Domino JNA

Cross-platform access to HCL Notes/Domino C API methods from Java

Features

The project provides functionality that is not available in the classic Java API of HCL Notes/Domino or that is poorly implemented, for example:

Document API

basic APIs to read and write note (document) item values like String/Double/NotesDateTime/Calendar single and multiple values
direct attachment streaming to create and extract files (HCL's Java API extracts files to temp disk space first to read attachment data and only supports adding files stored on disk as document attachments)
quick check if a document is editable by a specified user (without the need to scan through author items)
faster formula execution on documents with document modified/selected/deleted info, more than 64K of result data and applied security (e.g. no changes to Notes.ini)
creation of ghost notes (documents that do not appear in any views)

Database API

fulltext index creation with all available options
searching NSF data with formula (NSFSearch in the C API) with all parameters and return values, e.g. **get summary buffer data for each document matching a formula and compute your own field values like the view indexer does, including the readers list **
run agents bypassing ECL checks, pass an in-memory document for data exchange readable as Session.DocumentContext and redirect Agent output to a Java Writer
read/write replication data (change replica id and flags)
clearing the replication history
fast noteid / UNID bulk conversion with lookup of "modified in this file" property (part of the note OID - originator id)
API to read the item definition table of a database (all fieldnames and fieldtypes, useful to track FT Search issues and provide fieldname typeahead)
extended folder operations, e.g. create/delete/move/copy with content/rename, creation of new folder with specifying where to get the design
API to incrementally read folder changes (added/remove note ids)
efficient lookup of documents without lookup views by a value in the $Name field (see testcase TestNotePrimaryKey)

Domino views API

view lookups using different key formats (e.g. strings, numbers, dates, date ranges) and equality/inequality searches (e.g. find less or greater than a search key)
decodes all available types of view column data types (string, string list, number, number list, datetime, datetime list) and row data (e.g. just note id, UNID, counts, unread flag etc.)
read view data as another Notes user
extract read access information for rows in a view that Domino stores in the view index
separation of Notes views and their data into multiple databases (like programmatically creating private views and keeping them up-to-date incrementally)
dynamic filtering of view rows based on a Note id list with paging support (this really rocks!)
reading categorized views with expanded/collapsed entries and min/max level
differential view reads : on second view lookup, only read rows that have changed
support for view resorting (changing the collation in C API terms)

DQL / QueryResultsProcessor API

run DQL (Domino Query Language) queries against databases on Domino V10+ and return the result dynamically sorted
QueryResultsProcessor API to create JSON data and QRP views

Virtual View API

Our own replacement for Domino views and the QueryResultsProcessor!

See testcase TestVirtualView for code examples. And here is a recording of the OpenNTF Developer Variety Hour where I presented the Virtual View API in detail.

multi-DB views, even work between servers and client-server
view structure similar to Domino (multi level categorization, sorted columns)
support for sums / average values, child and descendant counts
compute column values via formula or Java code
incremental view updates, so no rebuilt required
full control when the view is updated, optional read locks to have exclusive access
view is populated by the server, shared across users
for each user we check which view entries the user is allowed to see (checks DB ACL level and compares user names list for each DB with computed list of document readers list)
for category entries we accumulate the readers of all descendant docs to quickly skip categories that would be empty for a user
these collected readers stats can be used for analysis purposes
several data sources can be combined to produce view data
- datasource 1: run NSF search with a formula (incrementally), can search data and design documents, optional post processing with FT search
- datasource 2: profile documents
- datasource 3: read note ids from a folder (incrementally)
- datasource 4: compute column values from any list of note ids
- custom datasources can be implemented and added to the virtual view
VirtualViewNavigator to read the view entries that a user is allowed to see, either all entries/category descendants, just docs/category, support for expanded/selected entries, upwards/downwards/paging like NIFReadEntries, keyword and range lookups
fast (e.g. processes 40.000 fakename docs and builds the view in 2-3 seconds)
VirtualView currently stored in Java heap (each VirtualViewEntry with a ConcurrentSkipListMap for the sorted children), might add support for serialization to disk later on
not a Domino specific implementation, would work in other environments as well

