@mcp-abap-adt/interfaces-adt-connection 1.0.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +19 -0
- package/COPYING +674 -0
- package/LICENSE +165 -0
- package/README.md +46 -0
- package/dist/IAbapConnection.d.ts +65 -0
- package/dist/IAbapConnection.d.ts.map +1 -0
- package/dist/IAbapConnection.js +2 -0
- package/dist/IAbapRequestOptions.d.ts +9 -0
- package/dist/IAbapRequestOptions.d.ts.map +1 -0
- package/dist/IAbapRequestOptions.js +2 -0
- package/dist/IConnectionCapabilities.d.ts +217 -0
- package/dist/IConnectionCapabilities.d.ts.map +1 -0
- package/dist/IConnectionCapabilities.js +29 -0
- package/dist/ITimeoutConfig.d.ts +15 -0
- package/dist/ITimeoutConfig.d.ts.map +1 -0
- package/dist/ITimeoutConfig.js +2 -0
- package/dist/index.d.ts +18 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +5 -0
- package/package.json +41 -0
package/LICENSE
ADDED
|
@@ -0,0 +1,165 @@
|
|
|
1
|
+
GNU LESSER GENERAL PUBLIC LICENSE
|
|
2
|
+
Version 3, 29 June 2007
|
|
3
|
+
|
|
4
|
+
Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
|
|
5
|
+
Everyone is permitted to copy and distribute verbatim copies
|
|
6
|
+
of this license document, but changing it is not allowed.
|
|
7
|
+
|
|
8
|
+
|
|
9
|
+
This version of the GNU Lesser General Public License incorporates
|
|
10
|
+
the terms and conditions of version 3 of the GNU General Public
|
|
11
|
+
License, supplemented by the additional permissions listed below.
|
|
12
|
+
|
|
13
|
+
0. Additional Definitions.
|
|
14
|
+
|
|
15
|
+
As used herein, "this License" refers to version 3 of the GNU Lesser
|
|
16
|
+
General Public License, and the "GNU GPL" refers to version 3 of the GNU
|
|
17
|
+
General Public License.
|
|
18
|
+
|
|
19
|
+
"The Library" refers to a covered work governed by this License,
|
|
20
|
+
other than an Application or a Combined Work as defined below.
|
|
21
|
+
|
|
22
|
+
An "Application" is any work that makes use of an interface provided
|
|
23
|
+
by the Library, but which is not otherwise based on the Library.
|
|
24
|
+
Defining a subclass of a class defined by the Library is deemed a mode
|
|
25
|
+
of using an interface provided by the Library.
|
|
26
|
+
|
|
27
|
+
A "Combined Work" is a work produced by combining or linking an
|
|
28
|
+
Application with the Library. The particular version of the Library
|
|
29
|
+
with which the Combined Work was made is also called the "Linked
|
|
30
|
+
Version".
|
|
31
|
+
|
|
32
|
+
The "Minimal Corresponding Source" for a Combined Work means the
|
|
33
|
+
Corresponding Source for the Combined Work, excluding any source code
|
|
34
|
+
for portions of the Combined Work that, considered in isolation, are
|
|
35
|
+
based on the Application, and not on the Linked Version.
|
|
36
|
+
|
|
37
|
+
The "Corresponding Application Code" for a Combined Work means the
|
|
38
|
+
object code and/or source code for the Application, including any data
|
|
39
|
+
and utility programs needed for reproducing the Combined Work from the
|
|
40
|
+
Application, but excluding the System Libraries of the Combined Work.
|
|
41
|
+
|
|
42
|
+
1. Exception to Section 3 of the GNU GPL.
|
|
43
|
+
|
|
44
|
+
You may convey a covered work under sections 3 and 4 of this License
|
|
45
|
+
without being bound by section 3 of the GNU GPL.
|
|
46
|
+
|
|
47
|
+
2. Conveying Modified Versions.
|
|
48
|
+
|
|
49
|
+
If you modify a copy of the Library, and, in your modifications, a
|
|
50
|
+
facility refers to a function or data to be supplied by an Application
|
|
51
|
+
that uses the facility (other than as an argument passed when the
|
|
52
|
+
facility is invoked), then you may convey a copy of the modified
|
|
53
|
+
version:
|
|
54
|
+
|
|
55
|
+
a) under this License, provided that you make a good faith effort to
|
|
56
|
+
ensure that, in the event an Application does not supply the
|
|
57
|
+
function or data, the facility still operates, and performs
|
|
58
|
+
whatever part of its purpose remains meaningful, or
|
|
59
|
+
|
|
60
|
+
b) under the GNU GPL, with none of the additional permissions of
|
|
61
|
+
this License applicable to that copy.
|
|
62
|
+
|
|
63
|
+
3. Object Code Incorporating Material from Library Header Files.
|
|
64
|
+
|
|
65
|
+
The object code form of an Application may incorporate material from
|
|
66
|
+
a header file that is part of the Library. You may convey such object
|
|
67
|
+
code under terms of your choice, provided that, if the incorporated
|
|
68
|
+
material is not limited to numerical parameters, data structure
|
|
69
|
+
layouts and accessors, or small macros, inline functions and templates
|
|
70
|
+
(ten or fewer lines in length), you do both of the following:
|
|
71
|
+
|
|
72
|
+
a) Give prominent notice with each copy of the object code that the
|
|
73
|
+
Library is used in it and that the Library and its use are
|
|
74
|
+
covered by this License.
|
|
75
|
+
|
|
76
|
+
b) Accompany the object code with a copy of the GNU GPL and this license
|
|
77
|
+
document.
