currently build of synocli-file fails due to #5463

Hi @hgy59
I've tested building from your branch hgy59:update_rmlint and manually applied patch for #5463 by @th0ma7 .

The build completes and packages the .spk , when I test on my DS-1821 . rmlint --version shows the new version 2.10.1 .
But when trying to use the handlers 'clone' or 'relink' I'm still getting

WARNING: reflink will not be emitted: please compile with blkid and gio-unix-2.0.

WARNING: reflink will not be emitted: please compile with blkid and gio-unix-2.0.

can you please check whether compiled with: contains +mounts ?
The warning is only displayed when one of the options HAVE_BLKID and HAVE_GIO_UNIX is missing, and mounts is only active when both are definded.

rmlint --version
version 2.10.1 compiled: Oct 21 2022 at [16:12:32] "Ludicrous Lemur" (rev 79c294ee)
compiled with: +mounts -nonstripped +fiemap +sha512 +bigfiles +intl -replay +xattr +btrfs-support

rmlint was written by Christopher <sahib> Pahl and Daniel <SeeSpotRun> Thomas.
The code at https://github.com/sahib/rmlint is licensed under the terms of the GPLv3.

    } features[] = {{.name = "mounts",         .enabled = HAVE_BLKID & HAVE_GIO_UNIX},
                    {.name = "nonstripped",    .enabled = HAVE_LIBELF},
                    {.name = "fiemap",         .enabled = HAVE_FIEMAP},
                    {.name = "sha512",         .enabled = HAVE_SHA512},
                    {.name = "bigfiles",       .enabled = HAVE_BIGFILES},
                    {.name = "intl",           .enabled = HAVE_LIBINTL},
                    {.name = "replay",         .enabled = HAVE_JSON_GLIB},
                    {.name = "xattr",          .enabled = HAVE_XATTR},
                    {.name = "btrfs-support",  .enabled = HAVE_BTRFS_H},
                    {.name = NULL,             .enabled = 0}};
    /* clang-format on */

So now only the replay feature is disabled, as we build without glib-json.

BTW: to build a package for rmlint only, you can build it in spksrc/diyspk/rmlint (diyspk = do it yourself spk). That is much faster and does not need the rust patch

Thanks @hgy59
I built diyspk/rmlint as suggested instead of the full spk/synocli-file .
And rmlint -c sh:handler:clone and handler:reflink are now working 🥳

Hopefully your PR can be included into the next release of synocli-file

Also I made a patch for rmlint/lib/format/sh.c.in as upstream the original code actually has the checks for reflink warning commented out and so always prints the warning when the clone or reflink handler is used, Please feel free to use

--- lib/formats/sh.c.in.orig    2020-06-13 19:12:45.000000000 +0930
+++ lib/formats/sh.c.in 2022-10-27 22:22:57.623108880 +1030
@@ -218,9 +218,9 @@
 };
 
 static void rm_sh_warn_if_reflink_not_compiled_in(void) {
-#if !HAVE_BLKID || !HAVE_GIO_UNIX
+  if(!HAVE_BLKID || !HAVE_GIO_UNIX) {
     g_printerr("\n%sWARNING:%s reflink will not be emitted: please compile with blkid and gio-unix-2.0.\n", YELLOW, RESET);
-#endif
+  }
 }
 
 static void rm_sh_parse_handlers(RmFmtHandlerShScript *self, const char *handler_cfg) {

@t-rutter reading your patch i doubt that there is an error in the source code.
Please regard that #if !HAVE_BLKID || !HAVE_GIO_UNIX it not commented out, but #if and #endif are a c preprocessing directives.
The WARNING output is removed at compile time if the condition evaluates to 0.

I built diyspk/rmlint as suggested instead of the full spk/synocli-file .
And rmlint -c sh:handler:clone and handler:reflink are now working

@t-rutter
Just found that in the context of synocli-file package, libblkid is missing.
This is caused by two packages compiling util-linux from source (cross/libuuid and cross/libblkid). After util-linux is compiled for libuuid the second build (for libblkid) is ignored, as the build folder is the same.
I am working on it...