Fix(bulk): Correctly append all the slices of cbuf while reduce phase

[ahsanbarkati]

The bulk loader was not considering the last slice of cbuf in appendToList
function. In condition next >= 0 && (next < end || end == -1) the part end == -1 ensures that the last slice is considered because for the last slice, end is -1 hence next < end is not true. This fixes the issue of index not being seen in a bulk loaded data, the index was missing because index key is the last key and this bug causes the last key to get missing.

---

This change is 

[netlify]

✔️ Deploy preview for dgraph-docs ready!

🔨 Explore the source changes: a2d4528

🔍 Inspect the deploy logs: https://app.netlify.com/sites/dgraph-docs/deploys/5fd248d98c35150007061a0a

😎 Browse the preview: https://deploy-preview-7110--dgraph-docs.netlify.app

[manishrjain]

Reviewed 1 of 1 files at r1, 1 of 1 files at r2.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @vvbalaji-dgraph)

[danielmai]

This bug fix applies to the data keys as well, not just the index. This is the count of nodes per predicate from bulk loader for the 21-million movie data set before and after this PR. Some predicates were previously missing a single node.

$ diff -u 1.txt 2.txt
--- 1.txt	2020-12-25 10:40:19.780857171 -0800
+++ 2.txt	2020-12-25 10:39:53.564583223 -0800
@@ -12,7 +12,7 @@
       6 collection.films_in_collection
     116 collections
     321 company.films
-     49 company_role_or_service.companies_performing_this_role_or_service
+     50 company_role_or_service.companies_performing_this_role_or_service
     142 content_rating.country
       9 content_rating.minimum_accompanied_age
     121 content_rating.minimum_unaccompanied_age
@@ -31,7 +31,7 @@
       1 cut.release_region
       4 cut.runtime
       2 cut.type_of_cut
-3508685 dgraph.type
+3508686 dgraph.type
   95043 director.film
      10 distribution_medium.films_distributed_in_this_medium
     877 distributor.films_distributed
@@ -66,13 +66,13 @@
      50 http://www.w3.org/2002/07/owl#inverseOf
  240856 initial_release_date
    7447 job.films_with_this_crew_job
- 191562 language
+ 191563 language
    2157 location.featured_in_films
      24 locations
    4952 metacritic_id
   89073 music
   34751 music_contributor.film
-2116315 name
+2116316 name
   72998 netflix_id
       1 nytimes_id
     386 other_companies

[ahsanbarkati]

Yes @danielmai, the issue was that the last slice of the buffer was not processed. It has nothing to do with the indices.