Revert "Prepare release"

This reverts commit dc57404252c47520f352834ad9be45ad684f96c9.

1 files changed, 0 insertions, 475 deletions

diff --git a/erts/doc/src/notes.xml b/erts/doc/src/notes.xml
index 628df1b9bd..acc19f250a 100644
--- a/erts/doc/src/notes.xml
+++ b/erts/doc/src/notes.xml
@@ -32,481 +32,6 @@
   <p>This document describes the changes made to the ERTS application.</p>
 
 
-<section><title>Erts 9.0</title>
-
-    <section><title>Fixed Bugs and Malfunctions</title>
-      <list>
-        <item>
-	    <p>Fix various bugs regarding loading, upgrade and purge
-	    of HiPE compiled code:</p> <list> <item>The native code
-	    memory for a purged module was never deallocated.</item>
-	    <item>Wrong functions could in some cases be called after
-	    a module upgrade.</item>
-	    <item><c>erlang:check_process_code</c> did not check for
-	    recursive calls made from native code.</item> </list>
-          <p>
-	    Own Id: OTP-13968</p>
-        </item>
-        <item>
-          <p>
-	    Hipe optional LLVM backend does require LLVM version 3.9
-	    or later as older versions forced strong dependencies on
-	    erts internals structures.</p>
-          <p>
-	    Own Id: OTP-14238</p>
-        </item>
-        <item>
-          <p>
-	    Slogans in crash dumps have been extended to print more
-	    complex terms.</p>
-          <p>
-	    Own Id: OTP-14303</p>
-        </item>
-        <item>
-          <p>
-	    Fixed bug when using <c>enif_inspect_binary</c> in
-	    combination with <c>enif_copy</c>. In some circumstances
-	    the inspected binary could be realloced by the
-	    <c>enif_copy</c> call when it shouldn't have been.</p>
-          <p>
-	    Own Id: OTP-14304</p>
-        </item>
-        <item>
-          <p>
-	    Use <c>-fno-PIE</c> for Gentoo Hardened and others that
-	    don't accept linker flag <c>-no-pie</c>.</p>
-          <p>
-	    Own Id: OTP-14307 Aux Id: PR-1379 </p>
-        </item>
-        <item>
-          <p>
-	    Disable hipe for <c>ppc64le</c> architecture (little
-	    endian) as it is not, and has never been, supported. It
-	    was earlier equated with <c>ppc64</c> (big endian) which
-	    lead to broken build without <c>--disable-hipe</c>.</p>
-          <p>
-	    Own Id: OTP-14314 Aux Id: ERL-369, PR-1394 </p>
-        </item>
-        <item>
-          <p>
-	    Fix 'epmd -kill' to return a failure exit status code if
-	    epmd was not killed because of some error.</p>
-          <p>
-	    Own Id: OTP-14324</p>
-        </item>
-        <item>
-	    <p>Fixed the following dirty scheduler related bugs:</p>
-	    <list> <item><p>the <c>+SDPcpu</c> command line argument
-	    could cause the amount of dirty CPU schedulers to be set
-	    to zero</p></item>
-	    <item><p><c>erlang:system_flag(multi_scheduling, _)</c>
-	    failed when only one normal scheduler was used together
-	    with dirty scheduler support</p></item> </list>
-          <p>
-	    Own Id: OTP-14335</p>
-        </item>
-        <item>
-          <p>
-	    Fix erlexec to handle mismatch in sysconf and proc fs
-	    when figuring out the cpu topology. This behaviour has
-	    been seen when using docker together with
-	    <c>--cpuset-cpus</c>.</p>
-          <p>
-	    Own Id: OTP-14352</p>
-        </item>
-        <item>
-          <p>
-	    Fixed memory segment cache used for multiblock carriers.
-	    Huge (> 2GB) memory segments could cause a VM crash.
-	    Creation of such huge memory segments used for multiblock
-	    carriers is however very uncommon.</p>
-          <p>
-	    Own Id: OTP-14360 Aux Id: ERL-401, PR-1417 </p>
-        </item>
-      </list>
-    </section>
-
-
-    <section><title>Improvements and New Features</title>
-      <list>
-        <item>
-          <p>
-	    Added <c>erlang:garbage_collect/2</c> that takes an
-	    option list as the last argument that can be used to
-	    control whether a minor or a major garbage collection is
-	    to be done. Doing a minor collection only collects terms
-	    that have recently died, but is cheaper than a mjor
-	    collection.</p>
-          <p>
-	    Own Id: OTP-11695</p>
-        </item>
-        <item>
-          <p>
-	    Optimized test for tuples with an atom as first element.</p>
-          <p>
-	    Own Id: OTP-12148</p>
-        </item>
-        <item>
-          <p>
-	    Erlang literals are no longer copied during process to
-	    process messaging.</p>
-          <p>
-	    Own Id: OTP-13529</p>
-        </item>
-        <item>
-	    <p>Add support in the <c>erl_nif</c> API for asynchronous
-	    message notifications when sockets or other file
-	    descriptors are ready to accept read or write operations.