|
|
78
|
+
|
|
79
|
+
4. Combined Works.
|
|
80
|
+
|
|
81
|
+
You may convey a Combined Work under terms of your choice that,
|
|
82
|
+
taken together, effectively do not restrict modification of the
|
|
83
|
+
portions of the Library contained in the Combined Work and reverse
|
|
84
|
+
engineering for debugging such modifications, if you also do each of
|
|
85
|
+
the following:
|
|
86
|
+
|
|
87
|
+
a) Give prominent notice with each copy of the Combined Work that
|
|
88
|
+
the Library is used in it and that the Library and its use are
|
|
89
|
+
covered by this License.
|
|
90
|
+
|
|
91
|
+
b) Accompany the Combined Work with a copy of the GNU GPL and this license
|
|
92
|
+
document.
|
|
93
|
+
|
|
94
|
+
c) For a Combined Work that displays copyright notices during
|
|
95
|
+
execution, include the copyright notice for the Library among
|
|
96
|
+
these notices, as well as a reference directing the user to the
|
|
97
|
+
copies of the GNU GPL and this license document.
|
|
98
|
+
|
|
99
|
+
d) Do one of the following:
|
|
100
|
+
|
|
101
|
+
0) Convey the Minimal Corresponding Source under the terms of this
|
|
102
|
+
License, and the Corresponding Application Code in a form
|
|
103
|
+
suitable for, and under terms that permit, the user to
|
|
104
|
+
recombine or relink the Application with a modified version of
|
|
105
|
+
the Linked Version to produce a modified Combined Work, in the
|
|
106
|
+
manner specified by section 6 of the GNU GPL for conveying
|
|
107
|
+
Corresponding Source.
|
|
108
|
+
|
|
109
|
+
1) Use a suitable shared library mechanism for linking with the
|
|
110
|
+
Library. A suitable mechanism is one that (a) uses at run time
|
|
111
|
+
a copy of the Library already present on the user's computer
|
|
112
|
+
system, and (b) will operate properly with a modified version
|
|
113
|
+
of the Library that is interface-compatible with the Linked
|
|
114
|
+
Version.
|
|
115
|
+
|
|
116
|
+
e) Provide Installation Information, but only if you would otherwise
|
|
117
|
+
be required to provide such information under section 6 of the
|
|
118
|
+
GNU GPL, and only to the extent that such information is
|
|
119
|
+
necessary to install and execute a modified version of the
|
|
120
|
+
Combined Work produced by recombining or relinking the
|
|
121
|
+
Application with a modified version of the Linked Version. (If
|
|
122
|
+
you use option 4d0, the Installation Information must accompany
|
|
123
|
+
the Minimal Corresponding Source and Corresponding Application
|
|
124
|
+
Code. If you use option 4d1, you must provide the Installation
|
|
125
|
+
Information in the manner specified by section 6 of the GNU GPL
|
|
126
|
+
for conveying Corresponding Source.)
|
|
127
|
+
|
|
128
|
+
5. Combined Libraries.
|
|
129
|
+
|
|
130
|
+
You may place library facilities that are a work based on the
|
|
131
|
+
Library side by side in a single library together with other library
|
|
132
|
+
facilities that are not Applications and are not covered by this
|
|
133
|
+
License, and convey such a combined library under terms of your
|
|
134
|
+
choice, if you do both of the following:
|
|
135
|
+
|
|
136
|
+
a) Accompany the combined library with a copy of the same work based
|
|
137
|
+
on the Library, uncombined with any other library facilities,
|
|
138
|
+
conveyed under the terms of this License.
|
|
139
|
+
|
|
140
|
+
b) Give prominent notice with the combined library that part of it
|
|
141
|
+
is a work based on the Library, and explaining where to find the
|
|
142
|
+
accompanying uncombined form of the same work.
|
|
143
|
+
|
|
144
|
+
6. Revised Versions of the GNU Lesser General Public License.
|
|
145
|
+
|
|
146
|
+
The Free Software Foundation may publish revised and/or new versions
|
|
147
|
+
of the GNU Lesser General Public License from time to time. Such new
|
|
148
|
+
versions will be similar in spirit to the present version, but may
|
|
149
|
+
differ in detail to address new problems or concerns.
|
|
150
|
+
|
|
151
|
+
Each version is given a distinguishing version number. If the
|
|
152
|
+
Library as you received it specifies that a certain numbered version
|
|
153
|
+
of the GNU Lesser General Public License "or any later version"
|
|
154
|
+
applies to it, you have the option of following the terms and
|
|
155
|
+
conditions either of that published version or of any later version
|
|
156
|
+
published by the Free Software Foundation. If the Library as you
|
|
157
|
+
received it does not specify a version number of the GNU Lesser
|
|
158
|
+
General Public License, you may choose any version of the GNU Lesser
|
|
159
|
+
General Public License ever published by the Free Software Foundation.
|
|
160
|
+
|
|
161
|
+
If the Library as you received it specifies that a proxy can decide
|
|
162
|
+
whether future versions of the GNU Lesser General Public License shall
|
|
163
|
+
apply, that proxy's public statement of acceptance of any version is
|
|
164
|
+
permanent authorization for you to choose that version for the
|
|
165
|
+
Library.
|
package/README.md
ADDED
|
@@ -0,0 +1,46 @@
|
|
|
1
|
+
# @mcp-abap-adt/interfaces-adt-connection
|
|
2
|
+
|
|
3
|
+
The ABAP connection contract: what an ADT request is sent through, and what
|
|
4
|
+
comes back off the wire. Types and one constant; no implementation.
