Re: [EGIT PATCH 1/9] Create a listener structure for changes to refs and index

From: Shawn O. Pearce <spearce@spearce.org>
Date: Jul 11, 2008, 04:22 UTC
Message-ID: <20080711042251.GB32633@spearce.org>
In-Reply-To: <1215729651-26781-2-git-send-email-robin.rosenberg@dewire.com>

Robin Rosenberg <robin.rosenberg@dewire.com> wrote:

> This version does not tell you which refs have changed, nor
> what changes have happened. There is not scanning for externally
> initiated changes either, though such changes can be found when
> a JGit client wants to read index or refs information.

Show 26 quoted lines

> diff --git a/org.spearce.jgit/src/org/spearce/jgit/lib/RefDatabase.java b/org.spearce.jgit/src/org/spearce/jgit/lib/RefDatabase.java
> index 9e3e020..4be33b8 100644
> --- a/org.spearce.jgit/src/org/spearce/jgit/lib/RefDatabase.java
> +++ b/org.spearce.jgit/src/org/spearce/jgit/lib/RefDatabase.java
> @@ -81,6 +81,10 @@ class RefDatabase {
>  
>  	private long packedRefsLength;
>  
> +	long lastRefModification;
> +
> +	long lastNotifiedRefModification;
> +
>  	RefDatabase(final Repository r) {
>  		db = r;
>  		gitDir = db.getDirectory();
> @@ -155,6 +161,12 @@ class RefDatabase {
>  		}
>  		if (!lck.commit())
>  			throw new ObjectWritingException("Unable to write " + name);
> +		setModified();
> +		db.fireRefsMaybeChanged();
> +	}
> +
> +	void setModified() {
> +		lastRefModification = System.currentTimeMillis();
>  	}

We don't ever actually need the time here as a time, do we? Wouldn't a simple counter work just as well, but ensure we don't lose notifications that occur more frequently then the resolution of currentTimeMillis()?

-- 
Shawn.