Richtext / MIME API

richtext item reading: convenience method to extract text content, advanced API to read all CD records (IRichTextNavigator)
richtext item writing: create richtext items with text, images, doclinks, by rendering other notes and appending other richtet items
add PNG images to richtext items: something that cannot easily be done yet in the Notes Client UI
richtext item conversion: multi-level conversion of richtext item content, e.g. to add/remove file hotspots (file icons with custom image that open file on click) independent from the actual file attachment or do mail merge with richtext
richtext-html conversion with advanced quality and access to embedded images
API to read and write MIME items

Design API

read design collection with design elements in a database
design richtext processing: e.g. to apply a string replacement and recompile formulas for computed text/subforms or hotspots and find all fields in a form
DXL exporter with the option to write the DXL into a stream (Notes.jar classes fill up the Java heap with a giant DXL string)

Sync API

incremental data synchronization with external databases or indexers. Generic design, sample implementation for CQEngine and SQLite
supports incremental synchronization of Domino databases by reading noteid lists of modified and deleted documents (HCL's Java API does not return ids of deleted docs)

Server API

quick reading of files and folders in the Domino data directory or in subdirectories (HCL's DbDirectory is slow for many DBs and does not support subdirectory scanning)
compute @Usernameslist values for any Notes user on local and remote server
APIs to read and write Out-of-Office information (OOO) of any user
API to read the remote server console and send commands.

Calendaring API

APIs to read extended busytime information like UNID/start/end of busytime entries (not just the freetime search that HCL provides)
APIs to create/read/update Domino appointments via iCal format and the option to only output selected fields into the generated iCal (e.g. only start/end/summary) and full meeting workflow action support (accept/decline invitation etc.)

Notes Client API

APIs to read and modify the ECL
read/write access for the Notes Client workspace (desktop8.nsk), e.g. read/write page and chicklet infos (titles, server, filename, tabindex, x, y, classic and modern icon), move pages with their icons, move replicas on top

Security API

SSO token computation (with tokens also working on Websphere)
APIs to get/put/sync IDs with the ID Vault and to sign/encrypt/decrypt documents and attachments

Warning

The project gives access to some really low level functions of HCL Notes/Domino. Using them in the wrong way or sending unexpected parameter values might crash your application server, so make sure you know what you are doing and test your code on a local machine first!

One reason for open sourcing all this stuff was to get more hands on it and make it as robust as possible.

Supported platforms

The code should run in 32 and 64 bit Notes Client and Domino server environments on Windows, Linux and Mac.

It is not expected to run without changes on other platforms, mainly because of little endian / big endian differences or memory alignments, but we don't currently have access to those platforms anyway.

XPages

Domino JNA can be used in XPages applications!

See the release section for ready to install builds. Those work similar to the XPages Extension Libary. So you need to install the provided OSGi plugins both in Domino Designer and the Domino Server.

Here are installation instructions how to do this: link.

The API is available in the code editor after you activate the com.mindoo.domino.jna.xsp.library entry in the xsp.properties file.

If you are having trouble getting Domino JNA to work in Designer 9.0.1 FP10 or later, see this blog posting for a workaround: link

In short, you need to add <notesdata>\workspace\applications\eclipse to your target platform (replace <notesdata> with your data directory path).

Maven

Domino JNA is available on Maven Central: https://mvnrepository.com/artifact/com.mindoo.domino/domino-jna.

<dependency>
    <groupId>com.mindoo.domino</groupId>
    <artifactId>domino-jna</artifactId>
    <version>see Maven Central for latest version number</version>
</dependency>

Snapshot releases may be provided on https://oss.sonatype.org/content/repositories/snapshots for bug analysis purpose, e.g. via a Github issue.

Standalone applications

There is a sample application available that demonstrates how to use Domino JNA in standalone Java applications.

Usage

Here is a code snippet for the API usage. It opens a database and filters view entries.