-	    The following functions have been added:</p> <list>
-	    <item><p>enif_select</p></item>
-	    <item><p>enif_monitor_process</p></item>
-	    <item><p>enif_demonitor_process</p></item>
-	    <item><p>enif_compare_monitors</p></item>
-	    <item><p>enif_open_resource_type_x</p></item> </list>
-          <p>
-	    Own Id: OTP-13684</p>
-        </item>
-        <item>
-	    <p>There are two new guard BIFs '<c>floor/1</c>' and
-	    '<c>ceil/1</c>'. They both return integers. In the
-	    '<c>math</c>' module, there are two new BIFs with the
-	    same names that return floating point values.</p>
-          <p>
-	    Own Id: OTP-13692</p>
-        </item>
-        <item>
-          <p>
-	    Remove deprecated <c>erlang:hash/2</c>.</p>
-          <p>
-	    *** POTENTIAL INCOMPATIBILITY ***</p>
-          <p>
-	    Own Id: OTP-13827</p>
-        </item>
-        <item>
-	    <p>Replaced usage of deprecated symbolic <seealso
-	    marker="erts:erlang#type-time_unit"><c>time
-	    unit</c></seealso> representations.</p>
-          <p>
-	    Own Id: OTP-13831 Aux Id: OTP-13735 </p>
-        </item>
-        <item>
-          <p>
-	    Added support in zlib for extraction of the inflation
-	    dictionary.</p>
-          <p>
-	    Own Id: OTP-13842</p>
-        </item>
-        <item>
-          <p>
-	    The previously used purge strategy has been removed. The
-	    optional purge strategy introduced in ERTS version 8.1 is
-	    now the only strategy available.</p>
-          <p>
-	    The new purge strategy is slightly incompatible with the
-	    old strategy. Previously processes holding <c>fun</c>s
-	    that referred to the module being purged either failed a
-	    soft purge, or was killed during a hard purge. The new
-	    strategy completely ignores <c>fun</c>s. If <c>fun</c>s
-	    referring to the code being purged exist, and are used
-	    after a purge, an exception will be raised upon usage.
-	    That is, the behavior will be exactly the same as the
-	    case when a <c>fun</c> is received by a process after the
-	    purge.</p>
-          <p>
-	    For more information see the documentation of <seealso
-	    marker="erts:erlang#check_process_code/3"><c>erlang:check_process_code/3</c></seealso>.</p>
-          <p>
-	    *** POTENTIAL INCOMPATIBILITY ***</p>
-          <p>
-	    Own Id: OTP-13844 Aux Id: OTP-13833 </p>
-        </item>
-        <item>
-          <p>
-	    Dirty schedulers are now enabled by default when the
-	    runtime system is built with SMP support.</p>
-          <p>
-	    Own Id: OTP-13860</p>
-        </item>
-        <item>
-          <p>
-	    Improved ETS lookup/insert/delete speed for large
-	    <c>set</c>, <c>bag</c> and <c>duplicate_bag</c> by a
-	    significant reduction of the hash load factor. This speed
-	    improvement comes at the expense of less than one word
-	    per table entry. Tables with less than 256 entries are
-	    not affected at all.</p>
-          <p>
-	    Own Id: OTP-13903</p>
-        </item>
-        <item>
-          <p>
-	    The NIF library <c>reload</c> feature is not supported
-	    anymore. It has been marked as deprecated since OTP R15B.