|
|
5
|
+
|
|
6
|
+
```bash
|
|
7
|
+
npm install @mcp-abap-adt/interfaces-adt-connection
|
|
8
|
+
```
|
|
9
|
+
|
|
10
|
+
| symbol | what it is |
|
|
11
|
+
|---|---|
|
|
12
|
+
| `IAbapConnection` | a connection an ADT request is made through: `makeAdtRequest`, connect/disconnect, the session type |
|
|
13
|
+
| `IAbapRequestOptions` | what one request carries |
|
|
14
|
+
| `IAdtWireResponse` | the answer as it came off the wire — extends `IHttpWireResponse` from `interfaces-network` with the headers ADT sends |
|
|
15
|
+
| `ISessionLifecycleAware`, `ICriticalSection`, `IRequestProfiling`, `IDeferredResponseConnection` | the connection's capability atoms — a connection honours the ones it can |
|
|
16
|
+
| `ADT_SESSION_ERROR` / `AdtSessionErrorCode` | what a session failure is called |
|
|
17
|
+
| `ITimeoutConfig` | the deadlines a caller gives a connection: `default`, `csrf`, `long` |
|
|
18
|
+
|
|
19
|
+
## Why this package exists
|
|
20
|
+
|
|
21
|
+
**It moves on its own schedule.** These four files were
|
|
22
|
+
`interfaces-adt/src/connection/` until `interfaces-adt` 11.0.0. The object
|
|
23
|
+
contracts beside them had three majors in one day — 9, 10 and 11 — and none
|
|
24
|
+
touched the connection; yet every one of them made the connector
|
|
25
|
+
(`@mcp-abap-adt/connection`) either release to follow or install a second copy
|
|
26
|
+
of the contract beside `adt-clients`. Two copies of one interface do not compare
|
|
27
|
+
equal in TypeScript, so that shows up as errors that read as impossible.
|
|
28
|
+
|
|
29
|
+
The subject is still ADT — `makeAdtRequest`, the stateful session, CSRF — so it
|
|
30
|
+
does not belong in `interfaces-network`, whose rule is that nothing there is
|
|
31
|
+
ADT-specific (decision 35). What separated it is the release rate (decision 38
|
|
32
|
+
in the repository's `docs/architecture/DECISIONS.md`).
|
|
33
|
+
|
|
34
|
+
## Who depends on it
|
|
35
|
+
|
|
36
|
+
- **A connector** — the implementation of `IAbapConnection` — depends on this
|
|
37
|
+
alone, and no longer on the object contracts.
|
|
38
|
+
- **`@mcp-abap-adt/interfaces-adt`** stands on it: its strategies are handed an
|
|
39
|
+
`IAdtWireResponse`. It does **not** re-export these names; import them from
|
|
40
|
+
here (decision 34).
|
|
41
|
+
|
|
42
|
+
Depends on `@mcp-abap-adt/interfaces-network`, and nothing else.
|
|
43
|
+
|
|
44
|
+
## Licence
|
|
45
|
+
|
|
46
|
+
`LGPL-3.0-only`.
|
|
@@ -0,0 +1,65 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Minimal response type for ADT requests.
|
|
3
|
+
*/
|
|
4
|
+
import type { IHttpHeaderValue, IHttpWireResponse } from '@mcp-abap-adt/interfaces-network';
|
|
5
|
+
import type { IAbapRequestOptions } from './IAbapRequestOptions';
|
|
6
|
+
export type IAdtHeaderValue = IHttpHeaderValue;
|
|
7
|
+
export interface IAdtWireResponse<T = any, D = any> extends Omit<IHttpWireResponse<T, D>, 'headers'> {
|
|
8
|
+
/**
|
|
9
|
+
* **The generic shape, narrowed by the headers ADT actually sends.** The
|
|
10
|
+
* frame itself — `data`, `status`, `statusText`, `config`, `request` — is
|
|
11
|
+
* `IHttpWireResponse` in `@mcp-abap-adt/interfaces-network`, because nothing
|
|
12
|
+
* about it is ABAP. What stays here is this list: `sap-adt-location` and the
|
|
13
|
+
* two spellings of `content-location` are where ADT puts the URI of what it
|
|
14
|
+
* just created, and a caller reads them by name.
|
|
15
|
+
*/
|
|
16
|
+
headers: Record<string, IAdtHeaderValue> & {
|
|
17
|
+
location?: string;
|
|
18
|
+
Location?: string;
|
|
19
|
+
'content-location'?: IAdtHeaderValue;
|
|
20
|
+
'Content-Location'?: IAdtHeaderValue;
|
|
21
|
+
'sap-adt-location'?: IAdtHeaderValue;
|
|
22
|
+
};
|
|
23
|
+
}
|
|
24
|
+
/**
|
|
25
|
+
* ABAP Connection interface
|
|
26
|
+
*
|
|
27
|
+
* Minimal interface for consumers to interact with SAP ADT.
|
|
28
|
+
* Implementation details (auth, token refresh, CSRF, cookies) are encapsulated.
|
|
29
|
+
*
|
|
30
|
+
* For JWT connections, token refresh is handled internally via ITokenRefresher.
|
|
31
|
+
* For Basic connections, no token refresh is needed.
|
|
32
|
+
*/
|
|
33
|
+
export interface IAbapConnection {
|
|
34
|
+
/**
|
|
35
|
+
* Initialize connection: fetch CSRF token and establish session cookies.
|
|
36
|
+
* Must be called before making any ADT requests.