NotesGC.runWithAutoGC(new Callable<Object>() {

	public Object call() throws Exception {
		//open database with the same access rights as a lotus.domino.Session
		NotesDatabase dbData = new NotesDatabase(session, "", "fakenames.nsf");

		//alternative: open database as another user (used for read and write access):
		//NotesDatabase dbData = new NotesDatabase("", "fakenames.nsf", "John Doe/Mindoo");

		//open database as the server:				
		//NotesDatabase dbData = new NotesDatabase("", "fakenames.nsf", "");

		//open People view (in C API called collection)
		NotesCollection peopleView = dbData.openCollectionByName("People");

		//read all note ids from the collection
		boolean includeCategoryIds = false;
		LinkedHashSet<Integer> allIds = peopleView.getAllIds(includeCategoryIds);
				
		//pick random note ids
		Integer[] allIdsArr = allIds.toArray(new Integer[allIds.size()]);
		Set<Integer> pickedNoteIds = new HashSet<Integer>();
		while (pickedNoteIds.size() < 1000) {
			int randomIndex = (int) (Math.random() * allIdsArr.length);
			int randomNoteId = allIdsArr[randomIndex];
			pickedNoteIds.add(randomNoteId);
		}
				
		//populate the collection's selected list with picked ids
		boolean clearPreviousSelection = true
		peopleView.select(pickedNoteIds, clearPreviousSelection);

		//next, traverse selected entries only, starting at position "0" (top of the view)
		String startPos = "0";
		//skip from "0" to the first entry that we are allowed to read
		//(its position could be different from "1" caused by reader fields)
		int entriesToSkip = 1;
		//add all read entries to the result list
		int entriesToReturn = Integer.MAX_VALUE;
		//tell the API how to navigate in the view: from one entry in the selectedList
		//to the next one (in view ordering)
		EnumSet<Navigate> returnNavigator = EnumSet.of(Navigate.NEXT_SELECTED);
		//preload the maximum number of entries, can be useful when implementing
		//filter method in EntriesAsListCallback
		int bufferSize = Integer.MAX_VALUE;
		//tell the API which data we want to read (in this case note ids and column values map)
		EnumSet<ReadMask> returnData = EnumSet.of(ReadMask.NOTEID, ReadMask.SUMMARYVALUES);
		
		List<NotesViewEntryData> selectedEntries = peopleView.getAllEntries(startPos, entriesToSkip,
				returnNavigator, Integer.MAX_VALUE,
				returnData, new EntriesAsListCallback(entriesToReturn));
				
		for (NotesViewEntryData currEntry : selectedEntries) {
		
			//check that all entries that we read were from our picked id list
			Assert.assertTrue("Entry read from view is contained in selected list",
				pickedNoteIds.contains(currEntry.getNoteId()));
				
			//read column values with their programmatic name
			String firstName = (String) currEntry.get("firstname");
			String lastName = (String) currEntry.get("lastname");
			//...
		}
		
		//now remove all read ids from pickedNoteIds and make sure that we found everything
		//we were searching for
		for (NotesViewEntryData currEntry : selectedEntries) {
			pickedNoteIds.remove(currEntry.getNoteId());
		}
		
		Assert.assertTrue("All ids from the selected list can be found in the view", pickedNoteIds.isEmpty());
		return null;
	}
});

Reducing the view to a specific selection (of note ids) is already the first big surprise, if you only know HCL's Java API for Domino.

Comparable to reading fulltext search results, but a lot more powerful! And the cool thing is that Domino handles the filtering and even the paging for you (entriesToSkip parameter). so you don't have to waste time to read and skip data slowly in your own code.

As you can see, all calls have to be wrapped in NotesGC.runWithAutoGC code blocks (which can also be nested).

We do this to automatically collect allocated C handles and free them when the code block is done.

In many cases, this should avoid manual recycling of API objects, but for some edge cases, objects like NotesCollection (which is the term for Notes View in the C API), NotesNote (a document) or NotesIDTable do have a recycle() method.