-	    This means that you are only allowed to do one successful
-	    call to <c>erlang:load_nif/2</c> for each module
-	    instance. A second call to <c>erlang:load_nif/2</c> will
-	    return <c>{error, {reload, _}}</c> even if the NIF
-	    library imlements the <c>reload</c> callback.</p>
-          <p>
-	    Runtime upgrade of a NIF library is still supported by
-	    using the Erlang module upgrade mechanics with a current
-	    and an old module instance existing at the same time with
-	    their corresponding NIF libraries.</p>
-          <p>
-	    *** POTENTIAL INCOMPATIBILITY ***</p>
-          <p>
-	    Own Id: OTP-13908</p>
-        </item>
-        <item>
-          <p>
-	    Add <c>erlang:system_info(atom_count)</c> and
-	    <c>erlang:system_info(atom_limit)</c> to provide a way to
-	    retrieve the current and maximum number of atoms.</p>
-          <p>
-	    Own Id: OTP-13976</p>
-        </item>
-        <item>
-	    <p>The function <c>fmod/2</c> has been added to the
-	    <c>math</c> module.</p>
-          <p>
-	    Own Id: OTP-14000</p>
-        </item>
-        <item>
-          <p>
-	    <c>erlang:load_nif/2</c> returns new error type
-	    <c>notsup</c> when called for a HiPE compiled module,
-	    which is not supported.</p>
-          <p>
-	    Own Id: OTP-14002</p>
-        </item>
-        <item>
-          <p>
-	    Add driver and nif lock instrumentation to lcnt</p>
-          <p>
-	    Own Id: OTP-14069</p>
-        </item>
-        <item>
-          <p>
-	    Reduce memory pressure by converting sub-binaries to
-	    heap-binaries when possible. This is done during garbage
-	    collection.</p>
-          <p>
-	    Own Id: OTP-14149</p>
-        </item>
-        <item>
-          <p>
-	    Dirty schedulers are now enabled and supported on Erlang
-	    runtime systems with SMP support.</p>
-          <p>
-	    Besides support for dirty NIFs also support for dirty
-	    BIFs and dirty garbage collection have been introduced.
-	    All garbage collections that potentially will take a long
-	    time to complete are now performed on dirty schedulers if
-	    enabled.</p>
-          <p>
-	    <seealso
-	    marker="erts:erlang#statistics_scheduler_wall_time"><c>erlang:statistics(scheduler_wall_time)</c></seealso>
-	    now also includes information about dirty CPU schedulers.
-	    Code using <c>erlang:statistics(scheduler_wall_time)</c>
-	    may have to be rewritten to take this incompatibility
-	    into consideration.</p>
-          <p>
-	    *** POTENTIAL INCOMPATIBILITY ***</p>
-          <p>
-	    Own Id: OTP-14152</p>
-        </item>
-        <item>
-	    <p>Atoms may now contain arbitrary Unicode
-	    characters.</p>
-          <p>
-	    Own Id: OTP-14178</p>
-        </item>
-        <item>
-          <p>
-	    Introduce an event manager in Erlang to handle OS
-	    signals. A subset of OS signals may be subscribed to and
-	    those are described in the Kernel application.</p>
-          <p>
-	    Own Id: OTP-14186</p>
-        </item>
-        <item>
-          <p>
-	    The <c>escript</c> program now handles symbolic links to
-	    escripts.</p>
-          <p>
-	    This is useful for standalone systems with
-	    <c>escript</c>s residing on a bin directory not included
-	    in the execution path (as it may cause their <c>erl</c>
-	    program(s) to override the desired one). Instead the
-	    <c>escript</c>s can be referred to via symbolic links
-	    from a bin directory in the path.</p>
-          <p>
-	    Own Id: OTP-14201 Aux Id: PR-1293 </p>
-        </item>
-        <item>
-          <p>
-	    All uses of the magic binary kludge has been replaced by
-	    uses of erlang references.</p>
-          <p>
-	    A magic binary was presented as an empty binary, but
-	    actually referred other data internally in the Erlang VM.
-	    Since they were presented as empty binaries, different
-	    magic binaries compared as equal, and also lost their
-	    internal data when passed out of an erlang node.</p>
-          <p>
-	    The new usage of references has not got any of these
-	    strange semantic issues, and the usage of these
-	    references has been optimized to give the same
-	    performance benefits as well as memory usage benefits as
-	    magic binaries had.</p>
-          <p>
-	    A couple of examples of previous uses of magic binaries
-	    are match specifications and NIF resources.</p>
-          <p>
-	    Own Id: OTP-14205</p>
-        </item>
-        <item>
-          <p>
-	    The non-smp emulators have been deprecated and are
-	    scheduled for removal in OTP-21.</p>
-          <p>
-	    In preperations for this, the threaded non-smp emulator
-	    is no longer built by default and has to be enabled using
-	    the --enable-plain-emulator to configure.</p>
-          <p>
-	    *** POTENTIAL INCOMPATIBILITY ***</p>
-          <p>
-	    Own Id: OTP-14272</p>
-        </item>
-        <item>
-          <p>
-	    Allow HiPE to run on VM built with
-	    <c>--enable-m32-build</c>.</p>
-          <p>
-	    Own Id: OTP-14330 Aux Id: PR-1397 </p>
-        </item>
-        <item>
-          <p>
-	    Upgraded the OTP internal PCRE library from version 8.33
-	    to version 8.40. This library is used for implementation
-	    of the <seealso marker="stdlib:re"><c>re</c></seealso>
-	    regular expressions module.</p>
-          <p>
-	    Besides various bug fixes, the new version allows for
-	    better stack protection. In order to utilize this
-	    feature, the stack size of normal scheduler threads is
-	    now by default set to 128 kilo words on all platforms.