|
|
37
|
+
*/
|
|
38
|
+
connect(): Promise<void>;
|
|
39
|
+
/**
|
|
40
|
+
* Get base URL of SAP system
|
|
41
|
+
*/
|
|
42
|
+
getBaseUrl(): Promise<string>;
|
|
43
|
+
/**
|
|
44
|
+
* Get current session ID (for stateful connections)
|
|
45
|
+
*/
|
|
46
|
+
getSessionId(): string | null;
|
|
47
|
+
/**
|
|
48
|
+
* Set session type for subsequent requests
|
|
49
|
+
* @param type - "stateful" for persistent session, "stateless" for independent requests
|
|
50
|
+
*/
|
|
51
|
+
setSessionType(type: 'stateful' | 'stateless'): void;
|
|
52
|
+
/**
|
|
53
|
+
* Make ADT request to SAP system
|
|
54
|
+
*
|
|
55
|
+
* Handles all auth concerns internally:
|
|
56
|
+
* - Adds authorization header (Basic or Bearer)
|
|
57
|
+
* - Manages CSRF token
|
|
58
|
+
* - Retries on 401/403 with token refresh (JWT only)
|
|
59
|
+
*
|
|
60
|
+
* @param options - Request options (url, method, data, etc.)
|
|
61
|
+
* @returns Promise with Axios response
|
|
62
|
+
*/
|
|
63
|
+
makeAdtRequest<T = any, D = any>(options: IAbapRequestOptions): Promise<IAdtWireResponse<T, D>>;
|
|
64
|
+
}
|
|
65
|
+
//# sourceMappingURL=IAbapConnection.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"IAbapConnection.d.ts","sourceRoot":"","sources":["../src/IAbapConnection.ts"],"names":[],"mappings":"AAAA;;GAEG;AACH,OAAO,KAAK,EACV,gBAAgB,EAChB,iBAAiB,EAClB,MAAM,kCAAkC,CAAC;AAC1C,OAAO,KAAK,EAAE,mBAAmB,EAAE,MAAM,uBAAuB,CAAC;AAEjE,MAAM,MAAM,eAAe,GAAG,gBAAgB,CAAC;AAE/C,MAAM,WAAW,gBAAgB,CAAC,CAAC,GAAG,GAAG,EAAE,CAAC,GAAG,GAAG,CAChD,SAAQ,IAAI,CAAC,iBAAiB,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,SAAS,CAAC;IAChD;;;;;;;OAOG;IACH,OAAO,EAAE,MAAM,CAAC,MAAM,EAAE,eAAe,CAAC,GAAG;QACzC,QAAQ,CAAC,EAAE,MAAM,CAAC;QAClB,QAAQ,CAAC,EAAE,MAAM,CAAC;QAClB,kBAAkB,CAAC,EAAE,eAAe,CAAC;QACrC,kBAAkB,CAAC,EAAE,eAAe,CAAC;QACrC,kBAAkB,CAAC,EAAE,eAAe,CAAC;KACtC,CAAC;CACH;AAED;;;;;;;;GAQG;AACH,MAAM,WAAW,eAAe;IAC9B;;;OAGG;IACH,OAAO,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;IAEzB;;OAEG;IACH,UAAU,IAAI,OAAO,CAAC,MAAM,CAAC,CAAC;IAE9B;;OAEG;IACH,YAAY,IAAI,MAAM,GAAG,IAAI,CAAC;IAE9B;;;OAGG;IACH,cAAc,CAAC,IAAI,EAAE,UAAU,GAAG,WAAW,GAAG,IAAI,CAAC;IAErD;;;;;;;;;;OAUG;IACH,cAAc,CAAC,CAAC,GAAG,GAAG,EAAE,CAAC,GAAG,GAAG,EAC7B,OAAO,EAAE,mBAAmB,GAC3B,OAAO,CAAC,gBAAgB,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;CACpC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"IAbapRequestOptions.d.ts","sourceRoot":"","sources":["../src/IAbapRequestOptions.ts"],"names":[],"mappings":"AAAA,MAAM,WAAW,mBAAmB;IAClC,GAAG,EAAE,MAAM,CAAC;IACZ,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,EAAE,MAAM,CAAC;IAChB,IAAI,CAAC,EAAE,OAAO,CAAC;IACf,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CAClC"}
|
|
@@ -0,0 +1,217 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Capability atoms for the connection.
|
|
3
|
+
*
|
|
4
|
+
* The same shape as the ADT capability atoms, for the same reason:
|
|
5
|
+
* `IAbapConnection` is the minimum every transport can honour, and this is a
|
|
6
|
+
* thing only some can. An RFC connection, a batch recorder and a test stub are
|
|
7
|
+
* all legitimate `IAbapConnection`s that own no HTTP session; making these
|
|
8
|
+
* methods mandatory would force each of them to implement a lie.
|
|
9
|
+
*
|
|
10
|
+
* ADDITIVE to `IAbapConnection`, which is unchanged — an implementation adds the
|
|
11
|
+
* atom when it genuinely supports it, and a consumer narrows to it.
|
|
12
|
+
*/
|
|
13
|
+
/** Error codes a session-aware connection raises. Match on these, not on messages. */
|
|
14
|
+
export declare const ADT_SESSION_ERROR: {
|
|
15
|
+
/** No usable session: never connected, or torn down since. */
|
|
16
|
+
readonly NOT_CONNECTED: "ADT_NOT_CONNECTED";
|
|
17
|
+
/**
|
|
18
|
+
* The server put us on a DIFFERENT session than the one we were using.
|
|
19
|
+
*
|
|
20
|
+
* Anything locked over the old session is orphaned: the lock handle refers to
|
|
21
|
+
* a session that no longer exists, so it can be neither used nor released.
|
|
22
|
+
*/
|
|
23
|
+
readonly SESSION_REPLACED: "ADT_SESSION_REPLACED";
|
|
24
|
+
/** A transport release is still pending; the session cannot yet be reused. */
|
|
25
|
+
readonly RELEASE_PENDING: "ADT_RELEASE_PENDING";
|
|
26
|
+
};
|
|
27
|
+
export type AdtSessionErrorCode = (typeof ADT_SESSION_ERROR)[keyof typeof ADT_SESSION_ERROR];
|
|
28
|
+
/**
|
|
29
|
+
* A connection whose session is owned, observable and explicitly torn down.
