Commit 2ec13188 authored by Niels Möller's avatar Niels Möller

Updated release plan for 2.8.

parent 66bf0965
......@@ -28,83 +28,17 @@
Leave for some later release!
</p>
<h1> Plan for the Nettle-2.7 release </h1>
<h2> Code </h2>
<p class='done'>
Support older GMP versions, as far as practical.
</p>
<p class='done'>
Make use of GMP's mpn_cnd_add_n and mpn_sub_n (which will be
available in the next GMP release).
</p>
<p class='done'>
Implement <tt>ecc_point_mul</tt> and <tt>ecc_point_mul_g</tt>.
</p>
<p class='done'>
Rewrite x86_64/sha3-permute.asm. Moves between xmm registers and
regular registers kill performance at least on some AMD
processors. (Not rewritten, only moves changed to go via memory).
</p>
<p class='done'>
Include UMAC. Needs a serious effort to produce test vectors.
</p>
<h2> Documentation </h2>
<p class='done'>
Document new ECC functions.
</p>
<p class='done'>
Document salsa20r12.
</p>
<p class='done'>
Document umac.
</p>
<p class='done'>
Update NEWS file.
</p>
<h2> Build system </h2>
<p class='done'>
Ensure that <tt>make install</tt> without a preceding
<tt>make</tt> works.
</p>
<p class='done'>
Make the assembly ALIGN-macro non-logarithmic, reusing the m4 log2
macro in GMP. Current macros have the problem that in m4 eval,
<tt>&lt;&lt;</tt> is not supported by Solaris, while <tt>**</tt>
is not supported by OpenBSD.
</p>
<h2> Testing </h2>
<p> Since xenofarm isn't up and running, do some manual testing:
</p>
<ul>
<li class='done'> x86_64-linux-gnu</li>
<li class='should'> x86_64-freebsd</li>
<li class='done'> x86-linux-gnu</li>
<li class='should'> x86-freebsd</li>
<li class='done'> x86-w*ndows (using cross compiler and wine)</li>
<li class='wish'> x86_64-w*ndows (seems a bit harder since 64-bit
wine is not yet available in debian)</li>
<li class='should'> x86-darwin (needs help from Nettle users)</li>
<li class='should'> x86_64-darwin (needs help from Nettle users)</li>
<li class='done'> armv7-linux-gnu (pandaboard test platform)</li>
<li class='wish'> armv7-android (possible test platform: N10 tablet)</li>
<li class='done'> sparc32-solaris10</li>
<li class='done'> sparc64-solaris10</li>
</ul>
<h1> Plans for nettle-2.8 </h1>
<p> nettle-2.8 is intended to be a minor API cleanup, with few new
features. </p>
<h2> Interface changes </h2>
<p class='should'>
<p class='done'>
Change the type of all lengths from <tt>unsigned</tt>
to <tt>size_t</tt>. An ABI change on most 64-bit platforms.
</p>
<p class='should'>
<p class='done'>
Change argument type of memxor and memxor3 from uint8_t * to void
* (like modern memcpy). Consider them unconditionally part of the
nettle library, with nettle_ prefix on the symbols, and no
......@@ -117,6 +51,12 @@
<p class='should'>
Move some internal-use macros from macros.h to nettle-internal.h.
</p>
<p class='should'>
Do separate aes128_ctx, etc, with smaller allocation for subkeys.
</p>
<p class='should'>
Don't require initialization of *dst_length for base*_decode_update.
</p>
<h2> Build system </h2>
<p class='should'>
......@@ -129,6 +69,25 @@
Delay building of test programs until <tt>make check</tt>.
</p>
<h2> Testing </h2>
<p> Since xenofarm isn't up and running, do some manual testing:
</p>
<ul>
<li class='should'> x86_64-linux-gnu</li>
<li class='should'> x86_64-freebsd</li>
<li class='should'> x86-linux-gnu</li>
<li class='should'> x86-freebsd</li>
<li class='should'> x86-w*ndows (using cross compiler and wine)</li>
<li class='wish'> x86_64-w*ndows (seems a bit harder since 64-bit
wine is not yet available in debian)</li>
<li class='should'> x86-darwin (needs help from Nettle users)</li>
<li class='should'> x86_64-darwin (needs help from Nettle users)</li>
<li class='should'> armv7-linux-gnu (pandaboard test platform)</li>
<li class='wish'> armv7-android (possible test platform: N10 tablet)</li>
<li class='should'> sparc32-solaris10</li>
<li class='should'> sparc64-solaris10</li>
</ul>
<h1> Plans for nettle-3.0 </h1>
<p> These are some larger API changes under consideration. </p>
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment