Release share v0.52.0 (2024-01-06)v0.52.0

=== New features

* ssh/config: add method MarshalText and WriteTo
* lib/ssh: implement method Output on Client
* ssh/sftp: implement method MkdirAll on Client
* cmd/httpdfs: implement [libhttp.Server] with [memfs.MemFS]

=== Breaking changes

* ssh/config: refactoring the Config merge
* ssh/config: add parameter Config to NewSection
* lib/ssh: add parameter context to Execute method
* lib/time: remove UnixMicro and UnixMilli
* lib/io: removed, this package has been merged into "lib/os"
* lib/parser: removed, this package has been merged into lib/strings

=== Bug fixes

* ssh/config: fix setting the default values
* ssh/config: set the Hostname if its not set on [Config.Get]
* http/sseclient: fix data race on [Client.Close]
* http/sseclient: fix Retry value not set to millisecond
* ssh/sftp: fix Stat on empty remote file name
* ssh/sftp: fix non-nil returned error on Close

=== Enhancements

* ssh/config: merge the Section slice values on [Section.merge]
* ssh/config: set the default UserKnownHostsFile in setDefaults

2 files changed, 1337 insertions, 0 deletions

diff --git a/_doc/CHANGELOG_2023.adoc b/_doc/CHANGELOG_2023.adoc
new file mode 100644
index 00000000..92dae827
--- /dev/null
+++ b/_doc/CHANGELOG_2023.adoc
@@ -0,0 +1,1334 @@
+=  CHANGELOG
+Shulhan <ms@kilabit.info>
+:toc:
+:sectanchors:
+:sectlinks:
+
+This library is released every month, usually at the first week of month.
+
+link:CHANGELOG_2022.html[Changelog in 2022^].
+This is changelog for share module since v0.33.0 until v0.42.0.
+
+link:CHANGELOG_2021.html[Changelog in 2021^].
+This is changelog for share module since v0.22.0 until v0.32.0.
+
+link:CHANGELOG_2020.html[Changelog in 2020^].
+This is changelog for share module since v0.12.0 until v0.21.0.
+
+link:CHANGELOG_2018-2019.html[Changelog from 2018 to 2019^].
+This is changelog for share module since v0.1.0 until v0.11.0.
+
+
+[#v0_51_0]
+== share v0.51.0 (2023-12-06)
+
+[#v0_51_0__new_features]
+=== New features
+
+lib/http: implement Server-Sent Events (SSE)::
++
+--
+For server SSE, we add new type SSEEndpoint, for registering endpoint
+that can handle SSE.
+
+For client SSE, we add it in new sub package "sseclient".
+--
+
+lib/net: add method Read::
++
+--
+The Read method read packet from raw connection.
+
+If the conn parameter is nil it will return [net.ErrClosed].
+
+The bufsize parameter set the size of buffer for each read operation,
+default to 1024 if not set or invalid (less than 0 or greater than
+65535).
+
+The timeout parameter set how long to wait for data before considering
+it as failed.
+If its not set, less or equal to 0, it will wait forever.
+If no data received and timeout is set, it will return [ErrReadTimeout].
+
+If there is data received and connection closed at the same time, it will
+return the data first without error.
+The subsequent Read will return empty packet with [ErrClosed].
+--
+
+lib/crypto: add support for reading passphrase using SSH_ASKPASS::
++
+--
+If the library failed to changes os.Stdin to raw, it will try to use
+a program defined in SSH_ASKPASS environment variable.
+
+The SSH_ASKPASS is controlled by environment SSH_ASKPASS_REQUIRE.
+
+- If SSH_ASKPASS_REQUIRE is empty the passphrase will read from
+  terminal first, if not possible then using SSH_ASKPASS program.
+
+- If SSH_ASKPASS_REQUIRE is set to "never" the passphrase will read
+  from terminal only.
+
+- If SSH_ASKPASS_REQUIRE is set to "prefer", the passphrase will read
+  using SSH_ASKPASS program not from terminal, but require
+  DISPLAY environment to be set.
+
+- If SSH_ASKPASS_REQUIRE is set to "force", the passphrase will read
+  using SSH_ASKPASS program not from terminal, without checking DISPLAY
+  environment.
+--
+
+This changes affect the [ssh.NewClientInteractive] indirectly.
+
+lib/memfs: add method JSON to Node::
++
+--
+The JSON method encode the Node into JSON.
+This method provides an alternative to MarshalJSON with granular options,
+depth and withoutModTime.
+
+The depth set the level of childs to be encoded.
+The depth=0 only encode the Root Node itself (with its childs), depth=1
+encode the Root node and its subdirectories, and so on.
+
+If withoutModTime is set to true, all of the node ModTime will not be
+included in output.
+--
+
+
+[#v0_51_0__enhancements]
+=== Enhancements
+
+lib/ini: create the file if its not exist on Open::
++
+Previously, Open will return an error if file not exist.
+This changes create the file if not exist.
+
+lib/net: use [time.Ticker] to loop in WaitAlive::
++
+--
+The Timeout field in net.Dialer sometimes does not have any effect.
+In the for-loop, we expect that each Dial at least consume 100 ms
+before returning an error.
+
+Turns out, on several occasion, the loop run too quickly, less than the
+passed timeout.
+This is probably because the listening address has not been
+active yet and we run it in the same machine, the OS (or Go?) return
+immediately without waiting for timeout.
+--
+
+lib/memfs: add sub file system::
++
+--
+This implemented simplified version of merging multiple MemFS instances.
