Fixes the problem reported in #3073. While trying to write a test
for this, I thought I'd use
Shell -c 'for i in $(seq 100) { echo $i }' | head -n 1
but that makes the cpu spin at 100% and doesn't terminate even
with this fix here. But at least piping disasm into head now works.
This has the nice side effect of fixing alias completion, because
cached_path is the source of truth for the completion system, and it was
only refreshed (with shell::cache_path()) in the shell's constructor,
before the rc files where loaded (ie no aliases)
This also means that shell::is_runnable can now rely on the cache, and
doesn't have to check the aliases itself.
Closes#2760.
This commit adds a 'for' loop, and tweaks the syntax slightly to make &&
bind more tightly than || (allowing for `expr && if_ok || if_bad`) :^)
This commit fixes job control by putting children in their own process
group, and proxying TTY signals to active jobs.
This also cleans up the code around builtin_disown a bit to use
the newer job interfaces.
This allows the parser to finally parse the entire source into a single
AST.
As a result of allowing comments inside sequences, Sequence is also
marked as would_execute if its left or right node would.
Take one small step towards #2357.
Handle completing barewords starting with '-' by piping the requests to
the Shell::complete_option(program_name, option) :^)
Also implements completion for a single builtin (setopt) until we figure out how
to handle #2357.
This builtin sets (and unsets) boolean flags that alter the behaviour of
the shell.
The only flags added are
- inline_exec_keep_empty_segments: Keep empty segments in the result of
splitting $(...) by $IFS
- verbose: Announce each command before executing it
It should be noted that the (rather extreme) verbosity of the names is
intentional, and will hopefully be alleviated by the next commit :^)
This commit also removes the ExecutionInputType and directly uses
RefPtr<Shell> instead, since nothing else is needed for execution
purposes, and also makes the shell refuse to evaluate commands with
any sort of syntax error.
This commit also completely reworks the execution, highlighting and
completion model to work with the new AST.
New additions:
- $(...) stdout captures
- fd>&fd redirections
- fd>&- redirections (close fd)
- read-write redirections (<> path)
- completely event-based execution
- the weird idea of allowing the user to redirect the shell's own fds
- variables in strings
- local variables
- minimal list support
- adding hyperlinks to all paths that exist
We were getting a little overly memey in some places, so let's scale
things back to business-casual.
Informal language is fine in comments, commits and debug logs,
but let's keep the runtime nice and presentable. :^)
Get rid of the weird old signature:
- int StringType::to_int(bool& ok) const
And replace it with sensible new signature:
- Optional<int> StringType::to_int() const
This commit allows the Shell to complete paths in redirections.
A closing quote is added if the path is an unclosed quote.
```
$ foo > "foob<tab>
$ foo > "foobar"
```
And move canonicalized_path() to a static method on LexicalPath.
This is to make it clear that FileSystemPath/canonicalized_path() only
perform *lexical* canonicalization.
This fixes a bunch of FIXME's in LibLine.
Also handles the case where read() would read zero bytes in vt_dsr() and
effectively block forever by erroring out.
Fixes#2370
