PRESENT Block Cipher

[gh-yhxnf]

Implementation of the ultra-lightweight block cipher PRESENT.

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​3c11154

removed ... in EXAMPLES section

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​027af50

added tests

[asante]

I would rename the doLastLinearLayer argument to doFinalRound, as this will fit/be more consistent, when implementing other block ciphers.

I also reviewed the ticket and it lgtm, but maybe someone else also want to take a look?

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​0ab40c0	fixed typos
​fd64dbd	renamed doLastLinearLayer to doFinalRound
​559ceac	updated disclaimer

[asante]

I'm setting this to positive review now.

[asante]

actually I realised that I would like to have some design changes. I'll add them in a minute

[git]

Branch pushed to git repo; I updated commit sha1. Last 10 new commits:

​5407256	moved determination of inputType out of _convert_to_vector
​1a1b6fe	added sbox_layer function
​18592f8	use little endian sbox
​3c174e0	expose sbox variable
​c1e355c	added linear_layer function
​d0ff01e	added round function
​6c9bf2d	renamed variables
​55cb61e	expose key schedule
​3fd87e1	added example
​b429007	Merge branch 'u/gh-yhxnf/present_block_cipher' of trac.sagemath.org:sage into present

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​01c2822	do not store inverse sbox
​3e9e673	use big endian PRESENT sbox
​e692eaf	add present.round example

[asante]

OK, LGTM with one last point:

If you do the following import in an example

sage: from sage.crypto.block_cipher.present import _convert_to_vector

I think the _convert_to_vector function should not be hidden by default.

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​32ec4a9

unhide _convert_to_vector

[asante]

lgtm

However, I have problems building the documentation, it fails with some error in the categorie docs -- which should actually not be related to this ticket? Maybe we should merge the latest development branch, for which I am able to build the documentation.

[asante]

I just did the rebase on the latest develop branch.

Docs now build correctly.

LGTM

---

New commits:

​6f1bfa7

Merge branch 'develop' into t/27573/present_block_cipher

[vbraun]

The PDF docs don't build

[docpdf] ! Package amsmath Error: \begin{aligned} allowed only in math mode.
[docpdf] 
[docpdf] See the amsmath package documentation for explanation.
[docpdf] Type  H <return>  for immediate help.
[docpdf]  ...                                              
[docpdf]                                                   
[docpdf] l.6727 \ 
[docpdf]          [k_{79}k_{78}\dots k_{1}k_{0}] &= [k_{18}k_{17}\dots k_{20}k_{19}] \\
[docpdf] ? 
[docpdf] ! Emergency stop.
[docpdf]  ...                                              
[docpdf]                                                   
[docpdf] l.6727 \ 
[docpdf]          [k_{79}k_{78}\dots k_{1}k_{0}] &= [k_{18}k_{17}\dots k_{20}k_{19}] \\
[docpdf] !  ==> Fatal error occurred, no output PDF file produced!
[docpdf] Transcript written on cryptography.log.

[gh-yhxnf]

New commits:

​65839f3	removed unneeded import
​798421b	remove :nowrap:

[gh-yhxnf]

Replying to vbraun:

The PDF docs don't build

should be fixed now

[asante]

LGTM

[jhpalmieri]

This causes a doctest failure with Python 3. Follow up at #28403.

[chapoton]

I recall that development should now be done with python3. This would prevent this to happen.