|
|
30
|
+
*
|
|
31
|
+
* Without this, "is my session still the one I locked over?" is unanswerable
|
|
32
|
+
* from the consumer's side: a connection can be handed a new server session
|
|
33
|
+
* silently, and a caller holding a lock has no way to notice.
|
|
34
|
+
*/
|
|
35
|
+
export interface ISessionLifecycleAware {
|
|
36
|
+
/**
|
|
37
|
+
* Tells the server this session is finished.
|
|
38
|
+
*
|
|
39
|
+
* **It notifies; it does not confirm.** Whether and when the session is
|
|
40
|
+
* actually freed is the server's affair, and nothing here checks afterwards.
|
|
41
|
+
* That is why there is no deadline to pass: waiting for the answer to a
|
|
42
|
+
* message whose answer is not acted on buys the caller nothing, and a
|
|
43
|
+
* goodbye carries no request timeout by design — so waiting is the one thing
|
|
44
|
+
* that could make a teardown unbounded.
|
|
45
|
+
*
|
|
46
|
+
* Resolves rather than throws, and **always settles**: whatever it could not
|
|
47
|
+
* finish is the connection's own state, not the caller's problem. A repeat
|
|
48
|
+
* call performs whatever is still owed.
|
|
49
|
+
*
|
|
50
|
+
* **Calling it again does not wait for the goodbye either**, which is what
|
|
51
|
+
* {@link flushGoodbye} is for. A caller about to reconnect wants both, in
|
|
52
|
+
* this order:
|
|
53
|
+
*
|
|
54
|
+
* ```typescript
|
|
55
|
+
* await conn.disconnect();
|
|
56
|
+
* await conn.flushGoodbye();
|
|
57
|
+
* await conn.connect();
|
|
58
|
+
* ```
|
|
59
|
+
*
|
|
60
|
+
* Without the middle line the next session opens while the previous one's
|
|
61
|
+
* goodbye is still being assembled, and the server keeps both. With it, the
|
|
62
|
+
* goodbye is given up to the budget to finish first.
|
|
63
|
+
*
|
|
64
|
+
* In-flight requests are NOT waited for. They continue as their caller
|
|
65
|
+
* arranged and nothing is aborted; their results can no longer affect this
|
|
66
|
+
* connection.
|
|
67
|
+
*/
|
|
68
|
+
disconnect(): Promise<void>;
|
|
69
|
+
/**
|
|
70
|
+
* Wait for the goodbye {@link disconnect} sent, within a budget.
|
|
71
|
+
*
|
|
72
|
+
* `disconnect()` dispatches the logoff and does not await it, deliberately: a
|
|
73
|
+
* goodbye carries no request timeout, and a server that never answers must
|
|
74
|
+
* not hold a teardown open. That is right for teardown and wrong for the one
|
|
75
|
+
* caller who reconnects — `disconnect()` then `connect()` opens the next
|
|
76
|
+
* session while the previous one's goodbye is still being assembled, and the
|
|
77
|
+
* server keeps both.
|
|
78
|
+
*
|
|
79
|
+
* Measured on E19 through `@mcp-abap-adt/adt-clients`, whose harness recycled
|
|
80
|
+
* the session after each test: a new ABAP session every one to two seconds
|
|
81
|
+
* for the length of a run, none released, each living to its own thirty
|
|
82
|
+
* minute idle timeout. Turning the recycling off left one session for the
|
|
83
|
+
* whole run, which is what the harness intended all along.
|
|
84
|
+
*
|
|
85
|
+
* **Bounded, and it resolves rather than throws.** The reason not to await
|
|
86
|
+
* inside `disconnect()` applies here too, so the wait has a budget and gives
|
|
87
|
+
* up quietly: "the goodbye has not arrived yet" is not a failure of whatever
|
|
88
|
+
* the caller does next. A goodbye that failed is likewise not the caller's
|
|
89
|
+
* problem — it is absorbed, not rethrown.
|
|
90
|
+
*
|
|
91
|
+
* **So a return is not a confirmation, and not even of dispatch.** Waiting out
|
|
92
|
+
* the budget resolves the same way as the goodbye finishing, and at that
|
|
93
|
+
* point the request may not have reached the wire at all — the promise it
|
|
94
|
+
* waits on covers assembling and sending, not just answering.
|
|
95
|
+
*
|
|
96
|
+
* The guarantee is exactly this: **the caller gives the goodbye up to the
|
|
97
|
+
* budget to finish before carrying on.** If it finishes in time there is no
|
|
98
|
+
* overlap with the next `connect()`. If it does not, the caller proceeds and
|
|
99
|
+
* the goodbye stays outstanding for as long as it takes — the budget bounds
|
|
100
|
+
* the waiting, not the overlap. Whether the server then releases the session
|
|
101
|
+
* is the server's affair, exactly as it is for `disconnect()`.
|
|
102
|
+
*
|
|
103
|
+
* Returns immediately when nothing was dispatched.
|
|
104
|
+
*
|
|
105
|
+
* @param timeoutMs how long to wait; an implementation chooses its default.
