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
|
// Copyright 2019 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#ifndef COMPONENTS_SERVICES_APP_SERVICE_PUBLIC_CPP_ICON_COALESCER_H_
#define COMPONENTS_SERVICES_APP_SERVICE_PUBLIC_CPP_ICON_COALESCER_H_
#include <map>
#include <memory>
#include <set>
#include <string>
#include <utility>
#include "base/memory/raw_ptr.h"
#include "base/memory/ref_counted.h"
#include "base/memory/scoped_refptr.h"
#include "base/memory/weak_ptr.h"
#include "base/sequence_checker.h"
#include "components/services/app_service/public/cpp/icon_loader.h"
namespace apps {
// An IconLoader that coalesces the apps::mojom::IconType::kUncompressed
// results of another (wrapped) IconLoader.
//
// This is similar to, but different from, an IconCache. Both types are related
// to the LoadIconFromIconKey Mojo call (the request and response), both reduce
// the number of requests made, and both re-use the response for requests with
// the same IconLoader::Key.
//
// An IconCache (another class) applies when the second request is sent *after*
// the first response is received. An IconCoalescer (this class) applies when
// the second request is sent *before* the first response is received (but
// after the first request is sent, obviously).
//
// Caching means that the second (and subsequent) requests can be satisfied
// immediately, sharing the previous response. Coalescing means that the second
// (and subsequent) requests are paused, and when the first request's response
// is finally received, those other requests are un-paused and share the same
// response.
//
// When there are no in-flight requests, a (memory-backed) cache can still have
// a significant memory cost, depending on how aggressive its cache eviction
// policy is, but a (memory-backed) coalescer will have a trivial memory cost.
// Much of its internal state (e.g. maps and multimaps) will be empty.
class IconCoalescer : public IconLoader {
public:
explicit IconCoalescer(IconLoader* wrapped_loader);
IconCoalescer(const IconCoalescer&) = delete;
IconCoalescer& operator=(const IconCoalescer&) = delete;
~IconCoalescer() override;
// IconLoader overrides.
std::optional<IconKey> GetIconKey(const std::string& id) override;
std::unique_ptr<Releaser> LoadIconFromIconKey(
const std::string& id,
const IconKey& icon_key,
IconType icon_type,
int32_t size_hint_in_dip,
bool allow_placeholder_icon,
apps::LoadIconCallback callback) override;
private:
class RefCountedReleaser;
using CallbackAndReleaser =
std::pair<apps::LoadIconCallback, scoped_refptr<RefCountedReleaser>>;
void OnLoadIcon(IconLoader::Key key,
uint64_t sequence_number,
IconValuePtr icon_value);
raw_ptr<IconLoader> wrapped_loader_;
// Every incoming LoadIconFromIconKey call gets its own sequence number.
uint64_t next_sequence_number_;
// Sequence numbers for outstanding requests to to the wrapped_loader_'s
// LoadIconFromIconKey. When the wrapped_loader_ returns, there will either
// be a matching entry in immediate_responses_ or an entry will be made in
// non_immediate_requests_, depending on whether LoadIconFromIconKey resolved
// (ran its callback) synchronously (immediately) or asynchronously
// (non-immediately).
std::set<uint64_t> possibly_immediate_requests_;
// Map from sequence number to the IconValue to give to the LoadIconCallback.
std::map<uint64_t, IconValuePtr> immediate_responses_;
// Multimap of pending LoadIconFromIconKey calls: those calls that were not
// resolved immediately.
std::multimap<IconLoader::Key, CallbackAndReleaser> non_immediate_requests_;
SEQUENCE_CHECKER(sequence_checker_);
base::WeakPtrFactory<IconCoalescer> weak_ptr_factory_{this};
};
} // namespace apps
#endif // COMPONENTS_SERVICES_APP_SERVICE_PUBLIC_CPP_ICON_COALESCER_H_
|
