Martin von Zweigbergk <martinvonz@google.com> [Tue, 11 Sep 2018 11:03:15 -0700] rev 39545
narrow: check "narrow" wire protocol capability, not bundle2 capability
It seems like the new "narrow" wire protocol capability should be what
determines if the server supports the "narrow" and
"{,old}{in,ex}cludepats" arguments to the getbundle request.
Differential Revision: https://phab.mercurial-scm.org/D4527
Pulkit Goyal <pulkit@yandex-team.ru> [Tue, 21 Aug 2018 16:11:17 +0300] rev 39544
sparse: add local files to temporaryfiles if they exist out of sparse
We get the f1 from args if it's merge and check that whether that exists in
sparse checkout or not. If that does not, we add that for merging.
The error comes from very low-level where we try to read data of a
working-filectx which does not exists in the working directory. It will be
extremely ugly to plug in logic to update sparse copy with new file at such a
low level.
We already have logic related to updating the checkout with required files in
calculateupdates() and let's handle this case there only. calculateupdates()
call sparse.filterupdatesactions() and the logic is added into the latter
function.
To get the exact traceback, this patch can be backed out and
test-sparse-merges.t can be run with ui.traceback=True.
Differential Revision: https://phab.mercurial-scm.org/D4341
Pulkit Goyal <pulkit@yandex-team.ru> [Tue, 21 Aug 2018 15:09:22 +0300] rev 39543
tests: show that merging with sparse is broken when rename is involved
This patch adds test to show that merging with sparse is broken when you have a
rename on one side and just modification on another side.
Differential Revision: https://phab.mercurial-scm.org/D4340
Pulkit Goyal <pulkit@yandex-team.ru> [Sat, 25 Aug 2018 22:19:42 +0300] rev 39542
narrowspec: fix a typoed 'supported'
Spotted by martinvonz.
Differential Revision: https://phab.mercurial-scm.org/D4374
Pulkit Goyal <pulkit@yandex-team.ru> [Mon, 10 Sep 2018 16:29:22 +0300] rev 39541
narrow: build the known set of nodes only when ellipses is enabled
We don't need to build the known set in non-ellipses case because we don't have
a shallow repo. In this patch, this checks whether the server has ellipses
enabled or not using the server capability and then build the known set of
nodes.
Building the known set of nodes can take ~3-4 minutes on repositories with
millions of csets so this patch speeds up extending a non-shallow narrow clone
on large repositories.
In future, we should first check whether local repository is an ellipses repo
using a new ellipses repo requirement and then control all the combinations
between local repo requirement and server capability.
Differential Revision: https://phab.mercurial-scm.org/D4520
Pulkit Goyal <pulkit@yandex-team.ru> [Mon, 10 Sep 2018 15:55:14 +0300] rev 39540
narrow: add narrow and ellipses as server capabilities
Right now we don't have a way to differentiate between whether a server can
serve ellipsis or not. The way we check whether a server is narrow enabled is by
checking bundle2 capability which does not scale outside of bundle2 world.
The goal is to use have wireprotocol commands just like remotefilelog for
widening the narrow clone, atleast in non-ellipses cases. Having a server
capability will help there as we can't rely on bundle2 capability there.
Also having a server capability is neat than having a bundle2 capability.
There are lot of things we can optimize locally on the client side by knowing
before hand that whether the server supports ellipses or not.
This will also help us in making sure that a client ellipses repo does not
communicate with a server repo without ellipses.
Differential Revision: https://phab.mercurial-scm.org/D4521