|
|
106
|
+
*/
|
|
107
|
+
flushGoodbye(timeoutMs?: number): Promise<void>;
|
|
108
|
+
/** Whether a caller may start work. False throughout a pending teardown. */
|
|
109
|
+
isConnected(): boolean;
|
|
110
|
+
/**
|
|
111
|
+
* Which server session the connection is on — or `null`, which says nothing
|
|
112
|
+
* about whether the connection is usable. **Use {@link isConnected} for that.**
|
|
113
|
+
*
|
|
114
|
+
* `null` means only that no identity is known, and there are two ways to get
|
|
115
|
+
* there: no session exists, or the connection is live over a server that
|
|
116
|
+
* issued no session cookie — some do not. So a live, working connection can
|
|
117
|
+
* return `null` here, and a consumer that reads `null` as "session lost"
|
|
118
|
+
* would tear down a connection that was fine.
|
|
119
|
+
*
|
|
120
|
+
* What it does identify is the SESSION, not the conversation: a stable
|
|
121
|
+
* client-side conversation id says nothing about whether the server replaced
|
|
122
|
+
* the session underneath it, which is precisely the failure this exists to
|
|
123
|
+
* expose. Compare two readings across an operation to detect a replacement.
|
|
124
|
+
*
|
|
125
|
+
* Only a CHANGED value is a replacement. `null` → non-null is an identity
|
|
126
|
+
* being learned, not a new session: the LOCK response is often the first to
|
|
127
|
+
* carry a session cookie, and reading that as a replacement would condemn the
|
|
128
|
+
* operation it just covered.
|
|
129
|
+
*/
|
|
130
|
+
getSessionIdentity(): string | null;
|
|
131
|
+
}
|
|
132
|
+
/**
|
|
133
|
+
* A connection whose responses resolve only after a later flush.
|
|
134
|
+
*
|
|
135
|
+
* A batch recorder collects requests and settles their promises when the batch
|
|
136
|
+
* executes. Awaiting one of those promises mid-recording deadlocks: the caller
|
|
137
|
+
* is blocked inside the very code that would have reached `execute()`.
|
|
138
|
+
*
|
|
139
|
+
* Deferral belongs to the connection, not to how a client was configured —
|
|
140
|
+
* which is why it is declared here rather than passed as an option. A consumer
|
|
141
|
+
* that wraps a recorder itself, without going through the batch client, still
|
|
142
|
+
* gets the guard.
|
|
143
|
+
*/
|
|
144
|
+
export interface IDeferredResponseConnection {
|
|
145
|
+
/** Always `true`. "Sometimes deferred" is not a state a caller could act on. */
|
|
146
|
+
readonly responsesAreDeferred: true;
|
|
147
|
+
}
|
|
148
|
+
/**
|
|
149
|
+
* A connection whose ordinary per-request deadline can be suspended.
|
|
150
|
+
*
|
|
151
|
+
* A `lock` → write → `unlock` sequence wants to run to completion. Aborting one
|
|
152
|
+
* of its requests part way ends nothing on the server — the server ends an ABAP
|
|
153
|
+
* session on its own idle timeout, which this side cannot see — it ends what
|
|
154
|
+
* this side *knows*: whether the write was applied becomes unanswerable, and
|
|
155
|
+
* the handle `unlock` needs is lost while the lock lives on in that session.
|
|
156
|
+
*
|
|
157
|
+
* Measured on a BTP trial: a `POST /sap/bc/adt/deletion/delete` abandoned at
|
|
158
|
+
* 45 s was followed by `400 … Session Timed Out or Not Found` carrying a *new*
|
|
159
|
+
* session cookie. Over HTTP a session is two layers — the ICF one the cookie
|
|
160
|
+
* addresses and the ABAP one beneath it holding the enqueue locks — and the
|
|
161
|
+
* abort replaces the first while stranding the second.
|
|
162
|
+
*
|
|
163
|
+
* **What this promises, exactly.** Inside a section the connection's ordinary
|
|
164
|
+
* per-request deadline does not apply. It does **not** promise that no request
|
|
165
|
+
* can ever be cut short: an implementation is free to keep a far larger ceiling
|
|
166
|
+
* — `@mcp-abap-adt/connection` raises it to `SAP_TIMEOUT_CRITICAL`, ten minutes
|
|
167
|
+
* by default — and a socket, a proxy or the process itself ends a request
|
|
168
|
+
* whatever a contract says. The guarantee is that the short deadline is out of
|
|
169
|
+
* the way, not that time is.
|
|
170
|
+
*
|
|
171
|
+
* Sections nest: the ceiling lifts on the outermost `begin` and ordinary
|
|
172
|
+
* deadlines resume after the matching `end`.
|
|
173
|
+
*
|
|
174
|
+
* Separate from {@link ISessionLifecycleAware} because it answers a different
|
|
175
|
+
* question: that one is *whose session is this*, this one is *may this request
|
|
176
|
+
* be cut short*. A batch recorder honours the first and has nothing to protect.
|
|
177
|
+
*/
|
|
178
|
+
export interface ICriticalSection {
|
|
179
|
+
/**
|
|
180
|
+
* Enter a section in which the ordinary per-request deadline is suspended.
|
|
181
|
+
*
|
|
182
|
+
* Nests. Every call needs its own {@link endCriticalSection}, and the usual
|
|
183
|
+
* shape is a `try`/`finally` so a throw inside the window still ends it.
|
|
184
|
+
*/
|
|
185
|
+
beginCriticalSection(): void;
|
|
186
|
+
/** Leave it. Ordinary deadlines resume when the outermost section ends. */
|
|
187
|
+
endCriticalSection(): void;
|
|
188
|
+
}
|
|
189
|
+
/**
|
|
190
|
+
* A connection that can be told what to ask the server to report about itself.
|
|
191
|
+
*
|
|
192
|
+
* `X-sap-adt-profiling` asks the server for its own processing time, and
|
|
193
|
+
* Eclipse sends it on every request. It is a per-connection default: a single
|
|
194
|
+
* request overrides it through `headers` on the request options, which every
|
|
195
|
+
* connection already accepts, so this atom exists for the *default* and not for
|
|
196
|
+
* the one-off.