+The sub file system (subfs) can be added to the parent MemFS instance
+by calling new method "Merge".
+
+  parent.Merge(other *MemfS)
+
+When Get method called, each subfs will be evaluated in order of Merge
+called.
+
+This deprecated the function Merge.
+--
+
+lib/memfs: improve the refresh method::
++
+--
+Instead of refreshing from the Root, find the directory that closes
+to the requested path.
+
+While at it, simplify the returned error.
+--
+
+lib/ini: append variable into section before any empty lines::
++
+--
+Previously, appending a new variable into section always create an empty
+line before after the variable.
+This changes fix this by appending variable only after non-empty lines.
+While at it, add empty space before variable value.
+--
+
+lib/memfs: stop all Watcher spawn by DirWatcher::
++
+--
+Previously, the Watcher goroutine will stopped only if the file is
+being deleted.
+If the DirWatcher stopped manually, by calling Stop method, the
+Watcher goroutine will still running in the background.
+
+This changes record all spawned Watcher and stop it when files inside
+a deleted directory or when Stop called.
+--
+
+lib/memfs: remove embedding os.FileInfo and http.File::
++
+The original idea of embedding those interface is to guard the Node
+implementation to always follow all methods in the os.FileInfo and
+http.File, in case there is additional methods in the future.
+But, embedding interfaces make the struct store it as field "FileInfo"
+and File with "<nil>" values, which add storage cost to Node.
+
+lib/memfs: refactoring, simplify DirWatcher logic::
++
+In this changes, we simplify re-scanning the directory content.
+Instead of looping on node Child to detect new or delete files, use map
+to store previous Child first.
+If node path exist in map then add it again as child, otherwise delete it.
+
+email/dkim: set ExpiredAt to MaxInt64 if value is greater than 12 digits::
++
+--
+According to RFC 6376,
+
+  To avoid denial-of-service attacks, implementations MAY consider any
+  value longer than 12 digits to be infinite.
+
+Since ExpiredAt type is uint64, we cannot set it to infinite, so set it
+to maximum value int64 (not maximum of uint64).
+--
+
+
+[#v0_50_1]
+== share v0.50.1 (2023-11-05)
+
+This release bring many enhancements to "lib/memfs", a library for caching
+file system in memory.
+
+[#v0_50_1__enhancements]
+===  Enhancements
+
+lib/memfs: return nil in AddChild if file not exist::
++
+--
+In case the node being added is a symlink, and the symlink is broken,
+the AddChild will return an error "fs.ErrNotExist".
+
+In this changes we check for fs.ErrNotExist and return nil without an
+error, so the user of AddChild can skip adding the node to the parent.
+--
+
+lib/memfs: quote the path in the returned error::
++
+--
+Error message should at least contains 4W in the following order,
+
+  WHEN WHO/WHERE WHAT WHY
+
+By quoting, user or machine can parse that the value inside quote
+is the value that cause the error (the what), not part of the cause of
+the error (the why).
+--
+
+lib/memfs: add method Child to Node::
++
+The Child method return the child node based on its node.
+
+lib/memfs: call the Init method in the embedded file::
++
+--
+By calling the Init method, the regular expression for include and
+excludes are initialized, which allow using any method of MemFS instance
+later, for example the Watch method.
+
+While at it, fix the fields alignment in template code format.
+--
+
+lib/memfs: include empty directory::
++
+--
+Even thought empty directory does not contains file, from the parent
+node _it is_ part of their content.
+
+Also, there is a use case where memfs use as virtual file system (VFS),
+as a layer with file system, where user can view list of directory,
+create a directory or file on the fly.
+If we skip scanning empty directory, that directory will not be visible.
+--
+
+lib/memfs: re-scan directory content on Node's Update::
++
+Previously, only node content get updated.
+In case in the MemFS, user set "Options.TryDirect" to true, the directory
+content should be updated too on "MemFS.Get".
+
+
+[#v0_50_0]
+== share v0.50.0 (2023-10-04)
+
+This release bring many enhancements thanks to linters like revive [1],
+fieldaligment [2], and shadow [3].
+
+This release also replace "math/rand.Seed" with "crypto/rand".
+Since Go 1.20 the "math/rand.Seed" is considered deprecated (the initial
+value of rand is seeded automatically, not zero).
+Now, it is the time to replace "math/rand" with more secure random number
+generator, from "crypto/rand".
+This changes affect tests in package "lib/email", "lib/http", and
+"lib/stmp".
+
+[#v0_50_0__breaking_changes]
+===  Breaking changes
+
+lib/test: refactoring, rename TestWriter to BufferWriter::
++
+--
+The name TestWriter is considered stutter if its called from external
+package, test.TestWriter.
+
+While at it, implement the Error and Errorf in BufferWriter and add
+comment to each exported methods.
+--
+
+ssh/sftp: rename method Close to CloseFile::
++
+Since the method accept FileHandle, that is returned from OpenFile,
+then the method should changed with the same suffix.
+This is also to make the method un-ambiguous later when we add Close method
+to close client connection.
+
+[#v0_50_0__new_features]
+===  New features
+
+lib/errors: implement method Is::
++
+The Is method will return true if the target error is instance of *E
+and the value of field Code and Name match with values in e.
+
+lib/email: add method ID to Header::
++
+The ID method return the Message-ID or empty if its not exist.
+
+test/mock: implement mock for testing io.ReadWriter or io.StringWriter::
++
+--
+The type ReadWriter provide two buffers, BufRead and BufWrite.
