Boris Feld <boris.feld@octobus.net> [Tue, 26 Sep 2017 15:55:01 +0200] rev 34330
pull: remove inadequate use of operations records to update stepdone
The 'stepdone' set is design to be a client side mechanism. If the client used
some advanced capabilities to request necessary information (changeset,
obsmarkers, phases, etc). It marks the steps as done to avoid having a less
advanced mechanism issue a duplicated request.
So, the "stepdone.add('phases')" should be the result of a client choice,
because only the client can know it has requested all it needed to request. In
4a08cf1a2cfe this principle was broken because any phase-heads part sent by
the server to the client would declare the phases retrieval complete.
Now that there is an official phases related capability and code associated to
it. We do not need the change in 4a08cf1a2cfe anymore and we can back it out.
This brings back 'stepdone' management for 'phases' in line with the rest of
the code (including other phases handing).
Here is an example of potential misbehavior that 4a08cf1a2cfe introduced:
Imagine a server that pre-computes bundles. The bundles contains a changegroup
part and an (advisory) 'phase-heads' part. When a pull occurs, precomputed
bundled are reused if available. As the phase part is advisory it can be sent
to all clients. However they could be relevant changesets without phase
information. Either because they are already common or because they had no
precomputed bundle for them yet.
If receiving any 'phase-heads' parts disable subsequent phases re-trivial
parts, the client will not request phase data for all relevant changesets. For
example common changesets will not turn public.
Boris Feld <boris.feld@octobus.net> [Sun, 24 Sep 2017 21:27:18 +0200] rev 34329
pull: use 'phase-heads' to retrieve phase information
A new bundle2 capability 'phases' has been added. If 'heads' is part of the
supported value for 'phases', the server supports reading and sending 'phase-
heads' bundle2 part.
Server is now able to process a 'phases' boolean parameter to 'getbundle'. If
'True', a 'phase-heads' bundle2 part will be included in the bundle with phase
information relevant to the whole pulled set. If this method is available the
phases listkey namespace will no longer be listed.
Beside the more efficient encoding of the data, this new method will greatly
improve the phase exchange efficiency for repositories with non-served
changesets (obsolete, secret) since we'll no longer send data about the
filtered heads.
Add a new 'devel.legacy.exchange' config item to allow fallback to the old
'listkey in bundle2' method.
Reminder: the pulled set is not just the changesets bundled by the pull. It
also contains changeset selected by the "pull specification" on the client
side (eg: everything for bare pull). One of the reason why the 'pulled set' is
important is to make sure we can move -common- nodes to public.
Boris Feld <boris.feld@octobus.net> [Wed, 20 Sep 2017 18:29:10 +0200] rev 34328
bundle2: only grab a transaction when 'phase-heads' affect the repository
The next patch will use the 'phase-heads' part to exchange phase data relevant to
the pulled set.
'handlephases' currently acquires a transaction even in case of no-op pull,
which would results in an empty transaction and messing with the existing
journal.
Pass the transaction fetcher to updatephases so it can fetch it if necessary.
Boris Feld <boris.feld@octobus.net> [Tue, 19 Sep 2017 22:23:41 +0200] rev 34327
phases: move the binary decoding function in the phases module
We move the decoding function near the encoding one in a place where they can
be reused in other place (current target, 'exchange.py').
Boris Feld <boris.feld@octobus.net> [Tue, 19 Sep 2017 22:01:31 +0200] rev 34326
phases: move binary encoding into a reusable function
We want to use binary phases for pushing and pulling. We extract the encoding
function out of the bundle2 module first.
Boris Feld <boris.feld@octobus.net> [Tue, 19 Sep 2017 22:08:09 +0200] rev 34325
phases: use a Struct object for binary encoding and decoding
We will move the binary encoding and decoding code to 'phases.py' in order to
make it easier to reuse. First, let's cleanup it a bit.
Boris Feld <boris.feld@octobus.net> [Wed, 20 Sep 2017 05:47:33 +0200] rev 34324
discovery: avoid dropping remote heads hidden locally
An extra post processing was added to recognize remote heads that are hidden
locally as "common" instead of "unknown". However, this processing was
removing such hidden heads from the remote heads sets.