When running in an XPages environment, NotesGC.runWithAutoGC can be omitted when the code processes a HTTP request (e.g. an XAgent). It is only required if you run code in separate threads, e.g. using the SessionCloner class.

Further information

Next steps

This project is not done yet, this is just the beginning. Here are some of the things that we plan to do:

write blog entries explaining the API internals
add more API methods, e.g. for new DQL features of Domino 11 and 12
write more testcases
add more syntactical sugar, hide complexity

Licence

The code is available under Apache 2.0 license.

Java Profiler

For development, we are using JProfiler Java profiler. It's the perfect tool to analyze performance bottlenecks and memory leaks.

Creating your own build

The following instructions are only relevant when you want to create your own Domino JNA release version.

Registration of local Notes.jar, lwpd.commons.jar and lwpd.domino.napi.jar

There are three JAR files that are part of every Notes Client installation and that are required to compile the Domino JNA code.

On macOS, you can find the files in these locations:

/Applications/HCL Notes.app/Contents/Resources/jvm/lib/ext/Notes.jar
/Applications/HCL Notes.app/Contents/Eclipse/shared/eclipse/plugins/com.ibm.commons_-version-/lwpd.commons.jar
/Applications/HCL Notes.app/Contents/Eclipse/shared/eclipse/plugins/com.ibm.domino.napi_-version-/lwpd.domino.napi.jar

On Windows you fine them here:

C:\Program Files (x86)\HCL\Notes\jvm\lib\ext\Notes.jar
C:\Program Files (x86)\HCL\Notes\osgi\shared\eclipse\plugins\com.ibm.commons_-version-\lwpd.commons.jar
C:\Program Files (x86)\HCL\Notes\osgi\shared\eclipse\plugins\com.ibm.domino.napi_-version-\lwpd.domino.napi.jar

These files need to be registered as Maven artifacts with the right groupId / artifactId on the local machine, because they are not available on Maven Central (com.ibm.commons is there, but outdated).

For the Mac, use this syntax (replace "-version-" with the right version on your machine):

mvn install:install-file -Dfile="/Applications/HCL Notes.app/Contents/Resources/jvm/lib/ext/Notes.jar" -DgroupId=com.ibm -DartifactId=domino-api-binaries -Dversion=11.0.0 -Dpackaging=jar

mvn install:install-file -Dfile="/Applications/HCL Notes.app/Contents/Eclipse/shared/eclipse/plugins/com.ibm.commons_-version-/lwpd.commons.jar" -DgroupId=com.ibm -DartifactId=ibm-commons -Dversion=11.0.0 -Dpackaging=jar

mvn install:install-file -Dfile="/Applications/HCL Notes.app/Contents/Eclipse/shared/eclipse/plugins/com.ibm.domino.napi_-version-/lwpd.domino.napi.jar" -DgroupId=com.ibm -DartifactId=napi -Dversion=11.0.0 -Dpackaging=jar

For Windows, use this syntax (replace "-version-" with the right version on your machine):

mvn install:install-file -Dfile="C:\Program Files (x86)\HCL\Notes\jvm\lib\ext\Notes.jar" -DgroupId=com.ibm -DartifactId=domino-api-binaries -Dversion=11.0.0 -Dpackaging=jar

mvn install:install-file -Dfile="C:\Program Files (x86)\HCL\Notes\osgi\shared\eclipse\plugins\com.ibm.commons_-version-\lwpd.commons.jar" -DgroupId=com.ibm -DartifactId=ibm-commons -Dversion=11.0.0 -Dpackaging=jar

mvn install:install-file -Dfile="C:\Program Files (x86)\HCL\Notes\osgi\shared\eclipse\plugins\com.ibm.domino.napi_-version-\lwpd.domino.napi.jar" -DgroupId=com.ibm -DartifactId=napi -Dversion=11.0.0 -Dpackaging=jar

Maven build

Mac: On Mac, use this syntax to build Domino JNA against the Notes Client:

mvn -DJVMPARAMS=-d64 -DDOMINOOSGIDIR=/Applications/HCL\ Notes.app/Contents/MacOS -DDOMINODIR=/Applications/HCL\ Notes.app/Contents/MacOS -DNOTESINI=~/Library/Preferences/Notes\ Preferences clean install -Dmaven.test.skip=true

