All the file actions stuff is still missing for now,
as is POSIX_SPAWN_SETSCHEDULER (not sure what that's
supposed to do) and POSIX_SPAWN_RESETIDS.
Implemented in userspace for now. Once there are users,
it'll likely make sense to make this a syscall for
performance reasons.
A simple test program of the form
extern char **environ;
int main(int argc, char* argv[])
{
pid_t pid;
char* args[] = { "ls", NULL };
posix_spawnp(&pid, "ls", nullptr, nullptr, args, environ);
}
works fine.
The "Reference" object is not just a counter, it also represents the
permission to map a shbuf itself.
Without this change, a shbuf could not be re-mapped by the same
process after it released all of its refs on it.
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. :^)
I booted the system on a much better screen than the one I normally use
and the variance in contrast between different icons bothered me.
Here's an attempt to fix that, while also redoing some icons that I've
wanted to redo for a while. :^)
* Port it to Core::ArgsParser - even though it doesn't support any options
at the moment, it's nice to handle --help and reject other options;
* Accept "-" to mean stdin;
* Make sure to write out all data that we read, looping if needed;
* Detect Useless Uses of Cat and print a cute warning :^)
This fixes a bug in how wc(1) would not consider certain things to be words,
and removes the unused "character" counter (the "bytes" counter is what's used
for `wc -c`).
This fixes a bug where the mode of a FIFO was reported as 001000 instead
of 0010000 (you see the difference? me nethier), and hopefully doesn't
introduce new bugs. I've left 0777 and similar in a few places, because
that is *more* readable than its symbolic version.
This is a convenient wrapper around readlink() that hides away the details
of buffers and buffer sizes, and simply returns a String. The best part is it
doesn't rely on PATH_MAX :D
It comes in two versions, for Serenity, where we can pass non-null-terminated
strings to syscalls, and where sys$readlink() returns the total link size, and
for other systems, where we have to copy out the string, and always have to do
two syscalls.
That's not how readlink() is supposed to work: it should copy as many bytes
as fit into the buffer, and return the number of bytes copied. So do that,
but add a twist: make sys$readlink() actually return the whole size, not
the number of bytes copied. We fix up this return value in userspace, to make
LibC's readlink() behave as expected, but this will also allow other code
to allocate a buffer of just the right size.
Also, avoid an extra copy of the link target.
This was showing up in Browser profiles, which is silly, so write a new
version that doesn't create a temporary String object.
There are a whole bunch of these and long-term it would be nice to find
a way to share all the very similar logic instead of duplicating it.
The more generic virtual variant is renamed to node_name() and now only
Element has tag_name(). This removes a huge amount of String ctor/dtor
churn in selector matching.
Toolchain/BuiltIt.sh expected a version of cmake which
apt did not provide for my system. Added note of the
expected version and where to find a newer version.