|
|
197
|
+
*
|
|
198
|
+
* `null` asks for nothing. The value is a string rather than an enum because
|
|
199
|
+
* what a server accepts there is the server's business and grows without this
|
|
200
|
+
* package: `'server-time'` is what Eclipse asks for and is the sensible
|
|
201
|
+
* default.
|
|
202
|
+
*
|
|
203
|
+
* Worth stating: an implementation is not expected to read the answer back.
|
|
204
|
+
* The measurement is for whoever is looking at the wire, and being able to turn
|
|
205
|
+
* it off matters more than being able to consume it.
|
|
206
|
+
*/
|
|
207
|
+
export interface IRequestProfiling {
|
|
208
|
+
/**
|
|
209
|
+
* What to ask for on every request from now on, or `null` to ask nothing.
|
|
210
|
+
*
|
|
211
|
+
* @param what the `X-sap-adt-profiling` value; `'server-time'` is Eclipse's.
|
|
212
|
+
*/
|
|
213
|
+
setProfilingRequest(what: string | null): void;
|
|
214
|
+
/** What this connection is currently asking for. */
|
|
215
|
+
getProfilingRequest(): string | null;
|
|
216
|
+
}
|
|
217
|
+
//# sourceMappingURL=IConnectionCapabilities.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"IConnectionCapabilities.d.ts","sourceRoot":"","sources":["../src/IConnectionCapabilities.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,sFAAsF;AACtF,eAAO,MAAM,iBAAiB;IAC5B,8DAA8D;;IAE9D;;;;;OAKG;;IAEH,8EAA8E;;CAEtE,CAAC;AAEX,MAAM,MAAM,mBAAmB,GAC7B,CAAC,OAAO,iBAAiB,CAAC,CAAC,MAAM,OAAO,iBAAiB,CAAC,CAAC;AAE7D;;;;;;GAMG;AACH,MAAM,WAAW,sBAAsB;IACrC;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;OA+BG;IACH,UAAU,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;IAE5B;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;OAqCG;IACH,YAAY,CAAC,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAEhD,4EAA4E;IAC5E,WAAW,IAAI,OAAO,CAAC;IAEvB;;;;;;;;;;;;;;;;;;;OAmBG;IACH,kBAAkB,IAAI,MAAM,GAAG,IAAI,CAAC;CACrC;AAED;;;;;;;;;;;GAWG;AACH,MAAM,WAAW,2BAA2B;IAC1C,gFAAgF;IAChF,QAAQ,CAAC,oBAAoB,EAAE,IAAI,CAAC;CACrC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AACH,MAAM,WAAW,gBAAgB;IAC/B;;;;;OAKG;IACH,oBAAoB,IAAI,IAAI,CAAC;IAE7B,2EAA2E;IAC3E,kBAAkB,IAAI,IAAI,CAAC;CAC5B;AAED;;;;;;;;;;;;;;;;;GAiBG;AACH,MAAM,WAAW,iBAAiB;IAChC;;;;OAIG;IACH,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,IAAI,GAAG,IAAI,CAAC;IAE/C,oDAAoD;IACpD,mBAAmB,IAAI,MAAM,GAAG,IAAI,CAAC;CACtC"}
|
|
@@ -0,0 +1,29 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
/**
|
|
3
|
+
* Capability atoms for the connection.
|
|
4
|
+
*
|
|
5
|
+
* The same shape as the ADT capability atoms, for the same reason:
|
|
6
|
+
* `IAbapConnection` is the minimum every transport can honour, and this is a
|
|
7
|
+
* thing only some can. An RFC connection, a batch recorder and a test stub are
|
|
8
|
+
* all legitimate `IAbapConnection`s that own no HTTP session; making these
|
|
9
|
+
* methods mandatory would force each of them to implement a lie.
|
|
10
|
+
*
|
|
11
|
+
* ADDITIVE to `IAbapConnection`, which is unchanged — an implementation adds the
|
|
12
|
+
* atom when it genuinely supports it, and a consumer narrows to it.
|
|
13
|
+
*/
|
|
14
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
15
|
+
exports.ADT_SESSION_ERROR = void 0;
|
|
16
|
+
/** Error codes a session-aware connection raises. Match on these, not on messages. */
|
|
17
|
+
exports.ADT_SESSION_ERROR = {
|
|
18
|
+
/** No usable session: never connected, or torn down since. */
|
|
19
|
+
NOT_CONNECTED: 'ADT_NOT_CONNECTED',
|
|
20
|
+
/**
|
|
21
|
+
* The server put us on a DIFFERENT session than the one we were using.
|
|
22
|
+
*
|
|
23
|
+
* Anything locked over the old session is orphaned: the lock handle refers to
|
|
24
|
+
* a session that no longer exists, so it can be neither used nor released.
|
|
25
|
+
*/
|
|
26
|
+
SESSION_REPLACED: 'ADT_SESSION_REPLACED',
|
|
27
|
+
/** A transport release is still pending; the session cannot yet be reused. */
|
|
28
|
+
RELEASE_PENDING: 'ADT_RELEASE_PENDING',
|
|
29
|
+
};
|
|
@@ -0,0 +1,15 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* The deadlines a caller gives an ADT connection.
|
|
3
|
+
*
|
|
4
|
+
* **`csrf` is why this is not in `@mcp-abap-adt/interfaces-network`.** Fetching
|
|
5
|
+
* a CSRF token is an SAP operation, not a transport primitive, and `long` is a
|
|
6
|
+
* client's policy for a long-polling read rather than anything the network
|
|
7
|
+
* layer knows. It was in that package until `interfaces-adt` 9.0.0, against its
|
|
8
|
+
* own README's claim that it holds nothing SAP-specific.
