 27 |
|
| 28 |
Having a sane default improves maintainability. That's the point of |
| 29 |
ebuild helpers and eclasses. |
| 30 |
|
| 31 |
> `fprefixify` is redundant. |
| 32 |
|
| 33 |
No, it's not redundant. An example of fprefixify is attached. |
| 34 |
|
| 35 |
Benda |
Archives