forgedb 0.3.0

ForgeDB — an application database generator. Compiles a declarative .forge schema into tailored Rust database code, a TypeScript SDK, and a REST API.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108

# git-cliff — CHANGELOG.md generator for the ForgeDB *core* release line (#211).
#
# The changelog is the backward-looking, mechanical record of what actually
# shipped in each `v*` release, generated from the repo's conventional commits
# (feat:/fix:/perf:/…). It is regenerated at release time (`make changelog`),
# committed, and consumed by TWO surfaces from one source:
#   - cargo-dist builds each GitHub Release body from this file (release.yml).
#   - the website renders it statically at /changelog (apps/website).
#
# Scope discipline: this is the CORE (`v[0-9]*`) changelog. The VS Code extension
# ships on its own `vscode-v*` tag line, and the marketing site / packaging
# channels are not product releases — commits scoped to them are filtered out
# below so they never headline a ForgeDB release.

[changelog]
# A stable preamble kept verbatim at the top of CHANGELOG.md.
header = """
# Changelog

All notable changes to the ForgeDB core (the `forgedb` CLI and its published
substrate crates) are documented here. This file is generated from conventional
commits with [git-cliff](https://git-cliff.org); do not edit it by hand — run
`make changelog`. The format follows [Keep a Changelog](https://keepachangelog.com),
and the project honors [Semantic Versioning](https://semver.org) per `docs/SEMVER.md`.

"""

# One section per release. Keep-a-Changelog headings (`## [x.y.z] - date`) so
# cargo-dist can locate the matching section and the website parser can split on
# them. Issue refs are linked by a preprocessor below.
body = """
{% if version %}\
## [{{ version | trim_start_matches(pat="v") }}] - {{ timestamp | date(format="%Y-%m-%d") }}
{% else %}\
## [Unreleased]
{% endif %}\
{% for group, commits in commits | group_by(attribute="group") %}
### {{ group | striptags | trim | upper_first }}
{% for commit in commits %}
- {% if commit.scope %}**{{ commit.scope }}:** {% endif %}\
{% if commit.breaking %}**breaking:** {% endif %}\
{{ commit.message | upper_first }}\
{% endfor %}
{% endfor %}
"""

footer = """
"""
trim = true

[git]
conventional_commits = true
# Drop commits that don't follow the conventional format — keeps the log clean
# rather than surfacing every stray subject.
filter_unconventional = true
split_commits = false

# Rewrite the raw subject before parsing/rendering.
commit_preprocessors = [
  # Link issue/PR refs so both the GitHub Release body and the website get
  # clickable references. Single-quoted TOML literal → the regex sees `\(`.
  { pattern = '\(#([0-9]+)\)', replace = "([#${1}](https://github.com/hoodiecollin/forgedb/issues/${1}))" },
]

# First match wins — order matters. Scope-based skips come first so the marketing
# site / extension / packaging channels never appear in the core changelog.
#
# Group names carry an `<!-- N -->` sort-key prefix so `group_by` orders sections
# logically (Features → Fixes → Performance) instead of alphabetically; the
# template strips the comment with `striptags`.
#
# Only the user-facing axes are kept — features, fixes, performance. Docs,
# refactors, chores, CI, and unconventional subjects are dropped as changelog
# noise (`filter_unconventional = true` discards anything with no matching
# parser). Breaking changes are the exception: `protect_breaking_commits = true`
# keeps them even when their type is skipped, and they are routed to their own
# top section here.
commit_parsers = [
  { message = "^chore\\(release\\)", skip = true },
  { message = "\\((website|vscode|npm|pypi|winget|docker|brew|homebrew)\\)", skip = true },
  # Any `type!:` / `type(scope)!:` breaking marker headlines its own section,
  # regardless of the underlying type.
  { message = "^[a-z]+(\\(.+\\))?!:", group = "<!-- 0 -->Breaking changes" },
  { message = "^feat", group = "<!-- 1 -->Features" },
  { message = "^fix", group = "<!-- 2 -->Bug Fixes" },
  { message = "^perf", group = "<!-- 3 -->Performance" },
  # Everything else (docs, refactors, chores, ci, tests, bench, merge, reverts,
  # and any other type) is changelog noise — skip it explicitly so a stray type
  # never defaults into a section of its own.
  { message = ".*", skip = true },
]

# Breaking changes are surfaced by the explicit `!:`-marker parser above (routed
# to their own section), so the auto-resurrection of skipped breaking commits is
# left off — otherwise a skipped `refactor!`/`bench!` reappears under a stray
# type-named section.
protect_breaking_commits = false
filter_commits = false

# Only the CORE release line counts as a release here. Anchored so `vscode-v*`
# (extension line) and any other prefix are excluded — this is the single point
# that keeps the two release lines from colliding (see docs/roadmap notes).
tag_pattern = "^v[0-9]+\\.[0-9]+\\.[0-9]+"
# Prereleases (`-rc`, `-beta`, …) are not public releases; keep them off the page.
ignore_tags = "-(rc|alpha|beta)"

topo_order = false
sort_commits = "oldest"