+The BufRead is for io.Reader and BufWrite for io.Writer or
+io.StringWriter.
+
+The type provide one method helper, Reset, to reset all buffers.
+--
+
+lib/crypto: add function LoadPrivateKeyInteractive::
++
+--
+The LoadPrivateKeyInteractive load the private key from file.
+If the private key file is encrypted, it will prompt for the passphrase
+from terminal.
+
+This function is taken from package "lib/ssh" with modification by adding
+parameter "io.ReadWriter" and removing parameter maxAttempt.
+--
+
+lib/crypto: implement RSA encrypt and decryption for large message size::
++
+--
+The EncryptOaep extend the "rsa.EncryptOAEP" to make it able to encrypt a
+message larger than its than (public modulus size - 2*hash.Size - 2).
+
+The function signature is the same with "rsa.EncryptOAEP" except the
+name, to make it distinguishable.
+
+The DecryptOaep function extend the "rsa.DecryptOAEP" to make it able to
+decrypt a message larger than its public modulus size.
+--
+
+ssh/sftp: add method to close client connection::
++
+The Close method close the client sftp session and release all its
+resources.
+
+lib/ssh: add method Close to Client::
++
+The Close method close the client connection and release all resources.
+
+
+[#v0_50_0__enhancements]
+===  Enhancements
+
+_AUR: include go-mod-tip.sh to package::
++
+The "go-mod-tip.sh" is a shell script to get and print the latest Go module
+version based on the latest tag and the latest commit hash from current
+working git directory.
+
+lib/crypto: rewrite LoadPrivateKey as wrapper of ssh.ParseRawPrivate::
++
+--
+Previously, the LoadPrivateKey function only able to load private key
+with PKCS#1 format.
+
+This changes make the function as a wrapper for ssh.ParseRawPrivate
+that can load RSA, DSA, ECDSA, and Ed25519 in PKCS#1, PKCS#8, OpenSSL,
+and OpenSSH formats.
+--
+
+lib/ssh: store and use the connected identity file for ScpPut and ScpGet::
++
+Previously, the ScpPut and ScpGet use the first IdentityFile in section
+to use in "-i" argument.
+This is not correct.
+If the section contains two or more IdentityFile and the one that can
+connect are the last one, the ScpPut and ScpGet method may return an
+error when executing the "scp" command.
+
+[1] https://github.com/mgechev/revive
+
+[2] https://pkg.go.dev/golang.org/x/tools/go/analysis/passes/fieldalignment
+
+[3] https://pkg.go.dev/golang.org/x/tools@v0.13.0/go/analysis/passes/shadow
+
+
+[#v0_49_1]
+== share v0.49.1 (2023-09-02)
+
+[#v0_49_1__bug_fixes]
+===  Bug fixes
+
+lib/dns: fix leaking internal zone::
++
+--
+Previously, if the server have internal zone "my.internal" and the client
+query "sub.my.internal" that does not exist in the zone, the server then
+forward the query to parent name server.
+This cause the internal zone and its domains leaked to parent name server.
+
+This changes fix this issue by checking if the query is subset of
+internal zone Origin if domain does not exist, and response with error
+code 3 (ERR_NAME) with the Zone SOA in Authority.
+--
+
+[#v0_49_1__enhancements]
+=== Enhancements
+
+lib/dns: update the SOA Serial when record added or removed from Zone::
++
+Any call to Zone Add or Remove methods will update the Zone.SOA.Serial to
+current epoch.
+
+lib/dns: add method AddAuthority to Message::
++
+--
+The AddAuthority add the rr to list of Authority.
+Calling this method mark the message as answer, instead of query.
+
+If the rr is SOA, it will replace the existing record if exist and set
+the flag authoritative answer (IsAA) in header to true.
+If the rr is NS, it will be added only if its not exist.
+
+It will return an error if the rr type is not SOA or NS or the size of
+records in Authority is full, maximum four records.
+--
+
+lib/dns: add method to populate internal caches by Zone::
++
+The InternalPopulateZone populate the internal caches from Zone's
+messages.
+
+ssh/config: handle key UserKnownHostsFile::
++
+The UserKnownHostsFile define list of the known_hosts files to be read,
+separated by spaces.
+If not set default to "~/.ssh/known_hosts" and "~/.ssh/known_hosts2".
+
+
+lib/ssh: use UserKnownHostFile from configuration in NewClientInteractive::
++
+--
+Previously, the ssh Client always use InsecureIgnoreHostKey in
+HostKeyCallback.
+This may post security issue, like man-in-the-middle attack, since we
+did not check the server host key with one of key that known by client
+from UserKnownHostFile (for example ~/.ssh/known_hosts).
+
+This changes use the SSH section UserKnownHostFile from configuration
+(default to ~/.ssh/known_hosts) to check if the server host key is
+valid.
+The NewClientInteractive will return an error, "key is unknown", if host
+key not exist in UserKnownHostFile or "key is mismatch" if host key
+not match with one registered in UserKnownHostFile.
+
+This changes depends on patch of golang.org/x/crypto [1] that has not
+reviewed yet, so we need to replace it with one that contains the patch.
+
+[1] https://go-review.googlesource.com/c/crypto/+/523555
+--
+
+
+[#v0_49_0]
+== share v0.49.0 (2023-08-04)
+
+[#v0_49_0__breaking_changes]
+=== Breaking changes
+
+lib/email: refactoring, replace field with type []byte to string::
++
+Using string provide safety, guaranteed that if we pass it as parameter
+the receiver will not be able to modify its content.
