[cleanup] Fix typos in source code (#1693)

- clarify some unclear comments
- Fix typos with `typos -w`
- Fix remaining typos manually

While the typos spell checker could report these typos
it couldn't automatically fix them since there are multiple possible
corrections, e.g. `agains` could be `again` or `against`.

Makefile

@@ -86,7 +86,7 @@ BYTECODE_ZIP := bytecode-opy.zip
 
 HAVE_OBJCOPY := $(shell command -v objcopy 2>/dev/null)
 
-# For faster tesing of builds
+# For faster testing of builds
 #default: _bin/oil.ovm-dbg
 
 # What the end user should build when they type 'make'.

README.md

@@ -274,7 +274,7 @@ These files make the slow "Oils Python" build, which is very different than the
     configure
     install
 
-These files are for teh C++ `oils-for-unix` tarball (in progress):
+These files are for the C++ `oils-for-unix` tarball (in progress):
 
     _build/
       oils.sh

asdl/gen_cpp.py

@@ -342,7 +342,7 @@ def Emit(s, depth=depth):
         # This is the base class.
         Emit('class %(sum_name)s_t {')
         # Can't be constructed directly.  Note: this shows up in uftrace in debug
-        # mode, e.g. when we intantiate Token.  Do we need it?
+        # mode, e.g. when we instantiate Token.  Do we need it?
         Emit(' protected:')
         Emit('  %s_t() {' % sum_name)
         Emit('  }')

asdl/gen_python.py

@@ -522,7 +522,7 @@ def VisitCompoundSum(self, sum, sum_name, depth):
             if variant.shared_type:
                 continue  # Don't generate a class for shared types.
             if len(variant.fields) == 0:
-                # We must use the old-style nameing here, ie. command__NoOp, in order
+                # We must use the old-style naming here, ie. command__NoOp, in order
                 # to support zero field variants as constants.
                 class_name = '%s__%s' % (sum_name, variant.name)
                 self._GenClass(variant, class_name, (sum_name + '_t',), i + 1)

benchmarks/osh-parser.sh

@@ -427,7 +427,7 @@ EOF
   fi
 
   if test -f $in_dir/elapsed.csv; then
-    cmark <<< '#### Elasped Time (milliseconds)'
+    cmark <<< '#### Elapsed Time (milliseconds)'
     echo
     csv2html $in_dir/elapsed.csv
   fi

benchmarks/pypy.sh

@@ -33,7 +33,7 @@ parse-with-cpython() {
 
 # ~4.8 seconds
 # NOTE: We could run it in a loop to see if the JIT warms up, but that would
-# only be for curiousity.  Most shell processes are short-lived, so it's the
+# only be for curiosity.  Most shell processes are short-lived, so it's the
 # wrong thing to optimize for.
 parse-with-pypy() {
   parse-abuild $PYPY

benchmarks/report.R

@@ -5,7 +5,7 @@
 # Usage:
 #   benchmarks/report.R OUT_DIR [TIMES_CSV...]
 
-# Supress warnings about functions masked from 'package:stats' and 'package:base'
+# Suppress warnings about functions masked from 'package:stats' and 'package:base'
 #   filter, lag
 #   intersect, setdiff, setequal, union
 library(dplyr, warn.conflicts = FALSE)
@@ -1198,7 +1198,7 @@ Percent = function(n, total) {
 }
 
 PrettyPrintLong = function(d) {
-  tr = t(d)  # tranpose
+  tr = t(d)  # transpose
 
   row_names = rownames(tr)
 

benchmarks/vm-baseline.sh

@@ -28,7 +28,7 @@ measure() {
   # TODO:
   # print-tasks should:
   # - use the whole shell path like _bin/osh
-  # - the host name shoudl be a column
+  # - the host name should be a column
   # - the join ID can be a file, and construct the task name from that
   # - Then maybe use tsv_columns_from_files.py like we do with cachegrind
 

build/common.sh

@@ -37,10 +37,10 @@ readonly CLANGXX=$CLANG_DIR/bin/clang++
 # I'm not sure if there's a GCC version of this?
 export ASAN_SYMBOLIZER_PATH=$CLANG_DIR_RELATIVE/bin/llvm-symbolizer
 
-# ThreadSanitizer doesn't always give us all locaitons, but this doesn't help
+# ThreadSanitizer doesn't always give us all locations, but this doesn't help
 # export TSAN_SYMBOLIZER_PATH=$ASAN_SYMBOLIZER_PATH
 
