1 |
On Fri, Sep 27, 2013 at 4:44 PM, Michał Górny <mgorny@g.o> wrote: |
2 |
> Dnia 2013-09-26, o godz. 17:24:49 |
3 |
> Davide Pesavento <pesa@g.o> napisał(a): |
4 |
> |
5 |
>> On Thu, Sep 26, 2013 at 4:04 PM, Kent Fredric <kentfredric@×××××.com> wrote: |
6 |
>> > |
7 |
>> > On 26 September 2013 19:53, Michał Górny <mgorny@g.o> wrote: |
8 |
>> >> |
9 |
>> >> How do we handle packages which install multiple libraries? I'm afraid |
10 |
>> >> forcing such a policy and/or hurrying developers to adapt will only |
11 |
>> >> cause more of poppler-like issues to occur. |
12 |
>> > |
13 |
>> > |
14 |
>> > Can you give a an example package which: |
15 |
>> > |
16 |
>> > - installs multiple libraries |
17 |
>> > - has an ABI that may change for only one of those libraries |
18 |
>> > - it is sane / plausible to expect one downstream dependent *not* to |
19 |
>> > forcibly rebuild as a result of a chane in one of those libaries |
20 |
>> > - it is sane / plausible to expect a different downstream to forcibly |
21 |
>> > rebuild as a result of changes in one of those libraries |
22 |
>> > |
23 |
>> |
24 |
>> dev-python/PyQt4 |
25 |
>> |
26 |
>> Each module is a separate library, and each has its own ABI that can |
27 |
>> change independently from the others. Downstream projects that rely |
28 |
>> only on PyQt4's python API are not affected by ABI changes, but those |
29 |
>> (very few) that link against one or more modules (e.g. kde-base/pykde4 |
30 |
>> I think) must be rebuilt. |
31 |
> |
32 |
> How often does ABI of pyqt4 libraries change in such a way that rebuild |
33 |
> of pykde4 is not required? |
34 |
|
35 |
Practically never (see below). |
36 |
|
37 |
> |
38 |
> Looking at the dep: |
39 |
> |
40 |
>>=dev-python/PyQt4-4.9.5[${PYTHON_USEDEP},dbus,declarative,script(+),sql,svg,webkit,X] |
41 |
> |
42 |
> I'd think it's fairly rare when only the libraries not listed above |
43 |
> change ABI without any of the remaining ones changing it. |
44 |
> |
45 |
|
46 |
Actually, most PyQt4 libraries never changed their ABIs since the |
47 |
initial 4.0 release, so yes, it's "fairly rare" :) |
48 |
That's also one of the reasons why I never bothered to add a subslot to PyQt4. |
49 |
|
50 |
Thanks, |
51 |
Davide |