| 1 | #!/usr/bin/env bash
|
| 2 | #
|
| 3 | # Junk drawer for refactoring. Also see test/lint.sh
|
| 4 | #
|
| 5 | # Usage:
|
| 6 | # devtools/refactor.sh <function name>
|
| 7 |
|
| 8 | : ${LIB_OSH=stdlib/osh}
|
| 9 | source $LIB_OSH/bash-strict.sh
|
| 10 | source $LIB_OSH/task-five.sh
|
| 11 |
|
| 12 | change-kind() {
|
| 13 | local kind=$1
|
| 14 | local kind2=${2:-$kind}
|
| 15 |
|
| 16 | # First make it all lower case
|
| 17 | sed -r -i "s/${kind}_([A-Z]+)/Id.${kind2}_\\L\\1/g" */*.py
|
| 18 |
|
| 19 | # Now make the first one upper case
|
| 20 | sed -r -i "s/${kind}_([a-zA-Z]+)/${kind2}_\\u\\1/g" */*.py
|
| 21 | }
|
| 22 |
|
| 23 | k2() {
|
| 24 | local kind=$1
|
| 25 | local lower=${kind,,}
|
| 26 |
|
| 27 | local title=${lower^}
|
| 28 | local replace=${2:-$title}
|
| 29 |
|
| 30 | sed -r -i "s/Id.${kind}_/Id.${replace}_/g" */*.py
|
| 31 | sed -r -i "s/TokenKind.${kind}/TokenKind.${replace}/g" */*.py
|
| 32 | }
|
| 33 |
|
| 34 | # Execute a bunch of find/replace pairs in a text file.
|
| 35 | replace() {
|
| 36 | local file=$1
|
| 37 | local include_asdl=${2:-}
|
| 38 |
|
| 39 | # NOTE: Escaping here is messed up. sed doesn't have --name like awk?
|
| 40 | # To match literal parentheses I had to double-escape like this
|
| 41 | # (shell-escape, then sed-escape).
|
| 42 | # MakeMatcher\\(\\) MATCHER
|
| 43 |
|
| 44 | local -a files=( */*.py )
|
| 45 | if test -n "$include_asdl"; then
|
| 46 | files+=( */*.asdl )
|
| 47 | fi
|
| 48 |
|
| 49 | while read pat replace; do
|
| 50 | sed -r -i "s/${pat}/${replace}/g" "${files[@]}"
|
| 51 |
|
| 52 | # word-anchored version
|
| 53 | #sed -r -i "s/\b${pat}\b/${replace}/g" "${files[@]}"
|
| 54 | done < $file
|
| 55 | }
|
| 56 |
|
| 57 | replace2() {
|
| 58 | #sed -r -i "s/^from osh import parse_lib/from frontend import parse_lib/g" */*.py
|
| 59 | #sed -r -i "s/^from core import libstr/from osh import string_ops/g" */*.py
|
| 60 | #sed -r -i "s/^from osh import word$/from osh import word_/g" */*.py
|
| 61 | #sed -r -i 's/from _devbuild.gen.syntax_asdl import word as osh_word/from _devbuild.gen.syntax_asdl import word/g' */*.py
|
| 62 | #sed -r -i 's/osh_word/word/g' */*.py
|
| 63 |
|
| 64 | if test -n ''; then
|
| 65 | sed -r -i 's/bool_expr.BoolUnary/bool_expr.Unary/g' */*.py
|
| 66 | sed -r -i 's/bool_expr.BoolBinary/bool_expr.Binary/g' */*.py
|
| 67 | sed -r -i 's/bool_expr_e.BoolUnary/bool_expr_e.Unary/g' */*.py
|
| 68 | sed -r -i 's/bool_expr_e.BoolBinary/bool_expr_e.Binary/g' */*.py
|
| 69 | sed -r -i 's/bool_expr__BoolUnary/bool_expr__Unary/g' */*.py
|
| 70 | sed -r -i 's/bool_expr__BoolBinary/bool_expr__Binary/g' */*.py
|
| 71 | fi
|
| 72 |
|
| 73 | sed -r -i 's/command.SimpleCommand/command.Simple/g' */*.py
|
| 74 | sed -r -i 's/command_e.SimpleCommand/command_e.Simple/g' */*.py
|
| 75 | sed -r -i 's/command__SimpleCommand/command__Simple/g' */*.py
|
| 76 | }
|
| 77 |
|
| 78 | trailing-ws() {
|
| 79 | sed -r -i 's/[ ]+$//g' "$@"
|
| 80 | }
|
| 81 |
|
| 82 | #
|
| 83 | # OLD STUFF
|
| 84 | #
|
| 85 |
|
| 86 | # Hm all of the solutions involve grep --perl or perl itself?