+
+ssh/config: refactoring, simplify the Section fields::
++
+--
+Instead of storing each Section value in separate field, store them
+inside a map, Field.
+This reduce the size of Section and simplify adding or getting the
+key that we are not supported but maybe usable by user in the future.
+
+This changes introduce several new methods as replacement of field:
+
+* CASignatureAlgorithms: a method that return list of signature
+  algorithms that Section set or the default
+* CanonicalDomains: a method that return CanonicalDomains set in Section
+* CanonicalizePermittedCNames: return the permitted CNAMEs set in Section,
+  from KeyCanonicalizePermittedCNames.
+* CertificateFile: return list of certificate file
+* Environments: return system and/or custom environment that will be
+  passed to remote machine.
+  The key and value is derived from "SendEnv" and "SetEnv".
+* FieldBool: return field value as boolean
+* FieldInt: return the field value as int
+* Hostname: return the Hostname in this Section
+* IdentityAgent: return the path to SSH agent socket to be used
+* Port: return the remote machine port
+* User: return the remote user name
+* Set: set the Field using key and value
+--
+
+lib/ssh: refactoring NewClientFromConfig, renamed to NewClientInteractive::
++
+--
+Previously, the NewClientInteractive blindly use the signers from
+Section.Signers.
+If one of the IdentityFile valid, it will add all the keys in
+IdentityFile to SSH agent.
+
+In this changes we try each IdentityFile independently.
+If the key is valid, client connected to remote machine, then only that
+key will be added to SSH agent.
+
+While at it we also rename the method to NewClientInteractive to
+indicate that the function will prompt for passphrase if one of the
+IdentityFile is encrypted.
+--
+
+[#v0_49_0__bug_fixes]
+=== Bug fixes
+
+lib/http: fix missing query when handling redirect in HandleFS::
++
+In 06e6cbdd511c, we redirect request by adding end slash to the path
+if the requested resource is directory, but somehow we miss adding the
+original request query.
+This changes fix this issue.
+
+lib/os: check for extract path in untar and unzip::
++
+--
+Once we Join-ed the directory output with the file name, we check if
+the result of join is still under directory output, if its not, return
+an error to prevent
+https://cwe.mitre.org/data/definitions/22.html[Zip Slip vulnerability].
+--
+
+lib/smtp: format the passed data in NewMailTx::
++
+--
+The following rules are applied to the data,
+
+* all lines must end with CRLF
+* if the line start with period, additional period is inserted before
+  the line. This recommendation based on RFC 5321 section 4.5.2 [1] to
+  prevent data that contains CRLF "." CRLF does not corrupt the message,
+  causing the server terminate reading the message where it should not.
+
+[1] https://datatracker.ietf.org/doc/html/rfc5321#section-4.5.2
+--
+
+[#v0_49_0__new_features]
+=== New features
+
+lib/os: add function PathFold and PathUnfold::
++
+--
+The PathFold replace the path "in" with tilde "~" if its prefix match
+with user's home directory from os.UserHomeDir.
+
+The PathUnfold expand the tilde "~/" prefix into user's home directory
+using os.UserHomeDir and environment variables using os.ExpandEnv
+inside the string path "in".
+--
+
+lib/os: add function Environments::
++
+The Environments function return list of system environment as map of
+key and value.
+
+lib/ssh: add function LoadPrivateKeyInteractive::
++
+The LoadPrivateKeyInteractive load private key from file.
+If key is encrypted, it will prompt the passphrase in terminal with
+maximum maxAttempt times.
+If the passphrase still invalid after maxAttempt it will return an error.
+
+[#v0_49_0__enhancements]
+=== Enhancements
+
+lib/smtp: set minimum Server TLS to v1.2::
++
+Using the TLS v1.1 is considered insecure and should not be used in
+server anymore.
+
+lib/memfs: check for refresh URL outside of Root SysPath::
++
+The case when refresh URL outside of Root SysPath is only exist when
+the memfs TryDirect is set to true, usually during development.
+In the production, the TryDirect should be false, hence the refresh
+always return nil Node.
+
+ssh/config: reorder struct fields for better alignment::
++
+--
+Changes,
+
+* Config: changes allocated size from 32 to 16 bytes (-8 bytes)
+* parser: changes allocated size from 40 to 32 bytes (-8 bytes)
+* Section: changes allocated size from 392 to 360 bytes (-32 bytes)
+
+The rest of changes that are not mentioned are from test files.
+--
+
+
+[#v0_48_0]
+==  share v0.48.0 (2023-07-07)
+
+This release bring many enhancements to lib/websocket including timeout,
+handling upgrade and read/write concurrently using goroutine.
+
+[#v0_48_0__breaking_changes]
+=== Breaking changes
+
+lib/net: changes the WaitRead/Event model on Poll::
++
+--
+Previously, the Pool's WaitRead and WaitReadEVent methods return list of
+file descriptor (fd) and keeps the fd in the pool.
+In case we want to process the returned fd concurrently, by running it
+in different goroutine, the next call WaitRead may return the same fd
+if its goroutine not fast enough to read from fd.
+
+This changes fix this issue by removing list of fd from poll and set the
+fd flag to blocking mode again after returning it from WaitRead or
+WaitReadEvent.
+
+This changes also remove the ReregisterRead and ReregisterEvent methods
+since it is not applicable anymore.
+--
+
+[#v0_48_0__bug_fixes]
+=== Bug fixes
+
+lib/websocket: call Quit when handshake contains close or invalid frame::
++
+If the HTTP handshake response contains trailing frame, handle it
+directly.
