Brave-browser: Pubs which are already in a-c table are getting deleted on upgraded profile after revisiting

Created on 30 Jan 2019  路  12Comments  路  Source: brave/brave-browser

Description

Follow up of #3134
Pubs which are already in a-c table are getting deleted on upgraded profile after revisiting

Steps to Reproduce

  1. Clean profile 0.58.21
  2. Enable rewards and restore wallet
  3. Change the Rewards Default settings ( Set the min page time to 5secs)
  4. Add few sites to a-c table ( I have added 5 verified+nonverified sites)
  5. Tip a verified publisher ( I have tipped only one site)
  6. Make sure all the sites are added to a-c table
  7. Upgrade profile to 0.59.32
  8. Revisit the pubs which are already in a-c table ( step 4)

Actual result:

when revisiting publishers that are already in the table, publishers are getting removed from a-c table.

Expected result:

pubs which are in a-c table should not get deleted after revisit

Reproduces how often:

Easy

Brave version (brave://version info)

Brave | 0.59.32 Chromium: 72.0.3626.81聽(Official Build)聽(64-bit)
-- | --
Revision | ac8b982e05014492d1bd7d317628a4f22a97ffa0-refs/branch-heads/3626@{#796}
OS | Windows 10

Reproducible on current release:

  • Does it reproduce on brave-browser dev/beta builds? no

Website problems only:

  • Does the issue resolve itself when disabling Brave Shields? na
  • Is the issue reproducible on the latest version of Chrome? na

Additional Information

@NejcZdovc @brave/legacy_qa

QA Pass-Linux QA Pass-Win64 QA Pass-macOS QYes bug featurrewards prioritP2 release-noteinclude

All 12 comments

@GeetaSarvadnya can you please paste before and after ac table

@NejcZdovc Please find the below screen shots:

Before:

UI:
image

publisher_info_db:

image

After:

UI:

image

publisher_info_db:
image

Reproduced on 0.58.21 --> 0.59.32 update on macOS. Here are my publisher_info_db tables
0.58.21:
screen shot 2019-01-30 at 10 10 38 am

0.59.32 (after visiting a site to trigger recalculation):
screen shot 2019-01-30 at 10 34 27 am

Steps:

  1. Install 0.58.21 and enable rewards. Change Auto-Contribute Settings default settings for page time (change from 8s to 5s)
  2. Visit 3 sites (clifton, ddg, google)
  3. Close and relaunch Brave
  4. Visit clifton --> a-c % does not change.
  5. Visit nytimes and cnn. They are added to a-c table with 0%.
  6. Visit publisher_info_db to see negative scores.
  7. Update to 0.59.32
  8. publisher_info_db still has negative scores.
  9. Visit a site in the a-c table (I used ddg) for at least min time.
  10. All % values recalculate but google is dropped from the table.

@LaurenWags Able to reproduce the above steps on Windows 10 x64

Negative scores are stores in a-c table when we relaunch and add few more sites to a-c table in 0.58.21

image

Additionally, if you follow these modified steps (no change to settings) you get % values which are off after updating to 0.59.32:

  1. Install 0.58.21 and enable rewards. Do not change Auto-Contribute Settings default settings for page time (no change of 8s to 5s)
  2. Visit 3 sites (clifton, ddg, google)
  3. Close and relaunch Brave
  4. Visit clifton --> a-c % does change.
  5. Visit nytimes and cnn. They are added to a-c table with appropriate % values
  6. Visit publisher_info_db (no negative scores)
  7. Update to 0.59.32
  8. publisher_info_db still fine.
  9. Visit a site in the a-c table (I used ddg) for at least min time.
  10. All % values recalculate. DDG appears to have a way off % value.