|
| 87 | # https://stackoverflow.com/questions/3001177/how-do-i-grep-for-all-non-ascii-characters-in-unix
|
| 88 |
|
| 89 | # Found a latin-1 character in Python-2.7.13/Lib/heapq.py. Had to add LC_ALL=C.
|
| 90 | grep-unicode() {
|
| 91 | LC_ALL=C grep --color='auto' --perl -n '[^\x00-\x7F]' "$@"
|
| 92 | }
|
| 93 |
|
| 94 | find-old-asdl() {
|
| 95 | egrep 'import.*\bruntime\b' */*.py || true
|
| 96 | echo ---
|
| 97 |
|
| 98 | # Only tests left
|
| 99 | egrep 'import.*\bast\b' */*.py || true
|
| 100 | }
|
| 101 |
|
| 102 | # This should be cleaned up
|
| 103 | grep-span-funcs() {
|
| 104 | grep MostSpan {osh,core,frontend}/*.py
|
| 105 | }
|
| 106 |
|
| 107 | cmd-val() {
|
| 108 | local file=$1
|
| 109 | sed -i 's/arg_vec.strs/cmd_val.argv/g' $file
|
| 110 | sed -i 's/arg_vec.spids/cmd_val.arg_spids/g' $file
|
| 111 | sed -i 's/arg_vector/cmd_value__Argv/g' $file
|
| 112 | sed -i 's/arg_vec/cmd_val/g' $file
|
| 113 | }
|
| 114 |
|
| 115 | id-kind() {
|
| 116 | sed --regexp-extended -i 's/import id_kind$/import id_kind_def/' */*.py
|
| 117 | sed --regexp-extended -i 's/id_kind[.]/id_kind_def./g' */*.py
|
| 118 | }
|
| 119 |
|
| 120 | options() {
|
| 121 | #sed -i 's/simple_echo/simple_echo/g' */*.{py,md,sh}
|
| 122 |
|
| 123 | #sed -i 's/simple_eval_builtin/simple_eval_builtin/g' */*.{py,md,sh}
|
| 124 | #sed -i 's/simple_trap/simple_trap/g' */*.{py,md,sh}
|
| 125 |
|
| 126 | sed -i 's/parse_backslash/parse_backslash/g' */*.{py,md,sh}
|
| 127 | sed -i 's/parse_backticks/parse_backticks/g' */*.{py,md,sh}
|
| 128 | sed -i 's/parse_dollar/parse_dollar/g' */*.{py,md,sh}
|
| 129 | }
|
| 130 |
|
| 131 | rename() {
|
| 132 | cat cpp/cc.txt | while read name; do
|
| 133 | local base=$(basename $name .cc)
|
| 134 | local new=leaky_${base%_leaky}.cc
|
| 135 | echo $name $new
|
| 136 |
|
| 137 | #sed -i "s/$name/$new/g" */*.{py,sh} || true
|
| 138 | git mv cpp/$name cpp/$new
|
| 139 | done
|
| 140 | return
|
| 141 |
|
| 142 | cat cpp/h.txt | while read name; do
|
| 143 | local base=$(basename $name .h)
|
| 144 | local new=leaky_${base%_leaky}.h
|
| 145 | echo $name $new
|
| 146 |
|
| 147 | #sed -i "s/$name/$new/g" */*.{sh,py,cc,h}
|
| 148 | git mv cpp/$name cpp/$new
|
| 149 | done
|
| 150 | return
|
| 151 | }
|
| 152 |
|
| 153 | revert() {
|
| 154 | cat cpp/cc.txt | while read name; do
|
| 155 | local new=$(basename $name .cc)_leaky.cc
|
| 156 | echo $name $new
|
| 157 |
|
| 158 | mv cpp/$new cpp/$name
|
| 159 | done
|
| 160 | }
|
| 161 |
|
| 162 | #
|
| 163 | # Things we want to get rid of
|
| 164 | #
|
| 165 |
|
| 166 | show-usages() {
|
| 167 | local out=$1
|
| 168 | shift
|
| 169 | "$@" | grep -v _test.py | tee $out
|
| 170 | echo
|
| 171 | wc -l $out
|
| 172 | }
|
| 173 |
|
| 174 | # 2024-02: 36 usages. Maybe use mylib.ToMachineInt() or mylib.ToInt32(). exit
|
| 175 | # status is a machine int?
