Ok, here's a fairly significant point posted on IRC.
<TimToady> wayland76: the point of using $*CWD would be (and would *have* to
be, given how context vars work) to give each thread its own working
directory, independent of the process as a whole
Now, given that chdir is an OS concept that applies to the process as
a whole, rather than individual threads, but $*CWD is designed for individual
threads, that pretty much means that it is a really bad idea to have $*CWD do
chdir() for us.
Now, having said that, I think the idea of a per-thread CWD is a great
idea, and I'd be in favour of using $*CWD over whatever getcwd retrns in Perl6
as much as possible. So I think it would be cool if all relative paths
specified via p{} were relative to $*CWD, not getcwd(). So something like
this:
chdir("/usr/bin");
$*CWD = "/home/username";
$path = p{Music};
say $path.canonpath();
# prints "/home/username/Music"
My question is, what else uses chdir/getcwd style paths? I know
qqx/$cmd @args[]/
...will pass its current dir to the child. Maybe we could have Perl6
do something like this pseudocode "under the hood" when someone does a qqx.
olddir = getcwd();
chdir($*CWD);
exec(qqx string here);
chdir(olddir);
That would mean that $*CWD would effectively act as a per-thread
working directory.
No doubt someone will write in and tell me how bad an idea this is. I
look forward to having my ignorance exposed :).
---------------------------------------------------------------------
| Name: Tim Nelson | Because the Creator is, |
| E-mail: wayl...@wayland.id.au | I am |
---------------------------------------------------------------------
----BEGIN GEEK CODE BLOCK----
Version 3.12
GCS d+++ s+: a- C++$ U+++$ P+++$ L+++ E- W+ N+ w--- V-
PE(+) Y+>++ PGP->+++ R(+) !tv b++ DI++++ D G+ e++>++++ h! y-
-----END GEEK CODE BLOCK-----