+If the frame is invalid or contains control close operation, call Quit
+directly to trigger the HandleQuit if its defined by user.
+
+lib/websocket: revert maxBuffer back to 1024::
++
+In v0.47.0 we increase the maxBuffer to 4096 to try increasing the
+performance when handling large payload.
+Turns out increasing this break the autobahn test suite.
+
+
+[#v0_48_0__new_features]
+=== New features
+
+lib/ascii: add type Set::
++
+--
+The Set type is a bitmap that represent list of ASCII characters for faster
+lookup.
+
+A Set is a 36-byte value, where each bit in the first 32-bytes represents
+the presence of a given ASCII character in the set.
+The remaining 4-bytes is a counter for the number of ASCII characters in the
+set.
+The 128-bits of the first 16 bytes, starting with the least-significant bit of
+the lowest word to the most-significant bit of the highest word, map to the
+full range of all 128 ASCII characters.
+The 128-bits of the next 16 bytes will be zeroed, ensuring that any non-ASCII
+character will be reported as not in the set.
+--
+
+lib/net: implement generic PollEvent::
++
+--
+The PollEvent contains file descriptor and the underlying event
+based on OS, unix.EpollEvent on Linux or unix.Kevent_t on BSD.
+
+The Poll interface provides two APIs to works with PollEvent,
+WaitReadEvents that return list of PollEvent ready for read, and
+ReregisterEvent to register the event back to poll (only for Linux).
+--
+
+[#v0_48_0__enhancements]
+=== Enhancements
+
+lib/websocket: add option to set read/write timeout on Server::
++
+--
+The ReadWriteTimeout define the maximum duration the server wait when
+receiving/sending packet from/to client before considering the
+connection as broken.
+
+Default read-write timeout is 30 seconds if not set.
+
+This changes affect the exported function Send and Recv by adding
+additional parameter timeout to both of them.
+--
+
+lib/websocket: handle concurrent upgrade using goroutine::
++
+--
+The maxGoroutineUpgrader define maximum goroutines running at the same
+time to handle client upgrade.
+The new goroutine only dispatched when others are full, so it will
+run incrementally not all at once.
+Default to defServerMaxGoroutineUpgrader (128) if its not set.
+--
+
+lib/websocket: handle concurrent Server read using goroutines::
++
+The Server now dispatch a goroutine to consume event from poll reader
+for each client connection that is ready to read.
+The maximum number of goroutine is defined in ServerOptions
+maxGoroutineReader, which currently set to 1024.
+
+lib/websocket: handle concurrent ping using goroutines::
++
+The maximum goroutines is quarter of max queue.
+The new goroutine for pinger will be dispatched when no goroutine can
+consume the current processed connection.
+
+
+[#v0_48_0__chores]
+===  Chores
+
+websocket/testdata: rewrite autobahn test using container::
++
+--
+Since the autobahn script can only run on Python 2, it become hard to
+setup and run the test on distro that does not provide Python 2 anymore.
+The autobahn repository recommend to use docker instead.
+
+When testing the server, we simplify it by using make task "test-server".
+The test-server task run our test server in background, and then run the
+autobahn fuzzingclient from container.
+Once the tests completed, we trigger the server to shutdown by sending
+text frame with payload "shutdown".
+
+When testing the client, we simplify it by using make task "test-client".
+The test-client task run the autobahn fuzzingserver and then
+we run our client.
+Once client finished, we trigger the server to generate the reports
+and cleanup the container.
+--
+
+
+[#v0_47_0]
+==  share v0.47.0 (2023-06-05)
+
+[#v0_47_0__breaking_changes]
+===  Breaking changes
+
+email/maildir: major refactoring plus adding unit tests::
++
+This changes remove all unneeded methods from Manager leave it with
+four methods: Delete, FetchNew, Incoming, OutgoingQueue.
++
+Also, we add the type filename to generate file name for tmp and new
+directory.
+
+lib/email: unexport the field ContentType in the field::
++
+The field ContentType will be set only when the field Name is
+"Content-Type" so it's not always exist on each field.
+To get the field ContentType, use Header.ContentType().
+
+
+[#v0_47_0__bug_fixes]
+=== Bug fixes
+
+lib/dns: fix zone parsing on SOA record with single line::
++
+Due to refactoring in c376eccd25, parsing SOA record with single line
+return an error: "parseSOA: line 2: incomplete SOA statement '0'".
+
+lib/memfs: ignore permission error when scanning directory content::
++
+Instead of returning error, skip the directory that we cannot read and
+continue to process the other.
+
+lib/memfs: fix panic when watched file deleted or renamed::
++
+When the file being watched is deleted, sometimes it will cause panic.
+
+lib/email: fix parsing multiple parameters in ContentType::
++
+While at it, also fix the ContentType String method to prefix ';' before
+appending parameter key and value.
+
+
+[#v0_47_0__new_features]
+=== New features
+
+cmd/bcrypt: CLI to compare or generate hash using bcrypt::
++
+The bcrypt command has two subcommand "compare" and "gen".
+The "compare" subcommand accept two parameter the hash and plain text.
+The "gen" subcommand accept only one parameter, the plain text to be hashed.
+
+lib/sql: add type DmlKind::
++
+The DmlKind define the kind for Data Manipulation Language.
+
+email/maildir: implement Folder::
++
+Folder is a directory under maildir that store messages per file.
+A folder contains three directories: tmp, new, and cur; and an empty
+file "maildirfolder".
