##// END OF EJS Templates
relnotes: add 6.8rc0
Raphaël Gomès -
r52540:56eb076a default
parent child Browse files
Show More
@@ -0,0 +1,72 b''
1 = Mercurial 6.8rc0 =
2
3 /!\ This is a tentative release, any and all notes below are subject to change or removal.
4
5 As usual, a *lot* of patches don't make it to this list.
6
7 == New Features or performance improvements ==
8
9 * Phases have been reworked to improve their general performance
10 * revset: stop serializing node when using "%ln"
11 * phases: convert remote phase root to node while reading them
12 * phases: use revision number in new_heads
13 * phases: use revision number in analyze_remote_phases
14 * phases: stop using `repo.set` in `remotephasessummary`
15 * phases: move RemotePhasesSummary to revision number
16 * phases: use revision number in `_pushdiscoveryphase`
17 * phases: introduce a performant efficient way to access revision in a set
18 * phases: rework the logic of _pushdiscoveryphase to bound complexity
19 * The Rust working copy code is being used by more places now:
20 * matchers: support patternmatcher in rust
21 * dirstate: remove the python-side whitelist of allowed matchers
22 * stream-clone: disable gc for `_entries_walk` duration
23 * stream-clone: disable gc for the initial section for the v3 format
24 * postincoming: avoid computing branchhead if no report will be posted
25 * stream-clone: disable gc for the entry listing section for the v2 format
26 * perf: allow profiling of more than one run
27 * perf: run the gc before each run
28 * perf: start recording total time after warming
29 * perf: clear vfs audit_cache before each run
30 * outgoing: rework the handling of the `missingroots` case to be faster
31 * outgoing: add a simple fastpath when there is no common
32 * tags-cache: skip the filternode step if we are not going to use it
33 * tags-cache: directly operate on rev-num warming hgtagsfnodescache
34 * tags-cache: directly perform a monimal walk for hgtagsfnodescache warming
35 * exchange: improve computation of relevant markers for large repos
36
37
38 == New Experimental Features ==
39
40 * Introduce a new experimental branch cache "v3":
41 * branchcache: add more test for the logic around obsolescence and branch heads
42 * branchcache: skip entries that are topological heads in the on disk file
43 * branchcache: add a "pure topological head" fast path
44 * branchcache: allow to detect "pure topological case" for branchmap
45
46
47 == Bug Fixes ==
48
49 * perf-stream-locked-section: actually use v1 generation when requested
50 * perf-stream-locked-section: fix the call to the v3 generator
51 * perf-stream-locked-section: advertise the right version key in the help
52 * stream: in v3, skip the "size" fast path if the entries have some unknown size
53 * stream-clone: stop getting the file size of all file in v3
54 * streamclone: stop listing files for entries that have no volatile files
55 * perf-stream-consume: use the source repository config when applying
56 * bundle: do no check the changegroup version if no changegroup is included
57 * perf: create the temporary target next to the source in stream-consume
58 * bundlespec: fix the "streamv2" and "streamv3-exp" variant
59 * push: rework the computation of fallbackheads to be correct
60 * profiler: flush after writing the profiler output
61 * base-revsets: use an author that actually exercises a lot of changesets
62 * hgrc: search XDG_CONFIG_HOME on mac
63 * clonebundles: add missing newline to legacy response
64 * narrow: add a test for linkrev computation done during widen
65
66 == Backwards Compatibility Changes ==
67
68 == Internal API Changes ==
69
70 == Miscellaneous ==
71
72 * obsolete: quote the feature name No newline at end of file
General Comments 0
You need to be logged in to leave comments. Login now