* Suggested `watch --color` as recommended by @ilyagr
* Answered monitoring `jj log` with watch and TUIs
* Minor wording fix.
* Adjusted watch(1) and hwatch links as suggested by @ilyagr
Signed-off-by: Tim Janik <timj@gnu.org>
I think most people would benefit from trying both an internal and an
external merge tool.
I was also considering mentioning `kdiff3` or `diffedit3` more heavily.
For `kdiff3`, I find its diff editing interface very confusing and I don't know of anybody who happily uses it as a diff
editor. For `diffedit3`, I'm not very objective.
I don't think recommending KDIff3 for diff editing is good advice in
most cases (as opposed to merge conflict resolution, for which KDiff3 is
very good). OTOH, trying Meld for diff editing is a good idea for many
people.
Historically, this part of the docs uses KDiff3 because it was written
when `meld` was the default and `:builtin` did not exist.
We only explained that working copies can become stale when updating a
commit from another workspace, but there are other cases too. I also
tried to better explain what the state actually means.
We had documented that we support `git.auto-local-bookmark` but we
don't. The documentation has been incorrect since d9c68e08b1a5. This
patch fixes it by adding support for `git.auto-local-bookmark` with
fallback to the old/current `git.auto-local-branch`.
.
This part of the git-comparison table was implying you need to manually
check and clear staged changes in order to only update the previous
commit's message. That isn't the case, though. As mentioned in the
`git-commit` documentation [1], this can be directly achieved by using
the `--only` flag with `--amend`:
> If this option is specified together with --amend, then no paths need
> to be specified, which can be used to amend the last commit without
> committing changes that have already been staged.
[1]: https://git-scm.com/docs/git-commit#Documentation/git-commit.txt--o
The current instructions ask the reader to convert the merge commit into
a single-parent commit, then create a new merge.
It is probably simpler to follow instructions for inserting a new commit
before the merge commit.
This was added at f5f61f6bfe36 "revset: resolve 'HEAD@git' just like other
pseudo @git branches." As I said in this patch, there was no practical use case
of the HEAD@git symbol.
Suppose we implement colocated workspaces/worktrees #4436, there may be multiple
Git HEAD revisions. This means HEAD can no longer be abstracted as a symbol of
the "git" remote.
As I said in the preceding patch, I settled on separate pad/truncate functions
instead of a function taking multiple optional parameters. It's less efficient
to process truncation and padding independently, but I don't think that would
matter.
The order of arguments follows the current f(..., content) convention. We can
also add a method syntax, but I'm not sure if it's useful. In order to call a
method of Template type, we'll need to coerce printable object to Template:
concat(author.email()).truncate_end(10).pad_end(10)
^^^^^^
String -> Template
FWIW, String type could provide more efficient truncate/pad methods.
Closes#3183
We have our MSRV recorded in `Cargo.toml`, but at least with Rust
versions < 1.71, the build fails before the MSRV is checked (IIUC),
and you get a cryptic message like this:
```
error: Package jj-lib v0.22.0 (<some path>) does not have feature git2.
```
The message is clearer from 1.71, but that doesn't help users who have
an older version. Let's mention the MSRV everywhere we tell the user
to run `cargo` in the installation instructions.
Users should generally use `builtin_immutable_heads()` when they
override `immutable_heads()`, so this patch updates the existing
example to use that. I also added another example showing how to
make an additional remote-tracking bookmark immutable.
This allows for more fine-grained control of timestamp formatting, for
example:
```
[template-aliases]
'format_timestamp(timestamp)' = '''
if(timestamp.before("1 week ago"),
timestamp.format("%b %d %Y %H:%M"),
timestamp.ago()
)
'''
```
Closes#3782.
The `coalesce` function takes a list of revsets and returns the commits in the
first revset in the list which evalutes to a non-empty set of commits.
It can be used to display fallbacks if a certain commit cannot be found,
e.g. `coalesce(present(user_configured_trunk), builtin_trunk)`.
This adds `raw_escape_sequence(...)` support for things that use
FormatRecorder like wrapped text / `fill(...)` / `indent(...)`.
Change-Id: Id00000004248b10feb2acd54d90115b783fac0ff
Templates can be formatted (using labels) and are usually sanitized
(unless for plain text output).
`raw_escape_sequence(content)` bypasses both.
```toml
'hyperlink(url, text)' = '''
raw_escape_sequence("\e]8;;" ++ url ++ "\e\\") ++
text ++
raw_escape_sequence("\e]8;;\e\\")
'''
```
In this example, `raw_escape_sequence` not only outputs the intended
escape codes, it also strips away any escape codes that might otherwise
be part of the `url` (from any labels attached to the `url` content).
Not all formatters (namely FormatRecorder) are supported yet.
Change-Id: Id00000004492dbf39e50f3b7090706839d1d8d45
One particular use case for these is escape sequences -- and to that
end, I'm also adding `\e` as a shorthand for `\x1b`.
Change-Id: Id000000040ea6fd8e2d720219931485960c570dd
This patch replaces all call sites with present(trunk()), and adds an explicit
check for unresolvable trunk(). If we add coalesce() expression, maybe it can
be rewritten to coalesce(present(trunk()), builtin_trunk()).
Fixes#4616
We didn't have any guidelines about what to include in a commit
message. Others have already written good guides for that. This commit
adds a link to one. I also added a sentence about explaining the
reason for a change, since I think that's particularly often missed
(I'm sure I also miss it sometimes - don't hesitate to point out when
that happens).
It's generally more useful to say something like "revset: use Self to
refer to expressions of the same type" than "lib: use Self to refer to
revset expressions of the same type ", so let's try to clarify that.
I forgot to add the `snapshot.auto-track` config option to `config.md`
when I added it. This patch copies it from `working-copy.md` and
modifies it slightly.
See discussion thread in linked issue.
With this PR, all revset functions in [BUILTIN_FUNCTION_MAP](8d166c7642/lib/src/revset.rs (L570))
that return multiple values are either named in plural or the naming is hard to misunderstand (e.g. `reachable`)
Fixes: #4122
