smiklosovic commented on code in PR #4278:
URL: https://github.com/apache/cassandra/pull/4278#discussion_r2254612569
##########
src/java/org/apache/cassandra/io/sstable/format/TOCComponent.java:
##########
@@ -78,20 +80,36 @@ public static Set<Component> loadTOC(Descriptor descriptor,
boolean skipMissing)
}
/**
- * Appends new component names to the TOC component.
+ * Updates the TOC file by reading existing component entries, merging
them with the given components,
+ * sorting the combined list in lexicographic order for deterministic
output.
+ *
+ * @param descriptor the SSTable descriptor for which to update the TOC
+ * @param components new components to merge into the TOC (existing TOC
entries are preserved)
+ * @throws FSWriteError if an I/O error occurs when creating or
overwriting the TOC file
*/
- public static void appendTOC(Descriptor descriptor, Collection<Component>
components)
+ public static void updateTOC(Descriptor descriptor, Collection<Component>
components)
{
+ if (components.isEmpty())
+ return;
+
File tocFile = descriptor.fileFor(Components.TOC);
- try (FileOutputStreamPlus out = tocFile.newOutputStream(APPEND);
- PrintWriter w = new PrintWriter(out))
+
+ Set<String> componentNames = new
TreeSet<>(Collections2.transform(components, Component::name));
+
+ if (tocFile.exists())
{
- for (Component component : components)
- w.println(component.name);
- w.flush();
- out.sync();
+ List<String> lines = FileUtils.readLines(tocFile);
+ if (lines.isEmpty())
Review Comment:
Also, if we fail to read, then it will return empty list. Empty list might
be returned both when file is genuinely empty as well as when it failed to be
read from.
So we might indeed just proceed and add empty `lines` to `componentNames` in
case we fail to read. But then, this file might already contain something, we
just failed to read it. So we would not update `lines` with its content and
then in `write` we might overwrite it even though it is quite improbable we
would read with error but we would write successfully.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: pr-unsubscr...@cassandra.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: pr-unsubscr...@cassandra.apache.org
For additional commands, e-mail: pr-h...@cassandra.apache.org