-# equivalent of 'cc' for C++ langauge
+# equivalent of 'cc' for C++ language
 # https://stackoverflow.com/questions/172587/what-is-the-difference-between-g-and-gcc
 CXX=${CXX:-'c++'}
 

build/cpython_defs.py

@@ -388,8 +388,8 @@ def __call__(self, rel_path, def_name, method_name):
 
     #log('= %s %s', def_name, method_name)
 
-    # If it doesn't appear in the .py source, it can't be used.  (Execption: it
-    # coudl be used in C source with dynamic lookup?  But I don't think CPython
+    # If it doesn't appear in the .py source, it can't be used.  (Exception: it
+    # could be used in C source with dynamic lookup?  But I don't think CPython
     # does that.)
     #if method_name not in self.py_names:
     if 0:

build/deps.sh

@@ -96,7 +96,7 @@ install-ubuntu-packages() {
   ### Packages for build/py.sh all, building wedges, etc.
 
   # python2-dev is no longer available on Debian 12
-  # python-dev als seems gone
+  # python-dev also seems gone
   #
   # g++: essential
   # libreadline-dev: needed for the build/prepare.sh Python build.

build/dev-shell-test.sh

@@ -50,7 +50,7 @@ test-cli() {
 test-python2() {
   banner "Testing python2"
 
-  # Can't do this beacuse of vendor/typing.py issue.
+  # Can't do this because of vendor/typing.py issue.
   # log "Testing oils_for_unix.py"
   # bin/oils_for_unix.py --help | head -n 2
 

build/ovm-actions.sh

@@ -182,7 +182,7 @@ join-modules() {
 
   # Filter out comments, print the first line.
   #
-  # TODO: I don't want to depend on egrep and GNU flags on the target sytems?
+  # TODO: I don't want to depend on egrep and GNU flags on the target systems?
   # Ship this file I guess.
   egrep --no-filename --only-matching '^[a-zA-Z0-9_\.]+' $static $discovered \
     | sort | uniq

core/alloc.py

@@ -205,7 +205,7 @@ def SnipCodeString(self, left, right):
 
         Used for ALIAS expansion, which happens in the PARSER.
 
-        The argument to aliases can span multiple lines, like htis:
+        The argument to aliases can span multiple lines, like this:
 
         $ myalias '1     2     3'
         """

core/comp_ui.py

@@ -572,7 +572,7 @@ def InitReadline(
     # This determines the boundaries you get back from get_begidx() and
     # get_endidx() at completion time!
     # We could be more conservative and set it to ' ', but then cases like
-    # 'ls|w<TAB>' would try to complete the whole thing, intead of just 'w'.
+    # 'ls|w<TAB>' would try to complete the whole thing, instead of just 'w'.
     #
     # Note that this should not affect the OSH completion algorithm.  It only
     # affects what we pass back to readline and what readline displays to the

core/dev.py

@@ -57,7 +57,7 @@ class CrashDumper(object):
 
     debug info for the source?  Or does that come elsewhere?
 
-    Yeah I think you sould have two separate files.
+    Yeah I think you should have two separate files.
     - debug info for a given piece of code (needs hash)
       - this could just be the raw source files?  Does it need anything else?
       - I think it needs a hash so the VM dump can refer to it.

core/main_loop.py

@@ -297,7 +297,7 @@ def Interactive(flag, cmd_ev, c_parser, display, prompt_plugin, waiter, errfmt):
 
         # TODO: Replace this with a shell hook?  with 'trap', or it could be just
         # like command_not_found.  The hook can be 'echo $?' or something more
-        # complicated, i.e. with timetamps.
+        # complicated, i.e. with timestamps.
         if flag.print_status:
             print('STATUS\t%r' % status)
 
@@ -311,7 +311,7 @@ def Batch(cmd_ev, c_parser, errfmt, cmd_flags=0):
     Returns:
       int status, e.g. 2 on parse error
 
-    Can this be combined with interative loop?  Differences:
+    Can this be combined with interactive loop?  Differences:
 
     - Handling of parse errors.
     - Have to detect here docs at the end?

core/process.py

@@ -290,7 +290,7 @@ def _PushDup(self, fd1, blame_loc):
             fd2 = cast(redir_loc.Fd, UP_loc).fd
 
             if fd1 == fd2:
-                # The user could have asked for it to be open on descrptor 3, but open()
+                # The user could have asked for it to be open on descriptor 3, but open()
                 # already returned 3, e.g. echo 3>out.txt
                 return NO_FD
 
