commit: tell user what to do with .hg/last-message.txt
I have always assumed that the message will be reused by the next `hg
commit`, but it seems it's just silently dropped on the next
commit. Let's try to be more helpful by telling the user that they
have to manually tell hg to reuse it.
The file will still be lost if the user runs some other operation in
between (like a non-in-memory rebase). That will be fixed once we've
switched all operations to be in-memory :)
I didn't include `$(hg root)/` in the path in the message to the user
because that would have made the message too long. Hopefully the user
will figure that part out themselves.
Differential Revision: https://phab.mercurial-scm.org/D8463
test-check-rust-format: specify --edition=2018
rustfmt doesn't read Cargo.toml unless it's executed by cargo.
https://github.com/rust-lang/rustfmt#rusts-editions
Added signature for changeset
26ce8e751503
Added tag 5.4rc0 for changeset
26ce8e751503
templatekw: fix shownames() to check if namespace exists in repo (
issue6301)
Namespace registration is dynamic, but the corresponding template keyword
is registered statically.
wait-on-file: use proper variable in math
This seems better and safer to be explicit.
Differential Revision: https://phab.mercurial-scm.org/D8426
wait-on-file: don't quote arithmetic argument
This is unnecessary and Mac OS X choke on them.
Differential Revision: https://phab.mercurial-scm.org/D8425
graft: exit 1 on conflicts, like merge
It's more consistent, and makes it nicer to script around hg if you
don't have to ignore exit code 255, which is the error code for
basically everything in hg.
Differential Revision: https://phab.mercurial-scm.org/D8423
tests: deal with "ls" vs "ls -A" difference on 2BSD derived systems
BSD-derived systems will use "ls -A" when running as root. Adjust the
test cases to always use the flag and include .hg and related output as
appropiately.
Differential Revision: https://phab.mercurial-scm.org/D8397