[PATCH 02 of 10] transaction: only generate file when we actually close the transaction

Matt Mackall mpm at selenic.com
Wed Oct 29 16:40:52 CDT 2014

On Tue, 2014-10-28 at 16:57 +0100, Pierre-Yves David wrote:
> On 10/28/2014 04:41 PM, Pierre-Yves David wrote:
> > # HG changeset patch
> > # User Pierre-Yves David <pierre-yves.david at fb.com>
> > # Date 1413606348 25200
> > #      Fri Oct 17 21:25:48 2014 -0700
> > # Branch stable
> > # Node ID 74f8131d57fbe539ed5fe16d84230b47ef328ddb
> > # Parent  5ccb5eb0a6dedf6910ed95434b6dde95b5e473c5
> > transaction: only generate file when we actually close the transaction
> These two first patch are simple and safe.

Queued these. Though see below about refactoring.

> The 8 others are more impactful but solve some of the current 
> inconsistency with changeglog and nested transaction. Nested transaction 
> are more common nowaday so even if we never had report for such bug they 
> may impact some of our user with 3.2

Pretty terrified to do this much rewriting to fairly subtle code in
stable. 'Refactor' is not a word I like to see in this context: if it's
complex enough that it needs refactoring first to fix, it's probably too
complex to fix in stable.

So now that I've seen these, I'm afraid this is going to have to wait
until after 3.2 (in smaller batches than 10, please).

> More transaction work is ready and scheduled for 3.3

I think we'll need a TransactionPlan page on the wiki that gives an
overview of the mechanics and semantics.

Mathematics is the supreme nostalgia of our time.

More information about the Mercurial-devel mailing list