- 19 8月, 2017 6 次提交
-
-
由 judds 提交于
------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=165717599
-
由 judds 提交于
From adb shell dumpsys meminfo com.google.android.apps.photos on a 2016 Pixel on OPM1.170816.001: Before: App Summary Pss(KB) ------ Java Heap: 20360 Native Heap: 66032 Code: 93336 Stack: 1192 Graphics: 85892 Private Other: 8956 System: 19944 TOTAL: 295712 TOTAL SWAP PSS: 82 After: App Summary Pss(KB) ------ Java Heap: 20456 Native Heap: 39756 Code: 93384 Stack: 1220 Graphics: 81460 Private Other: 9024 System: 11662 TOTAL: 256962 TOTAL SWAP PSS: 81 These numbers aren't super solid. Some extra invalidations can dump more Bitmaps into our Bitmap pool which will make it look like the steady state memory usage is increasing even though the maximum amount remains the same. That said, I see between a 33% and 50% improvement in native heap usage after this change. This was tested by flinging back and forth in the All grid in Photos. This change is limited to HARDWARE Bitmap support. We ought to also be able to reduce the bitmap pool size in O+ now that we only care about re-using Bitmaps for very small images or while generating thumbnails for the first time. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=165717498
-
由 judds 提交于
------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=165653043
-
由 judds 提交于
------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=165650311
-
由 loran 提交于
------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=165002367
-
由 judds 提交于
------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=164752684
-
- 18 8月, 2017 3 次提交
-
-
由 Sam Judd 提交于
Fixes #2266.
-
由 Sam 提交于
Running requests are left to complete to avoid cancelling and restarting expensive operations (network or I/O). Completed requests are asked to begin again, but short cut the normal request pipeline and redeliver the previously loaded resource to ensure that Targets and RequestListeners are notified as expected. Failed, cancelled, paused etc requests (all other states) are simply restarted.
-
由 Sam Judd 提交于
-
- 16 8月, 2017 2 次提交
-
-
由 Sam Judd 提交于
-
由 Sam Judd 提交于
This avoids writing partial data if the buffer is not at position 0 when the method is called. This fixes a bug where using DiskCacheStrategy.ALL or DiskCacheStrategy.RESOURCE with GifDrawables would result in partial and invalid gif files being written to Glide’s disk cache.
-
- 15 8月, 2017 4 次提交
-
-
由 Sam Judd 提交于
Fixes #2237.
-
由 Sam Judd 提交于
-
由 Dejan Stefanovic 提交于
* Added phone lookup for local URIs and updated sample to use it. Improved license handling in sample * Checkstyle fixes Fixing issues reported by checkstyle on travis * Code formatting Updated formatting after code review. Removed line of code that was commented-out. Added comment for permission request.
-
由 Sam Judd 提交于
Fixes #2240.
-
- 12 8月, 2017 2 次提交
-
-
由 Sam Judd 提交于
RecyclerView breaks badly when you run ViewAnimation on its views outside of an ItemAnimator. As a result, Glide’s cross fades are broken when used for loads into ImageViews in RecyclerView. Instead of using ViewAnimations when no placeholder is set in the cross fade, we can instead default to cross fading from a transparent ColorDrawable. Using the ColorDrawable avoids the issues with ViewAnimations in RecyclerView and provides a similar looking effect.
-
由 Sam Judd 提交于
Fixes #2182.
-
- 10 8月, 2017 1 次提交
-
-
由 Sam Judd 提交于
-
- 08 8月, 2017 5 次提交
- 07 8月, 2017 1 次提交
-
-
由 Sam Judd 提交于
Fixes #2212.
-
- 05 8月, 2017 2 次提交
-
-
由 Sam Judd 提交于
If a request is already in progress we should only cancel and restart it if the new request differs from the in progress request in some way. Prior to this change, we always cleared and restarted requests without checking to see if the request parameters had changed. Progress toward #2194.
-
由 Tolriq 提交于
Returning null from those will trigger obscure crash in MultiClassKey, annotation helps preventing that.
-
- 04 8月, 2017 1 次提交
-
-
由 Thomas Keller 提交于
See #2086
-
- 03 8月, 2017 2 次提交
-
-
由 christianw 提交于
replace_string \ 'import org.robolectric.res.builder.RobolectricPackageManager;' \ 'import org.robolectric.shadows.ShadowPackageManager;' replace_string \ 'import static org.robolectric.RuntimeEnvironment.getRobolectricPackageManager;' \ 'import static org.robolectric.Shadows.shadowOf;' replace_string \ 'RuntimeEnvironment.getRobolectricPackageManager\(\)' \ 'shadowOf(RuntimeEnvironment.application.getPackageManager())' replace_string 'RobolectricPackageManager' 'ShadowPackageManager'. ... plus fixes for line length and readability. Compatibility for [] and https://github.com/robolectric/robolectric/pull/3203 Tested: TAP --sample for global presubmit queue [] ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=162373742
-
由 Sam Judd 提交于
-
- 02 8月, 2017 1 次提交
-
-
由 Sam Judd 提交于
-
- 01 8月, 2017 4 次提交
-
-
由 Sam Judd 提交于
-
由 Sam Judd 提交于
Fixes #2121.
-
由 Robert Theis 提交于
-
由 Sam Judd 提交于
Fixes #1981.
-
- 21 7月, 2017 2 次提交
-
-
由 Sam Judd 提交于
-
由 François Blavoet 提交于
-
- 15 7月, 2017 1 次提交
-
-
由 Sam Judd 提交于
The Accept-Encoding header was added to ensure that we receive content length headers in responses. The Performance section of HttpUrlConnection’s documentation https://goo.gl/o2jeUM) indicates that unless Accept-Encoding is set to ‘identity’, the content length of the response will be cleared. In practice this appears to be incorrect for all API levels that Glide v4 supports (14+) and content length is always populated. Passing in ‘identity’ as the encoding header is problematic because it disables the default compression in HttpUrlConnection and some servers disregard the header and provide compressed data anyway. There’s not much point in adding extra complexity to Glide to handle compressed data if the ‘identity’ header isn’t actually required to get valid content lengths for images. Fixes #2132.
-
- 14 7月, 2017 1 次提交
-
-
由 Sam Judd 提交于
-
- 13 7月, 2017 2 次提交
-
-
由 judds 提交于
If we set the singleton before calling registerComponents, other threads that call get() while registerComponents is being called may receive a half initialized Glide object. Half initialized Glide objects may be missing critical components and using them in other threads may break lots of things. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=161232290
-
由 judds 提交于
Tested: TAP train for global presubmit queue [] ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=161135330
-