+
+lib/net: add function WaitAlive::
++
+--
+WaitAlive try to connect to network at address until timeout reached.
+If connection cannot established it will return an error.
+
+Unlike [net.DialTimeout], this function will retry not returning an error
+immediately if the address has not ready yet.
+--
+
+lib/smtp: implement Client SendEmail::
++
+--
+Somehow in 3a1a2715b25f, we include this method without implementing it.
+
+The SendEmail method simplify sending email by automatically create
+[MailTx] for passing it to method Client.MailTx.
+
+The test right now use live connection since the Server is not ready yet.
+--
+
+[#v0_47_0__enhancements]
+===  Enhancements
+
+lib/dns: add option to set debug level in ServerOptions::
++
+This options replace the global debug package.
+
+lib/dns: do not cache empty answers::
++
+The use case if one use and switch between two different
+networks with internal zone, frequently.
+For example, if on network Y they have domain MY.Y and
+current connection is X, request to MY.Y will return an
+empty answers.
+Once they connect to Y again, any request to MY.Y will not
+be possible because rescached caches contains empty answer
+for MY.Y.
+
+_bin/go-test-coverhtml: add parameter to run specific test::
++
+The second parameter is optional.
+It is passed to -run= argument in "go test".
+Default value is ".", or all functions.
+
+lib/http: redirect path with slash if request is directory::
++
+--
+If request path is a directory and it is not end with slash, redirect
+request to location with slash to allow relative links works inside the
+HTML content.
+
+For example, a "/page/index.html" contains links href="sub.html" (where
+"sub.html" is inside "/page" directory).
+If request to "/page" (without end with slash) return content of
+"/page/index.html", then when user click on sub.html it will request to
+"/sub.html" instead of "/page/sub.html".
+--
+
+lib/email: handle obsolete white spaces and comment when unpacking date::
++
+--
+In the obsolete syntax, white space and comments can appear between many
+more element, for example the folloing Date value are valid
+
+	Date  : Fri, 21 Nov 1997 09(comment):   55  :  06 -0600
+
+This changes handle this by sanitizing the Field value, removing comment
+and merge multiple spaces into one, before parsing it.
+--
+
+lib/email: set the Field Type and unpack its value on ParseField::
++
+--
+Once the field Name has detected and its Value is valid, we can unpack
+the Value based to type that it represent, for example to Date or Mailbox.
+
+This changes remove calling to unpack directly in some tests and check
+an error when testing ParseHeader.
+--
+
+lib/net: increase the maximum poll events::
++
+The maxQueue define the number of events that can be read from poll at
+one time.
+Using 128 seems to small for high throughput networks.
+Increasing this number also increase the memory consumed by process.
+Maybe later we can export this function as option when creating poll.
+
+lib/websocket: increase the max buffer and queue for better throughput::
++
+--
+The maxBuffer increased from 1024 to 4096 bytes.
+The reason that we use 1024 previously is related to MTU size and maximum
+payload in TCP (although its higher, 1460 bytes).
+
+The maxQueue increase from 128 to 4096.
+--
+
+[#v0_47_0__chores]
+=== Chores
+
+all: remove any usage of debug.Value in all packages::
++
+Using global debug value for all packages turns out is not a good
+idea.
+
+lib/test: update documentation related to Assert and Data::
++
+The documentation is based on the article published at
+https://kilabit.local/journal/2023/go_test_data/
+after reviewing and explain how to use both of them to public.
+
+all: record the contributors of this module in file AUTHORS::
+
+
+[#v0_46_0]
+==  share v0.46.0 (2023-05-02)
+
+This release deprecated lib/io and lib/parser.
+
+
+[#v0_46_0__breaking_changes]
+===  Breaking changes
+
+lib/reflect: remove the third return value from Marshal::
++
+The third, boolean, return value is redundant with the second error value.
+
+lib/bytes: changes the DumpPrettyTable output format::
++
+The change is to accommodate large bytes data, more than 0xFFFF.
+The hex address in the first column is increased to 8 digits, the
+characters compacted without space in between.
+
+
+[#v0_46_0__new_features]
+===  New features
+
+lib/os: merge some functions from lib/io::
++
+Functions like Copy, IsBinary, IsDirEmpty, IsFileExist, RmdirEmptyAll
+are read and operate on file and directory on operating system level, so
+it is not correct to put it in package io.
+
+lib/strings: merge lib/parser here::
++
+--
+The first idea of parser is to provide generic parser for both bytes and
+string.
+After we introduce lib/parser there is not much changes to that package.
+Also, since we create another Parser in lib/bytes that accept and
+return token as []byte, the lib/parser is not unique anymore.
+
+The following function/methods changes to minimize conflict in the future,
+
+* Lines become LinesOfFile
+* New become NewParser
+* Open become OpenForParser
+* Token become Read
+* TokenEscaped become ReadEscaped
+* TokenTrimSpace become ReadNoSpace
+--
+
+lib/bytes: implement function ParseHexDump::
++
+--
+The ParseHexDump parse the default output of [hexdump](1) utility from
+parameter in back into stream of byte.
+
+An example of default output of hexdump is
+
+	0000000 7865 5f70 6964 2f72 0000 0000 0000 0000
+	0000010 0000 0000 0000 0000 0000 0000 0000 0000
+	*
+	0000060 0000 0000 3030 3030 3537 0035 3030 3130
+
+The first column is the address and the rest of the column is the data.
+Each data column is 16-bit words in big-endian order, so in the above
+example, the first byte would be 65, second byte is 78 and so on.
