| 1 |
On 5/23/20 1:41 PM, magic-gentoo@××××××××××××××.de wrote: |
| 2 |
> Am 23.05.2020 um 22:20 schrieb Zac Medico: |
| 3 |
>> On 5/23/20 1:02 PM, magic-gentoo@××××××××××××××.de wrote: |
| 4 |
>>> I rewrote e-file in python by using the portage API [1]. But loading the |
| 5 |
>>> API slows down the whole script. Is there any way to speed up my |
| 6 |
>>> implementation? Have I done something fundamentally wrong? |
| 7 |
>> |
| 8 |
>> When I patched the portage API out of your script, I saw the run time |
| 9 |
>> drop from 4.2 seconds to 3.2 seconds with this patch: |
| 10 |
>> ... |
| 11 |
>> |
| 12 |
>> Are your results worse than mine? |
| 13 |
> |
| 14 |
> Nope, but maybe the phrase "loading the API" was misleading. I'd like to |
| 15 |
> replace it with: "But using the API slows down the whole script.". This |
| 16 |
> means it is much slower to get the additional informations by portage |
| 17 |
> API than just grep'ing throught the ebuild files. If I run the python |
| 18 |
> e-file on my machine it takes 3.2 seconds for a single file. The bash |
| 19 |
> e-file show the same result within a second or so. |
| 20 |
> |
| 21 |
> regards |
| 22 |
> Daniel |
| 23 |
> |
| 24 |
> |
| 25 |
|
| 26 |
Since the portage API only added about 1 second to the python script |
| 27 |
time, I guess it's on par with your bash implementation. ;-P |
| 28 |
-- |
| 29 |
Thanks, |
| 30 |
Zac |
Archives