resize() + onlyScaleDown() should not affect cache key if no resize happens.

[NightlyNexus]

No description provided.

[JakeWharton]

Makes sense. Implications are non-trivial I think...

[NightlyNexus]

yeah, pretty sure this is tricky. the cache key is created at request time but we'd need to edit it afterward for this.
i think it's a worthwhile optimization, though. i saw a case of it in our app.

maybe a new type of cache key.

[JakeWharton]

Changing to compute the actual key which represents the image instead of what was requested seems worthwhile

…

On Sat, Mar 16, 2019, 3:34 AM Eric Cochran ***@***.*** wrote: yeah, pretty sure this is tricky. the cache key is created at request time for the cache time but we'd need to edit it afterward for this. i think it's a worthwhile optimization, though. i saw a case of it in our app. maybe a new type of cache key. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2072 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEEEadyARUG4eUgTncbY9PY3OIKrpYUks5vXJ5xgaJpZM4b3EY9> .

[NightlyNexus]

on the other hand, if i call load(url).resize(500, 500).onlyScaleDown() and then call load(url).resize(500, 500).onlyScaleDown() again, i don't want to download the image again.
the cache key would need to match on both the original request and the result for this feature.

[JakeWharton]

Yep. We can start with something specific for this case, but it'd be nice to solve the case of loading a thumbnail from the cache for a larger image while the larger version decodes from disk cache. We could do this either with a series of secondary fallback keys or with some kind of cached key matcher.

…

On Mon, Mar 18, 2019 at 2:02 PM Eric Cochran ***@***.***> wrote: on the other hand, if i call load(url).resize(500, 500).onlyScaleDown() and then call load(url).resize(500, 500).onlyScaleDown() again, i don't want to redownload the image again. the cache key would need to match on both the original request and the result for this feature. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2072 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEEEUlzD577z9zb9pUWOcV3CFvpQQmvks5vX9S8gaJpZM4b3EY9> .