1 |
>>>>> On Fri, 27 Jan 2012, Ralph Sennhauser wrote: |
2 |
|
3 |
> Just recently Portage started to emit "* QA Notice: dosym target |
4 |
> omits basename ..." for "dosym /path_to_file/file /some_other_path/" |
5 |
|
6 |
> The PMS says: |
7 |
|
8 |
> dosym: Creates a symbolic link named as for its second parameter, |
9 |
> pointing to the first. If the directory containing the new link does |
10 |
> not exist, creates it. Failure behaviour is EAPI dependent as per |
11 |
> section 12.3.3.1. |
12 |
|
13 |
> The part "Creates a symbolic link named as for its second parameter" |
14 |
> could qualify it for requiring a basename. |
15 |
|
16 |
Exactly. |
17 |
|
18 |
> As the behaviour of an implicit basename of the second parameter was |
19 |
> used for a long time I wonder if the PMS needs to be updated and |
20 |
> clarify the dosym behaviour or if this new QA warning can be seen as |
21 |
> a long overdue implementation of the intended behaviour. |
22 |
|
23 |
This has been discussed in bug 379899 and the conclusion was that PMS |
24 |
specifies the intended behaviour. |
25 |
|
26 |
At the moment a directory argument produces only a QA warning in |
27 |
Portage. I think the long-term plan is to turn it into an error |
28 |
though. |