Updating README.md file with shell startup performance data (#20)

* Removing duplicated check for interactive shell in .zshrc * Updating README.md file with shell startup performance data
hpedrorodrigues · Dec 7, 2020 · 1d00a78 · 1d00a78
1 parent 9609b6e
commit 1d00a78
Show file tree

Hide file tree

Showing 3 changed files with 78 additions and 44 deletions.
diff --git a/.file/os/linux/.zshrc b/.file/os/linux/.zshrc
@@ -25,24 +25,23 @@ if [ -f "${HOME}/.zgen/zgen.zsh" ]; then
 
     zgen save
   fi
+fi
 
-  # Force regenerate the .zcompdump file with the sphynx completions
-  # if they were not included before
-  if ! grep -q sx "${HOME}/.zcompdump"; then
+if hash sx &>/dev/null; then
+  # Regenerating .zcompdump file when sphynx completions were not loaded
+  if ! grep -q 'sx' "${HOME}/.zcompdump"; then
     autoload -U compinit && compinit
   fi
-fi
 
-# Load the tmux "hack" session on zsh startup
-# Notes:
-# - not loading tmux in sphynx benchmarks because it's not important for it
-# - not loading tmux if a screen session is currently active
-# - loading the tmux session only when using Alacritty
-if [[ $- == *i* ]] \
-  && hash sx &>/dev/null \
-  && [ -z "${SX_SHELL_BENCHMARK}" ] \
-  && [ -z "${STY}" ] \
-  && [ -z "${TMUX}" ] \
-  && [ -n "${ALACRITTY_LOG}" ]; then
-  sx terminal tmux force-attach hack
+  # Load the tmux "hack" session on zsh startup
+  # Notes:
+  # - not loading tmux in sphynx benchmarks because it's not important for it
+  # - not loading tmux if a screen session is currently active
+  # - loading the tmux session only when using Alacritty
+  if [ -z "${SX_SHELL_BENCHMARK}" ] \
+    && [ -z "${STY}" ] \
+    && [ -z "${TMUX}" ] \
+    && [ -n "${ALACRITTY_LOG}" ]; then
+    sx terminal tmux force-attach hack
+  fi
 fi
diff --git a/.file/os/osx/.zshrc b/.file/os/osx/.zshrc
@@ -25,24 +25,23 @@ if [ -f "${HOME}/.zgen/zgen.zsh" ]; then
 
     zgen save
   fi
+fi
 
-  # Force regenerate the .zcompdump file with the sphynx completions
-  # if they were not included before
-  if ! grep -q sx "${HOME}/.zcompdump"; then
+if hash sx &>/dev/null; then
+  # Regenerating .zcompdump file when sphynx completions were not loaded
+  if ! grep -q 'sx' "${HOME}/.zcompdump"; then
     autoload -U compinit && compinit
   fi
-fi
 
-# Load the tmux "hack" session on zsh startup
-# Notes:
-# - not loading tmux in sphynx benchmarks because it's not important for it
-# - not loading tmux if a screen session is currently active
-# - loading the tmux session only when using Alacritty
-if [[ $- == *i* ]] \
-  && hash sx &>/dev/null \
-  && [ -z "${SX_SHELL_BENCHMARK}" ] \
-  && [ -z "${STY}" ] \
-  && [ -z "${TMUX}" ] \
-  && [ -n "${ALACRITTY_LOG}" ]; then
-  sx terminal tmux force-attach hack
+  # Load the tmux "hack" session on zsh startup
+  # Notes:
+  # - not loading tmux in sphynx benchmarks because it's not important for it
+  # - not loading tmux if a screen session is currently active
+  # - loading the tmux session only when using Alacritty
+  if [ -z "${SX_SHELL_BENCHMARK}" ] \
+    && [ -z "${STY}" ] \
+    && [ -z "${TMUX}" ] \
+    && [ -n "${ALACRITTY_LOG}" ]; then
+    sx terminal tmux force-attach hack
+  fi
 fi
diff --git a/README.md b/README.md
@@ -49,22 +49,58 @@ files! Beware!
 
 **zsh**
 ```bash
-♪ repeat 5 { /usr/bin/time zsh -i -c exit }
-        0.10 real         0.05 user         0.04 sys
-        0.11 real         0.06 user         0.04 sys
-        0.11 real         0.06 user         0.04 sys
-        0.11 real         0.06 user         0.04 sys
-        0.12 real         0.06 user         0.05 sys
+♪ sx shell benchmark zsh
+> Hyperfine
+
+Benchmark #1: zsh -i -c exit
+  Time (mean ± σ):     130.6 ms ±   2.4 ms    [User: 70.4 ms, System: 52.5 ms]
+  Range (min … max):   126.4 ms … 136.7 ms    21 runs
+
+> Bench
+
+benchmarking zsh -i -c exit
+time                 128.7 ms   (125.4 ms .. 131.6 ms)
+                     0.999 R²   (0.998 R² .. 1.000 R²)
+mean                 126.2 ms   (125.3 ms .. 127.6 ms)
+std dev              1.628 ms   (964.9 μs .. 2.354 ms)
+variance introduced by outliers: 11% (moderately inflated)
+
+> Time
+
+/usr/bin/time zsh -i -c exit (5x)
+        0.14 real         0.07 user         0.06 sys
+        0.13 real         0.07 user         0.05 sys
+        0.13 real         0.07 user         0.05 sys
+        0.14 real         0.07 user         0.05 sys
+        0.16 real         0.08 user         0.07 sys
 ```
 
 **bash**
 ```bash
-♪ repeat 5 { /usr/bin/time bash -i -c exit }
-        0.10 real         0.06 user         0.06 sys
-        0.10 real         0.06 user         0.05 sys
-        0.09 real         0.06 user         0.05 sys
-        0.09 real         0.06 user         0.05 sys
-        0.10 real         0.06 user         0.05 sys
+♪ sx shell benchmark bash
+> Hyperfine
+
+Benchmark #1: bash -i -c exit
+  Time (mean ± σ):     111.5 ms ±   2.7 ms    [User: 71.1 ms, System: 63.4 ms]
+  Range (min … max):   107.7 ms … 117.6 ms    24 runs
+
+> Bench
+
+benchmarking bash -i -c exit
+time                 104.7 ms   (97.34 ms .. 112.0 ms)
+                     0.989 R²   (0.962 R² .. 1.000 R²)
+mean                 103.8 ms   (100.4 ms .. 106.9 ms)
+std dev              5.138 ms   (3.102 ms .. 7.333 ms)
+variance introduced by outliers: 10% (moderately inflated)
+
+> Time
+
+/usr/bin/time bash -i -c exit (5x)
+        0.16 real         0.08 user         0.09 sys
+        0.11 real         0.07 user         0.06 sys
+        0.11 real         0.07 user         0.06 sys
+        0.11 real         0.07 user         0.07 sys
+        0.11 real         0.07 user         0.06 sys
 ```
 
 **Inspiration**: