代码拉取完成,页面将自动刷新
// Copyright 2021-present The Atlas Authors. All rights reserved.
// This source code is licensed under the Apache 2.0 license found
// in the LICENSE file in the root directory of this source tree.
package mysql
import (
"context"
"encoding/binary"
"fmt"
"regexp"
"sort"
"strconv"
"strings"
"gitee.com/damengde/atlas/sql/internal/sqlx"
"gitee.com/damengde/atlas/sql/migrate"
"gitee.com/damengde/atlas/sql/schema"
)
type (
// tplanApply decorates MySQL planApply.
tplanApply struct{ planApply }
// tdiff decorates MySQL diff.
tdiff struct{ diff }
// tinspect decorates MySQL inspect.
tinspect struct{ inspect }
)
// priority computes the priority of each change.
//
// TiDB does not support multischema ALTERs (i.e. multiple changes in a single ALTER statement).
// Therefore, we have to break down each alter. This function helps order the ALTERs so they work.
// e.g. priority gives precedence to DropForeignKey over DropColumn, because a column cannot be
// dropped if its foreign key was not dropped before.
func priority(change schema.Change) int {
switch c := change.(type) {
case *schema.ModifyTable:
// each modifyTable should have a single change since we apply `flat` before we sort.
return priority(c.Changes[0])
case *schema.ModifySchema:
// each modifyTable should have a single change since we apply `flat` before we sort.
return priority(c.Changes[0])
case *schema.AddColumn:
return 1
case *schema.DropIndex, *schema.DropForeignKey, *schema.DropAttr, *schema.DropCheck:
return 2
case *schema.ModifyIndex, *schema.ModifyForeignKey:
return 3
default:
return 4
}
}
// flat takes a list of changes and breaks them down to single atomic changes (e.g: no ModifyTable
// with multiple AddColumn inside it). Note that, the only "changes" that include sub-changes are
// `ModifyTable` and `ModifySchema`.
func flat(changes []schema.Change) []schema.Change {
var flat []schema.Change
for _, change := range changes {
switch m := change.(type) {
case *schema.ModifyTable:
for _, c := range m.Changes {
flat = append(flat, &schema.ModifyTable{
T: m.T,
Changes: []schema.Change{c},
})
}
case *schema.ModifySchema:
for _, c := range m.Changes {
flat = append(flat, &schema.ModifySchema{
S: m.S,
Changes: []schema.Change{c},
})
}
default:
flat = append(flat, change)
}
}
return flat
}
// PlanChanges returns a migration plan for the given schema changes.
func (p *tplanApply) PlanChanges(ctx context.Context, name string, changes []schema.Change, opts ...migrate.PlanOption) (*migrate.Plan, error) {
planned, err := sqlx.DetachCycles(changes)
if err != nil {
return nil, err
}
planned = flat(planned)
sort.SliceStable(planned, func(i, j int) bool {
return priority(planned[i]) < priority(planned[j])
})
s := &state{
conn: p.conn,
Plan: migrate.Plan{
Name: name,
// A plan is reversible, if all
// its changes are reversible.
Reversible: true,
Transactional: false,
},
}
for _, c := range planned {
// Use the planner of MySQL with each "atomic" change.
plan, err := p.planApply.PlanChanges(ctx, name, []schema.Change{c}, opts...)
if err != nil {
return nil, err
}
if !plan.Reversible {
s.Plan.Reversible = false
}
s.Plan.Changes = append(s.Plan.Changes, plan.Changes...)
}
return &s.Plan, nil
}
func (p *tplanApply) ApplyChanges(ctx context.Context, changes []schema.Change, opts ...migrate.PlanOption) error {
return sqlx.ApplyChanges(ctx, changes, p, opts...)
}
func (i *tinspect) InspectSchema(ctx context.Context, name string, opts *schema.InspectOptions) (*schema.Schema, error) {
s, err := i.inspect.InspectSchema(ctx, name, opts)
if err != nil {
return nil, err
}
return i.patchSchema(ctx, s)
}
func (i *tinspect) InspectRealm(ctx context.Context, opts *schema.InspectRealmOption) (*schema.Realm, error) {
r, err := i.inspect.InspectRealm(ctx, opts)
if err != nil {
return nil, err
}
for _, s := range r.Schemas {
if _, err := i.patchSchema(ctx, s); err != nil {
return nil, err
}
}
return r, nil
}
func (i *tinspect) patchSchema(ctx context.Context, s *schema.Schema) (*schema.Schema, error) {
for _, t := range s.Tables {
var createStmt CreateStmt
if ok := sqlx.Has(t.Attrs, &createStmt); !ok {
if _, err := i.createStmt(ctx, t); err != nil {
return nil, err
}
}
if err := i.setCollate(t); err != nil {
return nil, err
}
if err := i.setAutoIncrement(t); err != nil {
return nil, err
}
for _, c := range t.Columns {
i.patchColumn(ctx, c)
}
}
return s, nil
}
func (i *tinspect) patchColumn(_ context.Context, c *schema.Column) {
_, ok := c.Type.Type.(*BitType)
if !ok {
return
}
// TiDB has a bug where it does not format bit default value correctly.
if lit, ok := c.Default.(*schema.Literal); ok && !strings.HasPrefix(lit.V, "b'") {
lit.V = bytesToBitLiteral([]byte(lit.V))
}
}
// bytesToBitLiteral converts a bytes to MySQL bit literal.
// e.g. []byte{4} -> b'100', []byte{2,1} -> b'1000000001'.
// See: https://github.com/pingcap/tidb/issues/32655.
func bytesToBitLiteral(b []byte) string {
bytes := make([]byte, 8)
for i := 0; i < len(b); i++ {
bytes[8-len(b)+i] = b[i]
}
val := binary.BigEndian.Uint64(bytes)
return fmt.Sprintf("b'%b'", val)
}
// e.g CHARSET=utf8mb4 COLLATE=utf8mb4_bin
var reColl = regexp.MustCompile(`(?i)CHARSET\s*=\s*(\w+)\s*COLLATE\s*=\s*(\w+)`)
// setCollate extracts the updated Collation from CREATE TABLE statement.
func (i *tinspect) setCollate(t *schema.Table) error {
var c CreateStmt
if !sqlx.Has(t.Attrs, &c) {
return fmt.Errorf("missing CREATE TABLE statement in attributes for %q", t.Name)
}
matches := reColl.FindStringSubmatch(c.S)
if len(matches) != 3 {
return fmt.Errorf("missing COLLATE and/or CHARSET information on CREATE TABLE statement for %q", t.Name)
}
t.SetCharset(matches[1])
t.SetCollation(matches[2])
return nil
}
// setCollate extracts the updated Collation from CREATE TABLE statement.
func (i *tinspect) setAutoIncrement(t *schema.Table) error {
// patch only it is set (set falsely to '1' due to this bug:https://github.com/pingcap/tidb/issues/24702).
ai := &AutoIncrement{}
if !sqlx.Has(t.Attrs, ai) {
return nil
}
var c CreateStmt
if !sqlx.Has(t.Attrs, &c) {
return fmt.Errorf("missing CREATE TABLE statement in attributes for %q", t.Name)
}
matches := reAutoinc.FindStringSubmatch(c.S)
if len(matches) != 2 {
return nil
}
v, err := strconv.ParseInt(matches[1], 10, 64)
if err != nil {
return err
}
ai.V = v
schema.ReplaceOrAppend(&t.Attrs, ai)
return nil
}
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。