Is this change live/definite yet? Just updated to the latest K-9 Mail
version from F-Droid, and it is very clearly still using the K-9 branding,
package name, links, etc. The only thing that appears to have changed in
the app is the source code link (points to Thunderbird's GitHub).

Alternative thought off the wall: what if we stole
comptime from Zig? It is complex but a variation
could be used to specify complex behavior
in-language as opposed to passing on AST
parsing etc to the user and external tools.

On Fri Aug 2, 2024 at 12:03 PM EDT, Hugo Machet wrote:
> > Unfortunately, the test command now crashes when trying to use the
> > 
> > `grep` subcommand. Seems like iterating the directory doesn't work
> > 
> > anymore?
>
>
> I think the new way to do it since zig 0.12 is
>
> var dir = try fs.cwd().openDir(".", .{ .iterate = true });

Ah, thanks. Didn't catch that. Fixed in v2.

---
v1 -> v2: Fixed iteration crash, added std.fs.Dir.OpenOptions param

 .gitignore    |  2 ++
 README.md     |  2 +-
 build.zig     | 13 +++++++------
 build.zig.zon | 11 +++++++++++
 cmd.zig       | 13 +++++++------
 5 files changed, 28 insertions(+), 13 deletions(-)
 create mode 100644 .gitignore

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..3389c86
[message trimmed]

---
 .gitignore    |  2 ++
 README.md     |  2 +-
 build.zig     | 13 +++++++------
 build.zig.zon | 11 +++++++++++
 cmd.zig       | 13 +++++++------
 5 files changed, 28 insertions(+), 13 deletions(-)
 create mode 100644 .gitignore

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..3389c86
--- /dev/null
+++ b/.gitignore
[message trimmed]

Also related to this ticket: https://todo.sr.ht/~sircmpwn/hare/862

-1 -- if the user selects `-q`, I think they should only see output that
is explicitly from the program with `?:`.

If this patch is to be added, though, then it should at least be
modified to print the final result to stderr so that it can be filtered
away on the command line.

Neat, but I don't think these will catch on ultimately. And we
definitely don't need two identical registers each for I/O.
---
 src/modal.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/src/modal.c b/src/modal.c
index 3fd2094..5ec3139 100644
--- a/src/modal.c
+++ b/src/modal.c
@@ -111,10 +111,12 @@ write_reg(char r, char *reg)
 {
 	char c, *cap = walk(reg);
 	switch(r) {
[message trimmed]

---
 src/modal.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/src/modal.c b/src/modal.c
index e4f5b6f..8016af9 100644
--- a/src/modal.c
+++ b/src/modal.c
@@ -13,6 +13,17 @@ static char bank_a[0x4000], *src_ = bank_a;
 static char bank_b[0x4000], *dst_ = bank_b;
 static char *regs[0x100];
 
+static const int hex[128] = {
+	-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /*00*/
[message trimmed]

This commit spiritually reverts 392e05d.

With debug messages and interpreter state sent to stderr and the actual
output of the program sent to stdout, it is easier to read both
independently. For instance, one can redirect stdout or stderr to
another terminal window to keep an eye on both outputs independently:

        bin/modal examples/hello.modal >/dev/pts/5
---
 src/modal.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/src/modal.c b/src/modal.c
index e9b54fe..fe3767e 100644
[message trimmed]