| 1 |
On Mon, 09 Dec 2013 10:52:15 +0400 |
| 2 |
Sergey Popov <pinkbyte@g.o> wrote: |
| 3 |
|
| 4 |
> In short - please do NOT do this. |
| 5 |
|
| 6 |
Why not? What about other solutions? What do you think? |
| 7 |
|
| 8 |
There have come up several solutions, the one suggested in the first |
| 9 |
mail is no longer of interest due to SLOT="0" no longer being special; |
| 10 |
now ideas tend to focus towards disallowing slot-less dependencies, or |
| 11 |
otherwise implement under the form of a repoman warning. |
| 12 |
|
| 13 |
> More complicated answer - you break |
| 14 |
> the whole idea of slots. When user types 'emerge cat/foo' it means |
| 15 |
> now - "i want cat/foo, whatever versions it will be(depending on mask, |
| 16 |
> keywords, etc, etc)". Your proposal changes this behaviour |
| 17 |
> drastically, and reasons for such changes are not worth it. |
| 18 |
|
| 19 |
A package ATOM used in another place warrants another discussion; when |
| 20 |
an user types it, what does the user mean? We could opt to not change |
| 21 |
the behavior for users if their usage is different. |
| 22 |
|
| 23 |
Note that the syntax (and thus usage) is different by design: |
| 24 |
|
| 25 |
$ emerge -1pv '|| ( sys-apps/openrc sys-apps/systemd )' |
| 26 |
!!! '|| ( sys-apps/openrc sys-apps/systemd )' is not a valid |
| 27 |
package atom. |
| 28 |
!!! Please check ebuild(5) for full details. |
| 29 |
|
| 30 |
The reasons for changing dependency syntax are worth it: |
| 31 |
|
| 32 |
For the dependency syntax, having :* as a default breaks things or |
| 33 |
causes a lot of work. If explicit slots (or :0) were the default, it |
| 34 |
works and you spare out dealing with lots of reverse dependencies when |
| 35 |
you introduce a new slot. |
| 36 |
|
| 37 |
When you depend on libfoo; I think you mean to depend on the libfoo |
| 38 |
that is currently in the Portage tree (eg. libfoo:0 or so), whereas |
| 39 |
libfoo:* would break as soon as the libfoo maintainer starts a new slot. |
| 40 |
|
| 41 |
With libfoo:*, this then results in the following procedure: |
| 42 |
|
| 43 |
1. Add libfoo:2 to the Portage tree and package.mask. |
| 44 |
2. Inform everyone to update the dependency accordingly, wait for it. |
| 45 |
3. Unmask libfoo:2. (Somewhere in the future) |
| 46 |
|
| 47 |
Whereas with an implicit or explicit libfoo:${SLOT}, this results in: |
| 48 |
|
| 49 |
1. Add libfoo:2 to the Portage tree. (Right now, without mask) |
| 50 |
2. Maintainers support libfoo:2 over time as they bump packages. |
| 51 |
|
| 52 |
As the former is frustrating, it can have library maintainers create a |
| 53 |
new package instead of a new SLOT or not add the new version at all; |
| 54 |
whereas the latter would allow the library maintainer to just add it, |
| 55 |
especially as it is the task of individual maintainers to test their |
| 56 |
package against new libraries as they come along. |
| 57 |
|
| 58 |
Thus, that's why I'd like to see slot-less dependencies go and make it |
| 59 |
policy or at least a repoman QA warning that people specify them. |
| 60 |
|
| 61 |
Specifying it without a slot is moving the work of dependency testing |
| 62 |
on the library maintainer instead of the package maintainer, causing the |
| 63 |
library maintainer to either test tons of packages or file tons of bug. |
| 64 |
|
| 65 |
Why have the library maintainer do extra work when a package maintainer |
| 66 |
can just test it instead during a version bump? (As required anyway) |
| 67 |
|
| 68 |
What do you think? Why do you think we should still keep a default :*? |
| 69 |
|
| 70 |
-- |
| 71 |
With kind regards, |
| 72 |
|
| 73 |
Tom Wijsman (TomWij) |
| 74 |
Gentoo Developer |
| 75 |
|
| 76 |
E-mail address : TomWij@g.o |
| 77 |
GPG Public Key : 6D34E57D |
| 78 |
GPG Fingerprint : C165 AF18 AB4C 400B C3D2 ABF0 95B2 1FCD 6D34 E57D |
Archives