@@ -1260,7 +1260,7 @@ def StartPipeline(self, waiter):
 
             pid = proc.StartProcess(trace.PipelinePart)
             if i == 0 and pgid != INVALID_PGID:
-                # Mimick bash and use the PID of the first process as the group for the
+                # Mimic bash and use the PID of the first process as the group for the
                 # whole pipeline.
                 pgid = pid
 
@@ -1528,7 +1528,7 @@ def __init__(self):
         # Counter used to assign IDs to jobs. It is incremented every time a job
         # is created. Once all active jobs are done it is reset to 1. I'm not
         # sure if this reset behavior is mandated by POSIX, but other shells do
-        # it, so we mimick for the sake of compatability.
+        # it, so we mimic for the sake of compatibility.
         self.job_id = 1
 
     def AddJob(self, job):
@@ -1738,7 +1738,7 @@ def NumRunning(self):
         Used by 'wait' and 'wait -n'.
         """
         count = 0
-        for _, job in iteritems(self.jobs):  # mycpp rewite: from itervalues()
+        for _, job in iteritems(self.jobs):  # mycpp rewrite: from itervalues()
             if job.State() == job_state_e.Running:
                 count += 1
         return count

core/pyos.py

@@ -46,7 +46,7 @@ def WaitPid(waitpid_options):
         # - The arg -1 makes it like wait(), which waits for any process.
         # - WUNTRACED is necessary to get stopped jobs.  What about WCONTINUED?
         # - We don't retry on EINTR, because the 'wait' builtin should be
-        #   interruptable.
+        #   interruptible.
         # - waitpid_options can be WNOHANG
         pid, status = posix.waitpid(-1, WUNTRACED | waitpid_options)
     except OSError as e:
@@ -113,7 +113,7 @@ def ReadLine():
     # type: () -> str
     """Read a line from stdin.
 
-    This is a SLOW PYTHON implementation taht calls read(0, 1) too many times.  I
+    This is a SLOW PYTHON implementation that calls read(0, 1) too many times.  I
     tried to write libc.stdin_readline() which uses the getline() function, but
     somehow that makes spec/oil-builtins.test.sh fail.  We use Python's
     f.readline() in frontend/reader.py FileLineReader with f == stdin.
@@ -362,8 +362,8 @@ def TakePendingSignals(self):
         # `self.pending_signals`. In the worst case the signal handler might write to
         # `new_queue` and the corresponding trap handler won't get executed
         # until the main loop calls this function again.
-        # NOTE: It's important to distinguish between signal-saftey an
-        # thread-saftey here. Signals run in the same process context as the main
+        # NOTE: It's important to distinguish between signal-safety an
+        # thread-safety here. Signals run in the same process context as the main
         # loop, while concurrent threads do not and would have to worry about
         # cache-coherence and instruction reordering.
         new_queue = []  #  type: List[int]

core/shell.py

@@ -885,7 +885,7 @@ def Main(lang, arg_r, environ, login_shell, loader, readline):
         process.InitInteractiveShell()  # Set signal handlers
 
         # The interactive shell leads a process group which controls the terminal.
-        # It MUST give up the termianl afterward, otherwise we get SIGTTIN /
+        # It MUST give up the terminal afterward, otherwise we get SIGTTIN /
         # SIGTTOU bugs.
         with process.ctx_TerminalControl(job_control, errfmt):
 

core/state.py

@@ -402,7 +402,7 @@ def PushEval(self):
         node = self._MakeOutputNode()
         self.result_stack = [node]
 
-        self.output = None  # remove last reuslt
+        self.output = None  # remove last result
 
     def PopEval(self):
         # type: () -> None
@@ -826,7 +826,7 @@ def ShowShoptOptions(self, opt_names):
         # type: (List[str]) -> None
         """For 'shopt -p'."""
 
-        # Respect option gropus.
+        # Respect option groups.
         opt_nums = []  # type: List[int]
         for opt_name in opt_names:
             opt_group = consts.OptionGroupNum(opt_name)
@@ -1981,7 +1981,7 @@ def InternalSetGlobal(self, name, new_val):
 
     def GetValue(self, name, which_scopes=scope_e.Shopt):
         # type: (str, scope_t) -> value_t
-        """Used by the WordEvaluator, ArithEvalutor, ysh/expr_eval.py, etc.
+        """Used by the WordEvaluator, ArithEvaluator, ysh/expr_eval.py, etc.
 
         TODO:
         - Many of these should be value.Int, not value.Str

cpp/fanos_shared.h

@@ -5,7 +5,7 @@
 //
 // This library is shared between cpp/ and pyext/.
 