Windows: To build against the HCL Notes Client on Windows use this syntax:

mvn -DJVMPARAMS= -DDOMINOOSGIDIR="C:\Program Files (x86)\HCL\Notes\osgi" -DDOMINODIR="C:\Program Files (x86)\HCL\Notes" -DNOTESINI="C:\Program Files (x86)\HCL\Notes\Notes.ini" clean install -Dmaven.test.skip=true

After the build is done, the directory target/lib contains all recursive dependencies required to use the library, e.g. JNA and Apache tool libraries.

Running the test cases

The project contains a number of test cases that demonstrate how the API is used. The project is not ready to run the tests automatically as part of the build. That's why they are disabled by default and should only be run manually for now.

We are still working on the tests to make the more robust and let them set up their required test environment. In addition there are issues in macOS when running the tests via Surefire plugin, because DYLD_LIBRARY_PATH is not allowed to be set via bash scripts anymore, causing load errors for libnotes.dylib, libxml.dylib and others.

Sample databases

The test cases use sample databases that we provide for download and will update from time to time depending on the requirements of newer testcases.

You can download the two sample databases fakenames.nsf and fakenames-views.nsf from this URL:

Download Link

Next, place them in the data folder of your HCL Notes Client.

fakenames.nsf is a directory database that contains about 40,000 sample documents and some additional lookup views, fakenames-views.nsf uses the same database design, but does not contain any data.

We use fakenames-views.nsf to demonstrate indexing of external Domino data (take a local view and incrementally pull data from an external NSF database, like the Notes Client does with private views).

In Eclipse, make sure to add the following environment variables (with the right paths for your machine) to the Run Configurations to run testcases:

Windows:

PATH = C:\Program Files (x86)\IBM\Notes

Mac:

DYLD_LIBRARY_PATH=/Applications/HCL Notes.app/Contents/MacOS
Notes_ExecDirectory=/Applications/HCL Notes.app/Contents/MacOS
NOTESBIN=/Applications/HCL Notes.app/Contents/MacOS
NotesINI=/Users/klehmann/Library/Preferences/Notes Preferences
PATH=/Applications/HCL Notes.app/Contents/MacOS

Creating XPages plugin build

The projects com.mindoo.domino.jna.xsp.build and domino-target contain build scripts to use Domino JNA in XPages applications, similar to HCL's XPages Extension Library.

Please use the following steps to create a build or just download a binary build from the "releases" section.

1. Target platform

To create the build, you first need to create the Eclipse target platform that we will compile against. This step is only required once.

In project domino-target, call mvn clean install with the same parameters described above (JVMPARAMS, DOMINOOSGIDIR, DOMINODIR and NOTESINI).

When the build is done, the directory domino-target/target/repository contains a P2 Update Site containing the features and plugins of the installed HCL Notes Client that can be used by Maven/Tycho.

2. Build Update Site

Next call mvn clean install (also with parameters JVMPARAMS, DOMINOOSGIDIR, DOMINODIR and NOTESINI) in project com.mindoo.domino.jna.xsp.build.

This copies the current Domino JNA source code from project domino-jna into two Eclipse plugins com.mindoo.domino.jna.xsp/jna-src and com.mindoo.domino.jna.xsp.source/jna-src and starts the compilation.

com.mindoo.domino.jna.xsp provides the extension library for XPages and com.mindoo.domino.jna.xsp.source provides the source code for the Java editor of HCL Domino Designer.

You can find the created Update Site in directory com.mindoo.domino.jna.xsp-updatesite/target/site.

Dependencies

The code uses the following open source projects:

metadata-extractor for image file metadata extraction, available under Apache 2.0 license

Apache Commons Collections 4 for case insensitive maps, available under Apache 2.0 license

cglib for bytecode manipulation, available under Apache 2.0 license

klehmann/domino-jna