+The asterisk "*" means that the address from 0000020 to 0000050 is equal to
+the previous line, 0000010.
+
+[hexdump]: https://man.archlinux.org/man/hexdump.1
+--
+
+lib/bytes: implement tokenize Parser::
++
+The Parser type parse stream of byte using one or more delimiters as
+separator between token.
+
+lib/bytes: add function TrimNull::
++
+The TrimNull function remove 0 value ("\0" or NULL in C) at leading
+and trailing of input.
+
+lib/net: add method WriteTo to ResolvConf::
++
+The WriteTo method write the ResolvConf as text.
+
+
+[#v0_46_0__enhancements]
+===  Enhancements
+
+lib/time: calculate the next event before notify the user on Scheduler run::
++
+This allow user to call the Next method, to know the next time the
+scheduler will be triggered, after receiving the event.
+
+lib/reflect: add option to skip processing struct field in Do/IsEqual::
++
+A struct's field tagged with `noequal:""`, its value will not be processed for
+equality.
+
+
+[#v0_46_0__chores]
+===  Chores
+
+lib/reflect: use doEqual inside IsEqual::
++
+Previously, IsEqual internally use isEqual, which have the same logic as
+doEqual.
+This changes minimize duplicate code between IsEqual and DoEqual, by
+calling doEqual for both of functions.
+
+lib/time: replace lib/io#Reader with lib/bytes#Parser::
+
+lib/smtp: replace lib/io#Reader with lib/bytes#Parser::
+
+lib/dns: replace lib/io#Reader with lib/bytes#Parser::
+
+lib/http: replace lib/io#Reader with lib/bytes#Parser::
+
+lib/email: replace lib/io#Reader with lib/bytes#Parser::
+
+email/dkim: replace lib/io#Reader with lib/bytes#Parser::
+
+lib/hunspell: replace lib/io with lib/os::
+
+lib/hunspell: replace lib/parser with lib/strings::
+
+lib/http: replace lib/parser with lib/strings::
+
+lib/bytes: copy TokenFind to internal/bytes#TokenFind::
++
+This is to prevent import cycle later when we use lib/test in bytes.
+
+
+[#v0_45_0]
+==  share v0.45.0 (2023-04-01)
+
+This release set the Go version to 1.19.
+
+[#v0_45_0__breaking_changes]
+===  Breaking changes
+
+lib/net: changes the PopulateQuery logic::
++
+--
+Previously, PopulateQuery only add the passed dname if the number of dots
+is greater than 0.
+After inspecting the result from dig and getenv, the dots seems does not
+affect the query.
+For example, if we have A record for domain "kilabit", both of those tools
+query name "kilabit" without adding local domain or domain in search.
+--
+
+[#v0_45_0__new_features]
+===  New features
+
+_bin: add shell script go-test-lint.sh to run test and lint sequentially::
++
+--
+The go-test-lint.sh run Go test and if its success it will run
+predefined linter, in the current directory.
+
+Arg 1: the method or function to test, default to ".".
+
+The linter program and its argument is derived from environment variable
+GO_LINT.
+If its empty, it will try the following linter in order: revive and then
+golangci-lint.
+
+To add additional arguments to Go test set the environment variable
+GO_TEST_ARGS.
+--
+
+lib/bytes: add function DumpPrettyTable::
++
+The DumpPrettyTable write each byte in slice data as hexadecimal, ASCII
+character, and integer with 8 columns width.
+
+lib/bytes: add function SplitEach::
++
+The SplitEach funciton split the slice of byte into n number of bytes.
+If n is less or equal than zero, it will return the data as chunks.
+
+lib/dns: add function ParseZone::
++
+--
+The ParseZone parse the content of zone from raw bytes.
+
+Now that we have ParseZone, all tests that use zoneParser now can be
+replaced using combination of test.Data and ParseZone.
+--
+
+lib/dns: add method WriteTo to Zone::
++
+--
+The WriteTo method write the zone as text into io.Writer.
+
+The result of WriteTo will be different with original content of zone
+file, since it does not preserve comment and indentation.
+--
+
+lib/http: add function to parse multipart Range response for Client::
++
+The ParseMultipartRange parse the multipart/byteranges body or response
+from HTTP Range request.
+Each Content-Range position and body part in the multipart will be stored
+under RangePosition.
+
+lib/http: add support for HTTP Range in Server::
++
+--
+For HTTP Server using HandleFS, the Range request is handled
+automatically.
+For other HTTP server, user can use the HandleRange function.
+
+The HandleRange function handle
+https://developer.mozilla.org/en-US/docs/Web/HTTP/Range_requests[HTTP Range]
+request using "bytes" unit.
+The body parameter contains the content of resource being requested that
+accept Seek method.
+
+If the Request method is not GET, or no Range in header request it will
+return all the body
+https://datatracker.ietf.org/doc/html/rfc7233#section-3.1[RFC7233 S-3.1].
+
+The contentType is optional, if its empty, it will detected by
+http.ResponseWriter during Write.
+--
+
+lib/io: add method ReplaceAll on Reader::
++
+The ReplaceAll method behave like standard bytes.ReplaceAll but start
+from current index.
+
+lib/parser: add method TokenTrimSpace::
++
+The TokenTrimSpace read the next token until one of the delimiter found,
+with leading and trailing spaces are ignored.
+
+lib/parser: add method SetDelimiters::
++
+The SetDelimiters replace the current delimiters.
+
+lib/telemetry: package for collecting and forwarding metrics::
++
+Package telemetry is a library for collecting various Metric, for example
+from standard runtime/metrics, and send or write it to one or more
+Forwarder.