|
| 176 | # Sometimes the lexer will validate, as with converting 2>&1
|
| 177 |
|
| 178 | int-convert() {
|
| 179 | show-usages _tmp/int-convert \
|
| 180 | egrep -n '\bint\(\b' $(metrics/source-code.sh osh-files)
|
| 181 | }
|
| 182 |
|
| 183 | # TokenVal() is generally bad in evaluators. However most of these are in
|
| 184 | # error paths, which is OK.
|
| 185 | #
|
| 186 | # 2024-02: 11 instances
|
| 187 | TokenVal-eval() {
|
| 188 | show-usages _tmp/TokenVal-eval \
|
| 189 | grep -n -w TokenVal */*_eval.py
|
| 190 | }
|
| 191 |
|
| 192 | TokenVal-all() {
|
| 193 | show-usages _tmp/TokenVal-all \
|
| 194 | grep -n -w TokenVal */*.py
|
| 195 | }
|
| 196 |
|
| 197 | # 2024-02: 71 left, mostly in ysh_ify which is the only thing that uses it
|
| 198 | spid-all() {
|
| 199 | show-usages _tmp/spid-all \
|
| 200 | egrep -n 'span_id|spid' */*.py
|
| 201 | }
|
| 202 |
|
| 203 | # 2024-04: 4 left in ysh_ify
|
| 204 | spid-sig() {
|
| 205 | show-usages _tmp/spid-sig \
|
| 206 | egrep -n 'def.*(span_id|spid)' */*.py
|
| 207 | }
|
| 208 |
|
| 209 | # 2024-04: 12 usages, mostly ysh_ify
|
| 210 | no-spid() {
|
| 211 | show-usages _tmp/no-spid \
|
| 212 | egrep -n 'runtime.NO_SPID' */*.py
|
| 213 | }
|
| 214 |
|
| 215 | # 69 instances
|
| 216 | loc-word() {
|
| 217 | # should NOT wrap CompoundWord
|
| 218 | show-usages _tmp/loc-word \
|
| 219 | fgrep -n 'loc.Word(' */*.py
|
| 220 | }
|
| 221 |
|
| 222 | # 2023-08: 155
|
| 223 | loc-missing() {
|
| 224 | show-usages _tmp/loc-m \
|
| 225 | egrep -n 'loc.Missing' */*.py
|
| 226 | }
|
| 227 |
|
| 228 | mylib-python() {
|
| 229 | show-usages _tmp/py \
|
| 230 | egrep -n 'mylib.PYTHON' */*.py
|
| 231 | }
|
| 232 |
|
| 233 | asdl-create() {
|
| 234 | fgrep -n 'CreateNull(alloc' */*.py */*/*.py \
|
| 235 | | egrep -v '_devbuild|_test.py' | tee _tmp/asdl
|
| 236 | }
|
| 237 |
|
| 238 | readline() {
|
| 239 | metrics/source-code.sh oils-files | xargs fgrep -n 'readline('
|
| 240 | }
|
| 241 |
|
| 242 | #
|
| 243 | # To improve code formatting
|
| 244 | #
|
| 245 |
|
| 246 | long-sigs() {
|
| 247 | # 32 of these
|
| 248 | egrep --no-filename '^[ ]*# type' */*.py \
|
| 249 | | awk 'length($0) >= 80 { print length($0) $0 }' \
|
| 250 | | sort -n
|
| 251 | }
|
| 252 |
|
| 253 | long-sigs-where() {
|
| 254 | # jump to the file
|
| 255 | egrep -n '^[ ]*# type' */*.py \
|
| 256 | | awk 'length($0) >= 110 { print }' | tee _tmp/long
|
| 257 | }
|
| 258 |
|
| 259 | #
|
| 260 | # Refactor tests
|
| 261 | #
|
| 262 |
|
| 263 | print-names() {
|
| 264 | egrep -o '[a-zA-Z_-]+'
|
| 265 | }
|
| 266 |
|
| 267 | make-sed() {
|
| 268 | awk '{ print "s/" $0 "/unquoted-" $0 "/g;" }'
|
| 269 | }
|
| 270 |
|
| 271 | test-files() {
|
| 272 | cat _tmp/r | print-names | make-sed | tee _tmp/sedr
|
| 273 |
|
| 274 | sed -i -f _tmp/sedr test/runtime-errors.sh
|
| 275 | }
|
| 276 |
|
| 277 | #
|
| 278 | # Boxless optimization
|
| 279 | #
|
| 280 |
|
| 281 | singleton-list() {
|
| 282 | egrep -n '\(List\[[^]]+\] [a-z]+\)' */*.asdl
|
| 283 | echo
|
| 284 | egrep -n '\(Dict\[[^]]+\] [a-z]+\)' */*.asdl
|
| 285 | }
|
| 286 |
|
| 287 | task-five "$@"
|