commit 655b285db25226fa483323e0a4fa50fc1e603e7a
parent 8a4e54679eef4b81c1e2160528d969d528638d4d
Author: Dridi Boukelmoune <dridi.boukelmoune@gmail.com>
Date: Tue, 1 Oct 2024 18:03:30 +0200
cmd/hare: Print dependencies to a t
The new 'hare deps -t' option changes the output to a plain text one.
This format can easily be consumed by field-based programs such as
tsort(1) (to a t!) or the more frequent utilities like sh(1) (while
read loops) or awk(1).
To illustrate how it renders, here is the output for the hare-compress
package, as of a4bb204305ef67a39dc1af3e8e843a2a4f5bee45:
$ hare deps -Dst ~/src/hare-compress/
compress::flate endian
compress::flate errors
compress::flate io
compress::gzip bufio
compress::gzip bytes
compress::gzip compress::flate
compress::gzip endian
compress::gzip errors
compress::gzip hash
compress::gzip hash::crc32
compress::gzip io
compress::gzip memio
compress::gzip time
compress::zlib compress::flate
compress::zlib endian
compress::zlib errors
compress::zlib hash
compress::zlib hash::adler32
compress::zlib io
The first column gives the effective list of modules provided by this
package. The second column gives required modules, once we filter out
the provided modules.
There is no compress module, because there are no sources. Adding such a
source file without depending on other modules would slightly change the
output:
$ cat >~/src/hare-compress/compress/version.ha <<EOF
export const version_string = "hare-compress x.y.z";
EOF
$ hare deps -Dst ~/src/hare-compress/
compress -
compress::flate endian
compress::flate errors
compress::flate io
compress::gzip bufio
compress::gzip bytes
compress::gzip compress::flate
compress::gzip endian
compress::gzip errors
compress::gzip hash
compress::gzip hash::crc32
compress::gzip io
compress::gzip memio
compress::gzip time
compress::zlib compress::flate
compress::zlib endian
compress::zlib errors
compress::zlib hash
compress::zlib hash::adler32
compress::zlib io
Note the presence of a 'compress -' line expressing the presence of
the module and its lack of dependencies.
The motivation is to enable module processing in package managers,
in the non-recursive mode, to derive virtual provides from the first
field and virtual requires from the second one, from the authoritative
tool chain.
Signed-off-by: Dridi Boukelmoune <dridi.boukelmoune@gmail.com>
Diffstat:
3 files changed, 20 insertions(+), 0 deletions(-)
diff --git a/cmd/hare/deps.ha b/cmd/hare/deps.ha
@@ -14,6 +14,7 @@ use sort::cmp;
type deps_fmt = enum {
DOT,
TERM,
+ TEXT,
};
type link = struct {
@@ -41,6 +42,8 @@ fn deps(name: str, cmd: *getopt::command) (void | error) = {
merge_tags(&tags, opt.1)?;
case 's' =>
submodules = true;
+ case 't' =>
+ goal = deps_fmt::TEXT;
case =>
abort();
};
@@ -99,6 +102,15 @@ fn deps(name: str, cmd: *getopt::command) (void | error) = {
};
};
fmt::println("}")!;
+ case deps_fmt::TEXT =>
+ for (let mod .. mods) {
+ if (module::gathered(&mod) && len(mod.deps) == 0) {
+ fmt::printfln("{} -", mod.name)!;
+ } else for (let dep .. mod.deps) {
+ const child = mods[dep.0];
+ fmt::printfln("{} {}", mod.name, child.name)!;
+ };
+ };
};
};
diff --git a/cmd/hare/main.ha b/cmd/hare/main.ha
@@ -49,6 +49,7 @@ const help: []getopt::help = [
('d', "print dot syntax for use with graphviz"),
('s', "recursively collect submodules"),
('T', "tagset", "set/unset build tags"),
+ ('t', "print text output for field-based parsing"),
"[path|module]",
]: []getopt::help),
("run", [
diff --git a/docs/hare-deps.1.scd b/docs/hare-deps.1.scd
@@ -36,6 +36,13 @@ characters.
directory. A path should be a source directory to reliably collect
dependencies.
+*-t*
+ Print the dependency tree as a field-based plain text output. Each
+ line has two fields, a module and one of its dependencies. If the
+ module (or one of its submodules with the *-s* option) does not have
+ dependencies, the second field contains just a dash. The text output
+ is a suitable input for *tsort*(1).
+
*-T* _tagset_
Set or unset build tags. See *BUILD TAGS* in *hare-module*(5).