|
|
9
|
+
*/
|
|
10
|
+
export interface ITimeoutConfig {
|
|
11
|
+
default: number;
|
|
12
|
+
csrf: number;
|
|
13
|
+
long: number;
|
|
14
|
+
}
|
|
15
|
+
//# sourceMappingURL=ITimeoutConfig.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"ITimeoutConfig.d.ts","sourceRoot":"","sources":["../src/ITimeoutConfig.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AACH,MAAM,WAAW,cAAc;IAC7B,OAAO,EAAE,MAAM,CAAC;IAChB,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;CACd"}
|
package/dist/index.d.ts
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* @mcp-abap-adt/interfaces-adt-connection
|
|
3
|
+
*
|
|
4
|
+
* The ABAP connection: what an ADT request is sent through, and what comes
|
|
5
|
+
* back off the wire.
|
|
6
|
+
*
|
|
7
|
+
* **Its own package because it moves on its own schedule.** These four files
|
|
8
|
+
* were `interfaces-adt/src/connection/` until 11.0.0, and every release of the
|
|
9
|
+
* object contracts — 9, 10 and 11 in one day, none touching them — made the
|
|
10
|
+
* connector either follow or install a second copy. The subject is still ADT
|
|
11
|
+
* (decision 35); what separated it is the release rate (decision 38).
|
|
12
|
+
*/
|
|
13
|
+
export type { IAbapConnection, IAdtWireResponse } from './IAbapConnection';
|
|
14
|
+
export type { IAbapRequestOptions } from './IAbapRequestOptions';
|
|
15
|
+
export type { AdtSessionErrorCode, ICriticalSection, IDeferredResponseConnection, IRequestProfiling, ISessionLifecycleAware, } from './IConnectionCapabilities';
|
|
16
|
+
export { ADT_SESSION_ERROR } from './IConnectionCapabilities';
|
|
17
|
+
export type { ITimeoutConfig } from './ITimeoutConfig';
|
|
18
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AACH,YAAY,EAAE,eAAe,EAAE,gBAAgB,EAAE,MAAM,mBAAmB,CAAC;AAC3E,YAAY,EAAE,mBAAmB,EAAE,MAAM,uBAAuB,CAAC;AACjE,YAAY,EACV,mBAAmB,EACnB,gBAAgB,EAChB,2BAA2B,EAC3B,iBAAiB,EACjB,sBAAsB,GACvB,MAAM,2BAA2B,CAAC;AACnC,OAAO,EAAE,iBAAiB,EAAE,MAAM,2BAA2B,CAAC;AAC9D,YAAY,EAAE,cAAc,EAAE,MAAM,kBAAkB,CAAC"}
|
package/dist/index.js
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
+
exports.ADT_SESSION_ERROR = void 0;
|
|
4
|
+
var IConnectionCapabilities_1 = require("./IConnectionCapabilities");
|
|
5
|
+
Object.defineProperty(exports, "ADT_SESSION_ERROR", { enumerable: true, get: function () { return IConnectionCapabilities_1.ADT_SESSION_ERROR; } });
|
package/package.json
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
1
|
+
{
|
|
2
|
+
"name": "@mcp-abap-adt/interfaces-adt-connection",
|
|
3
|
+
"version": "1.0.0",
|
|
4
|
+
"description": "The ABAP connection contract: IAbapConnection, its capability atoms, the wire response and the session error codes",
|
|
5
|
+
"main": "dist/index.js",
|
|
6
|
+
"types": "dist/index.d.ts",
|
|
7
|
+
"files": [
|
|
8
|
+
"dist",
|
|
9
|
+
"README.md",
|
|
10
|
+
"CHANGELOG.md",
|
|
11
|
+
"LICENSE",
|
|
12
|
+
"COPYING"
|
|
13
|
+
],
|
|
14
|
+
"sideEffects": false,
|
|
15
|
+
"author": "Oleksii Kyslytsia <oleksij.kyslytsja@gmail.com>",
|
|
16
|
+
"license": "LGPL-3.0-only",
|
|
17
|
+
"homepage": "https://github.com/fr0ster/mcp-abap-adt-interfaces/tree/master/packages/interfaces-adt-connection#readme",
|
|
18
|
+
"bugs": {
|
|
19
|
+
"url": "https://github.com/fr0ster/mcp-abap-adt-interfaces/issues"
|
|
20
|
+
},
|
|
21
|
+
"repository": {
|
|
22
|
+
"type": "git",
|
|
23
|
+
"url": "git+https://github.com/fr0ster/mcp-abap-adt-interfaces.git",
|
|
24
|
+
"directory": "packages/interfaces-adt-connection"
|
|
25
|
+
},
|
|
26
|
+
"publishConfig": {
|
|
27
|
+
"access": "public"
|
|
28
|
+
},
|
|
29
|
+
"scripts": {
|
|
30
|
+
"clean": "rm -rf dist tsconfig.tsbuildinfo tsconfig.build.tsbuildinfo",
|
|
31
|
+
"build": "tsc -b tsconfig.build.json",
|
|
32
|
+
"test:check": "tsc -p tsconfig.json --noEmit",
|
|
33
|
+
"prepublishOnly": "npm run --prefix ../.. check"
|
|
34
|
+
},
|
|
35
|
+
"engines": {
|
|
36
|
+
"node": ">=18.0.0"
|
|
37
|
+
},
|
|
38
|
+
"dependencies": {
|
|
39
|
+
"@mcp-abap-adt/interfaces-network": "^2.0.0"
|
|
40
|
+
}
|
|
41
|
+
}
|