git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [EGIT PATCH 3/6] Add a method to get refs by object Id

From
Shawn O. Pearce <spearce@spearce.org>
Date
Oct 6, 2008, 08:15 UTC
Message-ID
<20081006081554.GD27516@spearce.org>
In-Reply-To
<1223249802-9959-4-git-send-email-robin.rosenberg@dewire.com>
Robin Rosenberg <robin.rosenberg@dewire.com> wrote:
Show 11 quoted lines
> diff --git a/org.spearce.jgit/src/org/spearce/jgit/lib/Repository.java b/org.spearce.jgit/src/org/spearce/jgit/lib/Repository.java
> index dfce1b8..3fc5236 100644
> --- a/org.spearce.jgit/src/org/spearce/jgit/lib/Repository.java
> +++ b/org.spearce.jgit/src/org/spearce/jgit/lib/Repository.java
> @@ -939,6 +940,33 @@ public String getBranch() throws IOException {
>  	}
>  
>  	/**
> +	 * @return a map with all objects referenced by a peeled ref.
> +	 */
> +	public Map<AnyObjectId, List<Ref>> getAllRefsByPeeledObjectId() {

Do we really want to promise List here? Can we make it just Collection instead?

> +		Map<String, Ref> allRefs = getAllRefs();
> +		HashMap<AnyObjectId, List<Ref>> ret = new HashMap<AnyObjectId, List<Ref>>(allRefs.size());
> +		for (Map.Entry<String,Ref> e : allRefs.entrySet()) {
> +			Ref ref = e.getValue();
I think this is cleaner:
	for (Ref ref : allRefs.values()) {
as you never use the key.
Show 15 quoted lines
> +			AnyObjectId target = ref.getPeeledObjectId();
> +			if (target == null)
> +				target = ref.getObjectId();
> +			List<Ref> list = ret.get(target);
> +			if (list == null) {
> +				list = Collections.singletonList(ref);
> +			} else {
> +				if (list.size() == 1) {
> +					ArrayList<Ref> list2 = new ArrayList<Ref>(2);
> +					list2.add(list.get(0));
> +					list = list2;
> +				}
> +				list.add(ref);
> +			}
> +			ret.put(target, list);

Hmm. Putting the list every time is pointless. This is may run faster because we (on average) only do one hash lookup per target, not 2:

	List<Ref> nl = Collections.singletonList(ref);
	List<Ref> ol = ret.put(target, nl);
	if (ol != null) {
		if (ol.size() == 1) {
			nl = new ArrayList<Ref>(2);
			nl.add(ol.get(0));
			nl.add(ref);
			ret.put(target, nl);
		} else {
			ol.add(ref)
			ret.put(target, ol);
		}
	}

The trick is that most targets have exactly one Ref, so the first ret.put will return null and we'll never worry about the expansion cases. In the rare cases where more than one Ref points at the same object we'll degrade into doing two hash lookups per target, which is no worse than what you have right now.

-- 
Shawn.
Previous: Shawn O. PearceNext: Robin Rosenberg
Message 11 of 14 in “(unknown)”
  1. Robin RosenbergOct 5, 2008
  2. 1/6 Keep original ref name when reading refsRobin Rosenberg, Oct 5, 2008
  3. 2/6 Peel annotated tags when getting all refsRobin Rosenberg, Oct 5, 2008
  4. 3/6 Add a method to get refs by object IdRobin Rosenberg, Oct 5, 2008
  5. 4/6 Add tags to the graphical history display.Robin Rosenberg, Oct 5, 2008
  6. 5/6 Add decorate option to log programRobin Rosenberg, Oct 5, 2008
  7. 6/6 Comment the getId method and hint for copy to actually get an ObjectIdRobin Rosenberg, Oct 5, 2008
  8. Shawn O. PearceOct 6, 2008
  9. Robin RosenbergOct 6, 2008
  10. Shawn O. PearceOct 6, 2008
  11. Shawn O. PearceOct 6, 2008
  12. Robin RosenbergOct 6, 2008
  13. Shawn O. PearceOct 6, 2008
  14. Shawn O. PearceOct 6, 2008

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.