It had no impact because we used to pull phase information from all remote
heads.
This series will replace the phase pulling operation to a more efficient
process but requires the unmodified pulled set information.
Jun Wu <quark@fb.com> [Fri, 29 Sep 2017 11:41:24 -0700] rev 34323
test-patchbomb: use mocktime
The test was using system time for displaying ETAs, which could be flaky if
the sysload is high. This patch extracts mocktime.py from test-progress.t to
make sure test-patchbomb.t is unaffected by system time.
Differential Revision: https://phab.mercurial-scm.org/D844
Jun Wu <quark@fb.com> [Thu, 28 Sep 2017 10:37:53 -0700] rev 34322
test-patchbomb: fix the test
With the experimental config `progress.estimate` removed, the progress
output in `test-patchbomb.t` has a minor change: it shows ETA since the
beginning.
(This could be folded into f428c3)
Jun Wu <quark@fb.com> [Wed, 27 Sep 2017 15:14:59 -0700] rev 34321
progress: make ETA only consider progress made in the last minute
This patch limits the estimate time interval to roughly the last minute
(configurable by `estimateinterval`) to be more practical. See the test
change for why this is better.
.. feature:: Estimated time is more accurate with non-linear progress
Differential Revision: https://phab.mercurial-scm.org/D820
Jun Wu <quark@fb.com> [Wed, 27 Sep 2017 14:30:58 -0700] rev 34320
progress: remove progress.estimate config
It was introduced by 98e4d39 ("progress: add speed format" 2011-5-9) and was
intended to hide ETA information for the first few seconds.
Later 5d261fd ("progress: add a changedelay to prevent parallel topics from
flapping (issue2698)" 2011-6-23) introduced `changedelay` config which hides
the entire progress bar for the first few seconds. So `progress.estimate` seems
somehow duplicated feature-wise. Since it's experimental and duplicated, let's
just remove it. This makes the next patch simpler - it no longer needs to make
sure `starttimes` is the real start time.
Differential Revision: https://phab.mercurial-scm.org/D828
Jun Wu <quark@fb.com> [Tue, 26 Sep 2017 12:48:15 -0700] rev 34319
progress: demonstrate non-linear progress has a bad ETA experience
Previously, the ETA and speed assumes the progress is linear. Often, due to
network or other issues, it could be fast for the most time, and suddenly
slow down:
[====================================================> ]
\___________________________________________/\______/
very fast suddenly much slower
This patch adds a test demonstrating the ETA could be way off in those
cases.
Differential Revision: https://phab.mercurial-scm.org/D819
Pulkit Goyal <7895pulkit@gmail.com> [Thu, 21 Sep 2017 15:58:44 +0530] rev 34318
copytrace: add a a new config to limit the number of drafts in heuristics
The heuristics options tries to the default full copytracing algorithm if both
the source and destination branches contains of non-public changesets only. But
this can be slow in cases when we have a lot of drafts.
This patch adds a new config option experimental.copytrace.sourcecommitlimit
which defaults to 100. This value will be the limit of number of drafts from c1
to base. Incase there are more changesets even though they are draft, the
heuristics algorithm will be used.
Differential Revision: https://phab.mercurial-scm.org/D763
Igor Ippolitov <iippolitov@gmail.com> [Tue, 26 Sep 2017 16:14:57 +0300] rev 34317
mail: encode long unicode lines in emails properly (issue5687)
3e544c074459 introduced a bug: emails Content-Transfer-Encoding
is silently replaced with 'quoted-printable' while any other
encoding could be used by underlying code. The problem is revealed
when a long unicode line is encoded.
The patch implements proper check which works for any text and
encoding.
Jun Wu <quark@fb.com> [Sat, 23 Sep 2017 14:58:40 -0700] rev 34316
chg: show timestamp with debug messages
Like `strace -tr`, this helps finding performance bottlenecks.
Differential Revision: https://phab.mercurial-scm.org/D807