-	    The stack size of normal scheduler threads can be set
-	    upon system start by passing the <seealso
-	    marker="erts:erl#sched_thread_stack_size"><c>+sss</c></seealso>
-	    command line argument to the <seealso
-	    marker="erts:erl"><c>erl</c></seealso> command.</p>
-          <p>
-	    See <url
-	    href="http://pcre.org/original/changelog.txt"><c>http://pcre.org/original/changelog.txt</c></url>
-	    for information about changes made to PCRE between the
-	    versions 8.33 and 8.40.</p>
-          <p>
-	    *** POTENTIAL INCOMPATIBILITY ***</p>
-          <p>
-	    Own Id: OTP-14331 Aux Id: ERL-208 </p>
-        </item>
-        <item>
-          <p>
-	    Remove generation of atoms in old latin1 external format.
-	    The old latin1 format can still be decoded but is never
-	    generated by <c>term_to_binary</c> or sent on a
-	    distribution channel. The new utf8 format for atoms was
-	    introduced in OTP R16. An OTP 20 node can therefore not
-	    connect to nodes older than R16.</p>
-          <p>
-	    Own Id: OTP-14337</p>
-        </item>
-        <item>
-          <p>
-	    Added function <c>re:version/0</c> which returns
-	    information about the OTP internal PCRE version used for
-	    implementation of the <c>re</c> module.</p>
-          <p>
-	    Own Id: OTP-14347 Aux Id: PR-1412 </p>
-        </item>
-        <item>
-          <p>
-	    Added new debug bif <c>erlang:list_to_port/1</c>.</p>
-          <p>
-	    Own Id: OTP-14348</p>
-        </item>
-        <item>
-          <p>
-	    Various improvements of timer management internally in
-	    the VM. These improvements both reduced memory
-	    consumption of timer wheels as well as reduce the amount
-	    of work that has to be performed in order to handle
-	    timers.</p>
-          <p>
-	    Own Id: OTP-14356</p>
-        </item>
-        <item>
-	    <p> Sockets can now be bound to device (SO_BINDTODEVICE)
-	    on platforms where it is supported. </p> <p> This has
-	    been implemented e.g to support VRF-Lite under Linux; see
-	    <url
-	    href="https://www.kernel.org/doc/Documentation/networking/vrf.txt">
-	    VRF </url>, and GitHub pull request <url
-	    href="https://github.com/erlang/otp/pull/1326">#1326</url>.
-	    </p>
-          <p>
-	    Own Id: OTP-14357 Aux Id: PR-1326 </p>
-        </item>
-        <item>
-	    <p>Added the following <seealso
-	    marker="erl"><c>erl</c></seealso> command line arguments
-	    with which you can set suggested stack for dirty
-	    schedulers:</p> <taglist> <tag><seealso
-	    marker="erl#dcpu_sched_thread_stack_size"><c>+sssdcpu</c></seealso></tag>
-	    <item><p>for dirty CPU schedulers</p></item>
-	    <tag><seealso
-	    marker="erl#dio_sched_thread_stack_size"><c>+sssdio</c></seealso></tag>
-	    <item><p>for dirty IO schedulers</p></item> </taglist>
-	    <p>The default suggested stack size for dirty schedulers
-	    is 40 kilo words.</p>
-          <p>
-	    Own Id: OTP-14380</p>
-        </item>
-        <item>
-          <p>
-	    Changed erts startup program name, argv 0, to use the
-	    environment variable <c>ESCRIPT_NAME</c> so that
-	    <c>erlc</c>, <c>dialyzer</c>, <c>typer</c>,
-	    <c>ct_run</c>, or the escript name can be seen with
-	    external programs, such as ps and htop (depending on
-	    options), on unix.</p>
-          <p>
-	    If environment variable <c>ESCRIPT_NAME</c> is not set,
-	    argv 0 of the invoking program will be used, e.g. the
-	    name will be <c>erl</c>, <c>cerl</c>, or the soft link
-	    name.</p>
-          <p>
-	    Own Id: OTP-14381</p>
-        </item>
-        <item>
-          <p>
-	    Improvements of <c>escript</c> documentation.</p>
-          <p>
-	    Own Id: OTP-14384 Aux Id: OTP-14201 </p>
-        </item>
-        <item>
-          <p>
-	    Add function <c>enif_hash</c> for NIFs to calculate hash
-	    values of arbitrary terms.</p>
-          <p>
-	    Own Id: OTP-14385 Aux Id: PR-1413 </p>
-        </item>
-      </list>
-    </section>
-
-</section>
-
 <section><title>Erts 8.3.3</title>
 
     <section><title>Fixed Bugs and Malfunctions</title>