0.58.21:
screen shot 2019-01-30 at 11 25 45 am

Update to 0.59.32 and visit DDG:
screen shot 2019-01-30 at 11 29 23 am

Verified passed with

Brave | 0.59.33 Chromium: 72.0.3626.81聽(Official Build)聽(64-bit)
-- | --
Revision | ac8b982e05014492d1bd7d317628a4f22a97ffa0-refs/branch-heads/3626@{#796}
OS | Mac OS X

  • Verified steps from https://github.com/brave/brave-browser/issues/3162#issuecomment-458991004 (Verified when a profile from 0.58.21 with issue as described in #3134 is updated to 0.59.33, and a new site is visited, the a-c table is correctly displayed - no sites lost, % values are accurate, weights in publisher_info_db are accurate)
  • Verified steps from https://github.com/brave/brave-browser/issues/3162#issuecomment-459011628 (Verified when a profile from 0.58.21 without issue as described in #3134 is updated to 0.59.33, and a new site is visited, the a-c table is correctly displayed - no sites lost, % values are accurate, weights in publisher_info_db are accurate)

Verification passed on

Brave | 0.59.33 Chromium: 72.0.3626.81聽(Official Build)聽(64-bit)
-- | --
Revision | ac8b982e05014492d1bd7d317628a4f22a97ffa0-refs/branch-heads/3626@{#796}
OS | Windows 10

  • Verified the STR mentioned in description
  • Verified steps from https://github.com/brave/brave-browser/issues/3162#issuecomment-458991004 (Verified when a profile from 0.58.21 with issue as described in #3134 is updated to 0.59.33, and a new site is visited, the a-c table is correctly displayed - no sites lost, % values are accurate, weights in publisher_info_db are accurate)
  • Verified steps from https://github.com/brave/brave-browser/issues/3162#issuecomment-459011628 (Verified when a profile from 0.58.21 without issue as described in #3134 is updated to 0.59.33, and a new site is visited, the a-c table is correctly displayed - no sites lost, % values are accurate, weights in publisher_info_db are accurate)

Verified passed with

Brave | 0.59.33 Chromium: 72.0.3626.81聽(Official Build)聽(64-bit)
-- | --
Revision | ac8b982e05014492d1bd7d317628a4f22a97ffa0-refs/branch-heads/3626@{#796}
OS | Linux

  • Verified steps from https://github.com/brave/brave-browser/issues/3162#issuecomment-458991004 (Verified when a profile from 0.58.21 with issue as described in #3134 is updated to 0.59.33, and a new site is visited, the a-c table is correctly displayed - no sites lost, % values are accurate, weights in publisher_info_db are accurate)
  • Verified steps from https://github.com/brave/brave-browser/issues/3162#issuecomment-459011628 (Verified when a profile from 0.58.21 without issue as described in #3134 is updated to 0.59.33, and a new site is visited, the a-c table is correctly displayed - no sites lost, % values are accurate, weights in publisher_info_db are accurate)

Verified comments from (https://github.com/brave/brave-browser/issues/3162#issuecomment-458991004).
verified a profile from 0.58.21 with issues described in #3134 is updated to 0.59.33. Below are my observations

  1. Rightafter update to 0.59.33, the sites (cnn and nytimes) added in step 5 got dropped? when i revisit any of the sites which are already added in a-c table (step2), the sites (cnn and nytimes) are getting re-added to a-c table. - FAIL
  2. If i re-visit the sites which are already added in a-c table, % value calculated correctly and weights displayed in DB are accurate - PASS
  3. If i add new site, sites are getting added with correct % attention value and and weights displayed in DB are accurate - PASS

@NejcZdovc can you please look into the 1st point, which seems wrong to me, correct me if i am wrong.

@NejcZdovc please correct if wrong, but I think what @GeetaSarvadnya is seeing is expected because she has not yet visited any sites and the code to correct the a-c table (scores, % values, and weights) does not run until you visit a site. (also, we're not seeing any 0% items in the a-c table in 0.59.x due to https://github.com/brave/brave-browser/issues/2365)

@LaurenWags you are correct and this is expected

@NejcZdovc shouldn't the recalculations happen automatically upon upgrade which basically fixes the issue? We shouldn't wait until the user visits a site for the recalculation to happen with this fix. Could this be fixed for 0.60.x ?

@srirambv I think it doesn't matter as it will automatically be fixed when user do any action. So data will not be wrong for any operation that follows.

Was this page helpful?
0 / 5 - 0 ratings

Related issues

fmarier picture fmarier  路  3Comments

bbondy picture bbondy  路  3Comments

kjozwiak picture kjozwiak  路  3Comments

AlexCombas picture AlexCombas  路  3Comments

simonhong picture simonhong  路  3Comments