+Each Forwarder has capability to format the Metric before sending or
+writing it using Formatter.
+
+
+[#v0_45_0__bug_Fixes]
+===  Bug fixes
+
+lib/dns: fix packing, parsing, and saving MINFO resource data::
++
+--
+Even thought the MINFO record not formally obsolete, according to
+https://en.wikipedia.org/wiki/List_of_DNS_record_types#Obsolete_record_types[Wikipedia],
+we still need to support this for backward compatibility.
+
+When packing the resource data length does not include total length.
+When parsing, the RMailBox and EmailBox should be added the origin suffix
+if its not end with dot.
+When saving, the origin should be trimmed from RMailBox and EmailBox.
+--
+
+lib/dns: fix packing and unpacking resource record HINFO::
++
+--
+The rdata for HINFO contains two character-strings: CPU and OS.
+Previously, we pack the rdata section sequentially, without adding length
+on each of them: <RDLEN><CPU><OS>.
+The correct pack format should <RDLEN><LENGTH><CPU><LENGTH><OS>.
+--
+
+lib/dns: fix parsing SRV record from zone file::
++
+--
+Previous parseSRV start by parsing the _Service from tok, but the
+actual value of parameter tok is the Priority.
+
+This changes fix this and as testing we use the example from RFC 2782.
+--
+
+[#v0_45_0__enhancements]
+===  Enhancements
+
+lib/dns: allow parsing TXT rdata without quote in zone file::
++
+--
+Previously, the zone only parsing TXT record with double quote since
+most of the example that we found during implementation all use double
+quote.
+
+This changes allow non-double quoted text in zone file with consequence
+that any spaces will terminated the rdata immediately.
+
+Fixes #6
+--
+
+lib/dns: handle zone file with CRLF line ending::
++
+--
+While at it, fix parsing multiline SOA record where closing parentheses
+end on next lines.
+
+Fixes #6
+--
+
+
+
+
+lib/test: simplify the string diff output from Assert::
++
+--
+In the output, instead of using %q we replace it with %s, because printing
+string with double quote cause escaping and hard to read
+This change may cause difference in white spaces not showed in the
+terminal.
+
+In the diff changes, only print the Old and New, without printing each
+chunk.
+--
+
+
+[#v0_44_0]
+==  share v0.44.0 (2023-03-02)
+
+[#v0_44_0__new_features]
+===  New features
+
+lib/time: implement Scheduler::
++
+--
+Scheduler is a timer that run periodically based on calendar or day time.
+
+A schedule is divided into monthly, weekly, daily, hourly, and minutely.
+An empty schedule is equal to minutely, a schedule that run every minute.
+--
+
+lib/time: add new type Clock::
++
+Clock represent 24 hours time with hour, minute, and second.
+An hour value is from 0 to 23, a minute value is from 0 to 59, and
+a second value is from 0 to 59.
+
+[#v0_44_0__bug_fixes]
+===  Bug fixes
+
+lib/clise: fix potential data race between Push and Slice::
++
+The data race may occur if Push is called, the .last field is incremented
+and at the same time an other goroutine call Slice that access the .last
+field.
+
+lib/memfs: minimize data race on DirWatcher::
++
+Calling DirWatcher Stop while the start method set dw.ticker can cause
+data race.  This changes fix this issue.
+
+go.mod: update all dependencies::
++
+This update use "go get all" which resolve to semver for each dependencies.
+
+[#v0_44_0__chores]
+=== Chores
+
+all: set the test timeout to 1m::
++
+Some flaky test, caused by waiting for channel, require waiting for 10m
+before it considered fail.
+This changes we cut the default timeout to 1 minute.
+
+all: access embedded field or methods using the type name::
++
+This is for clarity in the code, better to be explicit by typing where
+the field or methods come from.
+
+lib/email: add an example for Filter method on Header::
+
+cmd: temporarily hide unused commands::
++
+The cart, cascaded-random-forest, lnsmote, random-forest, and smote
+are part of completing thesis and they never used anymore.
+
+AUR: add go-test-coverhtml.sh and go-bench.sh into package::
++
+While at it, changes the go-bench count to 10.
+
+lib/websocket: replace math/rand.Read with crypto/rand.Read::
++
+The math/rand will be deprecated in Go 1.20.
+
+
+[#v0_43_0]
+==  share v0.43.0 (2023-02-07)
+
+[#v0_43_0__new_features]
+===  New features
+
+lib/http: add function MarshalForm::
++
+--
+The MarshalForm marshal struct fields tagged with `form:` into url.Values.
+
+The rules for marshaling follow the same rules as in [UnmarshalForm].
+
+It will return an error if the input is not pointer to or a struct.
+--
+
+clise: implement io Closer, Writer, StringWriter, and ByteWriter::
+
+clise: add method UnmarshalJSON::
++
+The UnmarshalJSON convert JSON array into Clise.
diff --git a/_doc/index.adoc b/_doc/index.adoc
index 794f2755..39bfa6bd 100644
--- a/_doc/index.adoc
+++ b/_doc/index.adoc
@@ -41,6 +41,9 @@ This library is released every month, usually at the first week of month.
 
 link:CHANGELOG.html[Latest changelog^].
 
+link:CHANGELOG_2023.html[Changelog in 2023^].
+This is changelog for share module since v0.43.0 until v0.51.0.
+
 link:CHANGELOG_2022.html[Changelog in 2022^].
 This is changelog for share module since v0.33.0 until v0.42.0.