On Sat, May 13, 2017 at 07:55:44PM +1200, Evan Hanson wrote:
> The first changes the way explicitly-namespaced symbols are handled
> (that is, symbols with a module name directly included, like "foo#bar"),
> making them behave essentially like qualified symbols as far as variable
> resolution is
Hi Peter,
Looks good, applied.
Thanks for noticing that. I don't remember where that original list of
exports came from, but I'd always just assumed it was correct without
ever checking against the "real" API.
Best regards,
Evan
signature.asc
Description: PGP signature
---
tcp.scm | 152
1 file changed, 75 insertions(+), 77 deletions(-)
diff --git a/tcp.scm b/tcp.scm
index 8e3d0e4c..a806ac99 100644
--- a/tcp.scm
+++ b/tcp.scm
@@ -165,30 +165,28 @@ EOF
(define-foreign-variable _invalid_socket int
This adds an "escape hatch" to variable resolution for namespaced
symbols (e.g. `foo#bar`), allowing them to be used across module
boundaries just like qualified symbols.
This is done by simple string scanning of the identifier's name, which
is emphatically not ideal as it means the compiler has
---
continuation.scm | 4
1 file changed, 4 deletions(-)
diff --git a/continuation.scm b/continuation.scm
index e8cc4ff0..d1cb29de 100644
--- a/continuation.scm
+++ b/continuation.scm
@@ -30,10 +30,6 @@
(foreign-declare "#define C_direct_continuation(dummy) t1")
-;; XXX Is this still
Replace `##sys#read-string/port' and `##sys#read-string!/port', both of
which are defined in extras.scm but also used in srfi-4.scm, with
module-namespaced (but still internal) names.
Also, reshuffle their argument checks so that the internal procedures
trust their argument types and all the
Hi folks,
Here are a few patches that deal with symbols and namespaces.
The first changes the way explicitly-namespaced symbols are handled
(that is, symbols with a module name directly included, like "foo#bar"),
making them behave essentially like qualified symbols as far as variable
resolution