[MediaWiki-commits] [Gerrit] Fixed bugs that manifested for diff current dumps - change (operations...incremental)

[Petr Onderka (Code Review)]

Petr Onderka has uploaded a new change for review.

  https://gerrit.wikimedia.org/r/85879


Change subject: Fixed bugs that manifested for diff current dumps
......................................................................

Fixed bugs that manifested for diff current dumps

Change-Id: Id4b020505d73e63c7f65f8c278b492b13551428b
---
M Diff/Changes/DeleteRevisionChange.h
M DumpObjects/TextGroup.cpp
M DumpObjects/TextGroup.h
M TextGroupsManager.cpp
4 files changed, 14 insertions(+), 8 deletions(-)


  git pull ssh://gerrit.wikimedia.org:29418/operations/dumps/incremental 
refs/changes/79/85879/1

diff --git a/Diff/Changes/DeleteRevisionChange.h 
b/Diff/Changes/DeleteRevisionChange.h
index 9f0c157..0ff300b 100644
--- a/Diff/Changes/DeleteRevisionChange.h
+++ b/Diff/Changes/DeleteRevisionChange.h
@@ -5,7 +5,7 @@
 /**
  * A change that represents deleting a revision.
  *
- * In the (probably impossible) case when a revision is deleted but its page 
isn't,
+ * In the case when a revision is deleted but its page isn't,
  * this change has to follow the corresponding page change.
  */
 class DeleteRevisionChange : public Change
diff --git a/DumpObjects/TextGroup.cpp b/DumpObjects/TextGroup.cpp
index bf7372f..8c09d73 100644
--- a/DumpObjects/TextGroup.cpp
+++ b/DumpObjects/TextGroup.cpp
@@ -34,15 +34,13 @@
     }
     else
     {
-        texts = Read(revisionOffset);
-
+        texts = Read(revisionOffset, savedLength);
         savedOffset = revisionOffset.value;
-        savedLength = NewLength();
         isEditable = false;
     }
 }
 
-std::vector<std::string> TextGroup::Read(Offset offset)
+std::vector<std::string> TextGroup::Read(Offset offset, std::uint32_t& length)
 {
     auto &stream = *(dump.lock()->stream);
     stream.seekp(offset.value);
@@ -57,6 +55,9 @@
 
     auto compressed = DumpTraits<std::string>::ReadLong(stream);
     auto textsString = SevenZip::Decompress(compressed);
+
+    length = (std::uint64_t)stream.tellp() - offset.value;
+
     return split(textsString, '\0');
 }
 
diff --git a/DumpObjects/TextGroup.h b/DumpObjects/TextGroup.h
index 41bd1e4..66ea643 100644
--- a/DumpObjects/TextGroup.h
+++ b/DumpObjects/TextGroup.h
@@ -25,7 +25,7 @@
 
     void EnsureCompressed();
     void Load();
-    std::vector<std::string> Read(Offset offset);
+    std::vector<std::string> Read(Offset offset, std::uint32_t& length);
 
     bool IsEmpty() const;
 
diff --git a/TextGroupsManager.cpp b/TextGroupsManager.cpp
index 02ace26..e1b9229 100644
--- a/TextGroupsManager.cpp
+++ b/TextGroupsManager.cpp
@@ -37,11 +37,16 @@
 
 void TextGroupsManager::EndGroup(DiffWriter* diffWriter)
 {
+    if (currentGroup == nullptr)
+        return;
+
     if (textGroupModified)
+    {
         currentGroup->Write();
 
-    if (diffWriter != nullptr)
-        diffWriter->SetTextGroup(currentGroup->GetCompressedTexts());
+        if (diffWriter != nullptr)
+            diffWriter->SetTextGroup(currentGroup->GetCompressedTexts());
+    }
 
     currentGroup = nullptr;
     textGroupModified = false;

-- 
To view, visit https://gerrit.wikimedia.org/r/85879
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings

Gerrit-MessageType: newchange
Gerrit-Change-Id: Id4b020505d73e63c7f65f8c278b492b13551428b
Gerrit-PatchSet: 1
Gerrit-Project: operations/dumps/incremental
Gerrit-Branch: gsoc
Gerrit-Owner: Petr Onderka <gsv...@gmail.com>

_______________________________________________
MediaWiki-commits mailing list
MediaWiki-commits@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/mediawiki-commits