ubuntu-phone team mailing list archive
-
ubuntu-phone team
-
Mailing list archive
-
Message #11321
Re: webapps and automated cache expiration
On Thu, Feb 26, 2015 at 9:46 AM, Christian Dywan
<christian.dywan@xxxxxxxxxxxxx> wrote:
> On 24.02.2015 11:42, Olivier Tilloy wrote:
>> On Mon, Feb 23, 2015 at 5:48 PM, Olivier Tilloy
>> <olivier.tilloy@xxxxxxxxxxxxx> wrote:
>>> On Mon, Feb 23, 2015 at 5:40 PM, Ted Gould <ted@xxxxxxxxxx> wrote:
>>>> On Mon, 2015-02-23 at 16:35 +0100, Oliver Grawert wrote:
>>>>
>>>> after now about 3 months of constant use of my bq phone, i took a deeper
>>>> look on the weekend at the accumulated cache data:
>>>>
>>>> phablet@ubuntu-phablet:~$ du -hcs .local/share
>>>> 2,0G .local/share
>>>>
>>>> Seems really odd that these are in ~/.local/share ? Shouldn't it be ~/.cache
>>>> ? I don't think we should ever autoexpire anything in ~/.local/share because
>>>> that's in theory where user data lives.
>>> Yup, that’s https://launchpad.net/bugs/1424726 (just filed it, I
>>> thought it had been reported already but apparently not).
>> And fixed with https://code.launchpad.net/~osomon/webbrowser-app/cacheLocation/+merge/250730.
>>
>> However while testing the fix, I observed that oxide writes the
>> "Cache" directory to ~/.cache/webbrowser-app/ (good), but it still
>> writes other directories like "GPUCache" and "Local Storage" to
>> ~/.local/share/webbrowser-app/, which looks suspicious to me. At least
>> it is consistent with what chromium does on my desktop.
>>
> That makes perfect sense as far as local storage goes because user data
> is stored there. If apps also use it for caching that's because there's
> no cache API.
Right, "Local Storage" makes sense. Not so sure about GPUCache (unless
the name is misleading), but it’s consistent with what chromium does
on desktop anyway, so it should be fine.
References