| 1 |
On 01/26/2018 12:47 AM, Michał Górny wrote: |
| 2 |
> Dnia 26 stycznia 2018 06:45:25 CET, Zac Medico <zmedico@g.o> napisał(a): |
| 3 |
>> On 01/23/2018 03:25 PM, Alec Warner wrote: |
| 4 |
>>> |
| 5 |
>>> |
| 6 |
>>> On Tue, Jan 23, 2018 at 4:48 PM, Michał Górny <mgorny@g.o |
| 7 |
>>> <mailto:mgorny@g.o>> wrote: |
| 8 |
>>> |
| 9 |
>>> Always group all Prefix keywords after other types of keywords. |
| 10 |
>> This |
| 11 |
>>> not only ensures that fbsd sorts first but more importantly |
| 12 |
>> stabilizes |
| 13 |
>>> the LHS output between regular and -P variant -- that is, -P |
| 14 |
>> always adds |
| 15 |
>>> additional keywords at the end. |
| 16 |
>>> --- |
| 17 |
>>> pym/gentoolkit/eshowkw/keywords_header.py | 16 ++++++++++------ |
| 18 |
>>> 1 file changed, 10 insertions(+), 6 deletions(-) |
| 19 |
>>> |
| 20 |
>>> diff --git a/pym/gentoolkit/eshowkw/keywords_header.py |
| 21 |
>>> b/pym/gentoolkit/eshowkw/keywords_header.py |
| 22 |
>>> index 41b8ba4..1b64bfd 100644 |
| 23 |
>>> --- a/pym/gentoolkit/eshowkw/keywords_header.py |
| 24 |
>>> +++ b/pym/gentoolkit/eshowkw/keywords_header.py |
| 25 |
>>> @@ -142,12 +142,16 @@ class keywords_header: |
| 26 |
>>> break |
| 27 |
>>> |
| 28 |
>>> # sort by, in order (to match Bugzilla): |
| 29 |
>>> - # 1. arch, then ~arch |
| 30 |
>>> - # 2. profile stability |
| 31 |
>>> - # 3. short keywords, then long (prefix, fbsd) |
| 32 |
>>> - # 4. keyword name in reverse component order |
| 33 |
>>> - normal.sort(key=lambda kw: (kw in |
| 34 |
>>> self.__TESTING_KW_ARCHS, |
| 35 |
>>> - levels.get(kw, 99), kw.count('-'), |
| 36 |
>>> list(reversed(kw.split('-'))))) |
| 37 |
>>> + # 1. non-prefix, then prefix (stable output |
| 38 |
>> between |
| 39 |
>>> -P and not) |
| 40 |
>>> + # 2. arch, then ~arch |
| 41 |
>>> + # 3. profile stability |
| 42 |
>>> + # 4. short keywords, then long (prefix, fbsd) |
| 43 |
>>> + # 5. keyword name in reverse component order |
| 44 |
>>> + normal.sort(key=lambda kw: (self.__isPrefix(kw), |
| 45 |
>>> + kw in self.__TESTING_KW_ARCHS, |
| 46 |
>>> + levels.get(kw, 99), |
| 47 |
>>> + kw.count('-'), |
| 48 |
>>> + list(reversed(kw.split('-'))))) |
| 49 |
>>> |
| 50 |
>>> |
| 51 |
>>> I'm a bit sad about this lambda because its ended up a bit long. |
| 52 |
>>> |
| 53 |
>>> What are your thoughts on splitting it out? |
| 54 |
>> |
| 55 |
>> The fact that it's a lambda doesn't bother me so much as the |
| 56 |
>> inefficiency of regenerating the key on every call. I've found this |
| 57 |
>> cute |
| 58 |
>> little memodict decorator that will optimize it nicely: |
| 59 |
>> |
| 60 |
>> http://code.activestate.com/recipes/578231-probably-the-fastest-memoization-decorator-in-the-/ |
| 61 |
>> |
| 62 |
>> def memodict(f): |
| 63 |
>> class memodict(dict): |
| 64 |
>> def __missing__(self, key): |
| 65 |
>> ret = self[key] = f(key) |
| 66 |
>> return ret |
| 67 |
>> return memodict().__getitem__ |
| 68 |
> |
| 69 |
> Do you really think saving a few microseconds is worth the added complexity? Because if you do, then I'd rather turn keyword into a class and cache everything properly rather than adding ugly hacks to optimize one call site. |
| 70 |
|
| 71 |
Actually, memoization is pointless here, because python's sort method |
| 72 |
only calls the key function once per element. |
| 73 |
-- |
| 74 |
Thanks, |
| 75 |
Zac |
Archives