We use pexpect version 2.0. Shouldn't we upgrade to the current version 4.0.1?

See also #10294.

New upstream tarballs:

• https://pypi.python.org/packages/source/p/pexpect/pexpect-4.0.1.tar.gz
• https://pypi.python.org/packages/source/p/ptyprocess/ptyprocess-0.5.tar.gz
• http://www.math-cs.gordon.edu/~kcrisman/sagenb-0.11.6.1.tar

Patches included and submitted upstream:

• Add new delayafterread class attribute, for removing Superfluous sleep pexpect/pexpect#291
• Allow overriding the call to PtyProcess.spawn() pexpect/pexpect#303
• Read more bytes pexpect/pexpect#304
• Allow delaybeforesend=None to skip the sleep completely pexpect/pexpect#307

Note that this branch includes #19616.

Throughput timings (best out of 5):

pexpect 2.0 upstream:

sage: %time _ = str(gp("2^2^22"))
CPU times: user 184 ms, sys: 4 ms, total: 188 ms
Wall time: 384 ms

pexpect 4.0.1 upstream:

sage: %time _ = str(gp("2^2^22"))
CPU times: user 208 ms, sys: 4 ms, total: 212 ms
Wall time: 405 ms

pexpect 4.0.1 + Sage patches:

sage: %time _ = str(gp("2^2^22"))
CPU times: user 9 ms, sys: 7 ms, total: 16 ms
Wall time: 209 ms

Latency timings (best out of 5):

pexpect 2.0 upstream:

sage: gp(1);
sage: %time _ = [gp(i) for i in range(10^4)]
CPU times: user 2.96 s, sys: 855 ms, total: 3.82 s
Wall time: 4.52 s

pexpect 4.0.1 upstream:

sage: gp(1);
sage: %time _ = [gp(i) for i in range(10^4)]
CPU times: user 2.61 s, sys: 758 ms, total: 3.37 s
Wall time: 3.48 s

pexpect 4.0.1 + Sage patches:

sage: gp(1);
sage: %time _ = [gp(i) for i in range(10^4)]
CPU times: user 2.3 s, sys: 999 ms, total: 3.3 s
Wall time: 3.4 s

Depends on #19671

Upstream: Reported upstream. Developers acknowledge bug.

CC: @sagetrac-drkirkby @nexttime @sagetrac-Vincent-Neri

Component: interfaces

Keywords: pexpect upgrade

Author: François Bissey, Bill Page, Jeroen Demeyer

Branch/Commit: 9351ccb

Reviewer: Jeroen Demeyer, Bill Page

Issue created by migration from https://trac.sagemath.org/ticket/10295