-// Callers should initalize
+// Callers should initialize
 //   FanosError to { 0, NULL }, and
 //   FanosResult to { NULL, FANOS_INVALID_LEN }
 

cpp/frontend_pyreadline.h

@@ -5,13 +5,13 @@
 
 #include "mycpp/runtime.h"
 
-// hacky foward decl
+// hacky forward decl
 namespace completion {
 class ReadlineCallback;
 Str* ExecuteReadlineCallback(ReadlineCallback*, Str*, int);
 }  // namespace completion
 
-// hacky foward decl
+// hacky forward decl
 namespace comp_ui {
 class _IDisplay;
 void ExecutePrintCandidates(_IDisplay*, Str*, List<Str*>*, int);

cpp/stdlib.cc

@@ -122,7 +122,7 @@ void execve(Str* argv0, List<Str*>* argv, Dict<Str*, Str*>* environ) {
     throw Alloc<OSError>(errno);
   }
 
-  // ::execve() never returns on succcess
+  // ::execve() never returns on success
   FAIL(kShouldNotGetHere);
 }
 
@@ -188,7 +188,7 @@ time_t time() {
 
 // NOTE(Jesse): time_t is specified to be an arithmetic type by C++. On most
 // systems it's a 64-bit integer.  64 bits is used because 32 will overflow in
-// 2038.  Someone on a comittee somewhere thought of that when moving to
+// 2038.  Someone on a committee somewhere thought of that when moving to
 // 64-bit architectures to prevent breaking ABI again; on 32-bit systems it's
 // usually 32 bits.  Point being, using anything but the time_t typedef here
 // could (unlikely, but possible) produce weird behavior.

data_lang/demo.sh

@@ -84,7 +84,7 @@ test-programs() {
   ls --escape
   echo
   # Test out error message
-  # It's basicallly correct, but ugly.  There are too many segments, and
+  # It's basically correct, but ugly.  There are too many segments, and
   # there's an unnecessary leading ''.
   # QSN is shorter and more consistent.
 
@@ -141,7 +141,7 @@ test-errors() {
   grep z "$byte_then_char" || true
   grep z "$char_then_byte" || true
 
-  # python doens't print it smehow?
+  # python doesn't print it somehow?
   banner 'python'
   # BUG: Python prints terminal sequences
   #python "$BOLD" || true

demo/complete.sh

@@ -39,7 +39,7 @@ audit() {
 
   echo
   echo --
-  # Search for special complation var usage
+  # Search for special completion var usage
   grep -E --color 'COMP_[A-Z]+' $file
 
   echo

demo/coproc.sh

@@ -55,7 +55,7 @@ simple-demo() {
   # In ksh or zsh, the pipes to and from the co-process are accessed with >&p
   # and <&p.
   # But in bash, the file descriptors of the pipe from the co-process and the
-  # other pipe to the co-proccess are returned in the $COPROC array
+  # other pipe to the co-process are returned in the $COPROC array
   # (respectively ${COPROC[0]} and ${COPROC[1]}.
 
   argv ${COPROC[@]}

demo/old/benchmarks-oheap.sh

@@ -37,7 +37,7 @@ encode-all() {
     $0 encode-one
 }
 
-# Out of curiousity, compress oheap and originals.
+# Out of curiosity, compress oheap and originals.
 
 compress-oheap() {
   local c_dir=$BASE_DIR/oheap-compressed

demo/old/gen_oheap_cpp.py

@@ -21,7 +21,7 @@
 address space.  If 1, then we have 16MiB of code.  If 4, then we have 64 MiB.
 
 Everything is decoded on the fly, or is a char*, which I don't think has to be
-aligned (because the natural alignment woudl be 1 byte anyway.)
+aligned (because the natural alignment would be 1 byte anyway.)
 """
 from __future__ import print_function
 

demo/old/old_code.py

@@ -68,14 +68,14 @@ def _Led(self, token):
 #
 
 
-# Possible optmization for later:
+# Possible optimization for later:
 def _TreeCount(tree_word):
   """Count output size for allocation purposes.
 
   We can count the number of words expanded into, and the max number of parts
   in a word.
 
-  Every word can have a differnt number of parts, e.g. -{'a'b,c}- expands into
+  Every word can have a different number of parts, e.g. -{'a'b,c}- expands into
   words of 4 parts, then 3 parts.
   """
   # TODO: Copy the structure of _BraceExpand and _BraceExpandOne.