hook/reloc: fix up2k jank

* wark landed in the wrong registry when moved to another volume
   (harmless; upload would succeed on the next handshake)

* dedup did not apply correctly when moved into another volume,
   since all the checks were done based on the previous vol;
   fix this by recursing the whole thing

also update the reloc example after some real-world experience

Reported-by: @daniiooo

bin/hooks/reloc-by-ext.py

@@ -2,11 +2,12 @@
 
 import json
 import os
+import re
 import sys
 
 
 _ = r"""
-relocate/redirect incoming uploads according to file extension
+relocate/redirect incoming uploads according to file extension or name
 
 example usage as global config:
     --xbu j,c1,bin/hooks/reloc-by-ext.py
@@ -52,11 +53,19 @@ def main():
     try:
         fn, ext = fn.rsplit(".", 1)
     except:
-        # no file extension; abort
+        # no file extension; pretend it's "bin"
-        return
+        ext = "bin"
 
     ext = ext.lower()
 
+    # this function must end by printing the action to perform;
+    # that's handled by the print(json.dumps(... at the bottom
+    #
+    # the action can contain the following keys:
+    # "vp" is the folder URL to move the upload to,
+    # "ap" is the filesystem-path to move it to (but "vp" is safer),
+    # "fn" overrides the final filename to use
+
     ##
     ## some example actions to take; pick one by
     ## selecting it inside the print at the end:
@@ -80,11 +89,37 @@ def main():
     elif ext in MUSIC.split():
         by_category = {"vp": "/just/tunes"}
     else:
-        by_category = {}
+        by_category = {}  # no action
 
-    # now choose the effect to apply; can be any of these:
+    # now choose the default effect to apply; can be any of these:
     # into_subfolder  into_toplevel  into_sibling  by_category
-    effect = into_subfolder
+    effect = {"vp": "/junk"}
+
+    ##
+    ## but we can keep going, adding more speicifc rules
+    ## which can take precedence, replacing the fallback
+    ## effect we just specified:
+    ##
+
+    fn = fn.lower()  # lowercase filename to make this easier
+
+    if "screenshot" in fn:
+        effect = {"vp": "/ss"}
+    if "mpv_" in fn:
+        effect = {"vp": "/anishots"}
+    elif "debian" in fn or "biebian" in fn:
+        effect = {"vp": "/linux-ISOs"}
+    elif re.search(r"ep(isode |\.)?[0-9]", fn):
+        effect = {"vp": "/podcasts"}
+
+    # regex lets you grab a part of the matching
+    # text and use that in the upload path:
+    m = re.search(r"\b(op|ed)([^a-z]|$)", fn)
+    if m:
+        # the regex matched; use "anime-op" or "anime-ed"
+        effect = {"vp": "/anime-" + m[1]}
+
+    # aaand DO IT
     print(json.dumps({"reloc": effect}))
 
 

copyparty/up2k.py

@@ -2564,7 +2564,10 @@ class Up2k(object):
 
         return ret
 
-    def _handle_json(self, cj: dict[str, Any]) -> dict[str, Any]:
+    def _handle_json(self, cj: dict[str, Any], depth: int = 1) -> dict[str, Any]:
+        if depth > 16:
+            raise Pebkac(500, "too many xbu relocs, giving up")
+
         ptop = cj["ptop"]
         if not self.register_vpath(ptop, cj["vcfg"]):
             if ptop not in self.registry:
@@ -2794,12 +2797,16 @@ class Up2k(object):
                             if hr.get("reloc"):
                                 x = pathmod(self.asrv.vfs, dst, vp, hr["reloc"])
                                 if x:
+                                    zvfs = vfs
                                     pdir, _, job["name"], (vfs, rem) = x
                                     dst = os.path.join(pdir, job["name"])
+                                    job["vcfg"] = vfs.flags
                                     job["ptop"] = vfs.realpath
                                     job["vtop"] = vfs.vpath
                                     job["prel"] = rem
-                                    bos.makedirs(pdir)
+                                    if zvfs.vpath != vfs.vpath:
+                                        self.log("xbu reloc %d..." % (depth,), 6)
+                                        return self._handle_json(job, depth + 1)
 
                         job["name"] = self._untaken(pdir, job, now)
 
@@ -2880,7 +2887,9 @@ class Up2k(object):
                         lut.add(k)
 
                 try:
-                    self._new_upload(job)
+                    ret = self._new_upload(job, vfs, depth)
+                    if ret:
+                        return ret  # xbu recursed
                 except:
                     self.registry[job["ptop"]].pop(job["wark"], None)
                     raise
@@ -4176,22 +4185,17 @@ class Up2k(object):
 
         return ret
 
-    def _new_upload(self, job: dict[str, Any]) -> None:
+    def _new_upload(self, job: dict[str, Any], vfs: VFS, depth: int) -> dict[str, str]:
         pdir = djoin(job["ptop"], job["prel"])
         if not job["size"]:
             try:
                 inf = bos.stat(djoin(pdir, job["name"]))
                 if stat.S_ISREG(inf.st_mode):
                     job["lmod"] = inf.st_size
-                    return
+                    return {}
             except:
                 pass
 
-        self.registry[job["ptop"]][job["wark"]] = job
-        job["name"] = self._untaken(pdir, job, job["t0"])
-        # if len(job["name"].split(".")) > 8:
-        #    raise Exception("aaa")
-
         xbu = self.flags[job["ptop"]].get("xbu")
         ap_chk = djoin(pdir, job["name"])
         vp_chk = djoin(job["vtop"], job["prel"], job["name"])
@@ -4220,10 +4224,18 @@ class Up2k(object):
             if hr.get("reloc"):
                 x = pathmod(self.asrv.vfs, ap_chk, vp_chk, hr["reloc"])
                 if x:
+                    zvfs = vfs
                     pdir, _, job["name"], (vfs, rem) = x
+                    job["vcfg"] = vfs.flags
                     job["ptop"] = vfs.realpath
                     job["vtop"] = vfs.vpath
                     job["prel"] = rem
+                    if zvfs.vpath != vfs.vpath:
+                        self.log("xbu reloc %d..." % (depth,), 6)
+                        return self._handle_json(job, depth + 1)
+
+        job["name"] = self._untaken(pdir, job, job["t0"])
+        self.registry[job["ptop"]][job["wark"]] = job
 
         tnam = job["name"] + ".PARTIAL"
         if self.args.dotpart:
@@ -4233,7 +4245,7 @@ class Up2k(object):
             job["tnam"] = tnam
             if not job["hash"]:
                 del self.registry[job["ptop"]][job["wark"]]
-            return
+            return {}
 
         if self.args.plain_ip:
             dip = job["addr"].replace(":", ".")
@@ -4293,6 +4305,8 @@ class Up2k(object):
         if not job["hash"]:
             self._finish_upload(job["ptop"], job["wark"])
 
+        return {}
+
     def _snapshot(self) -> None:
         slp = self.args.snap_wri
         if not slp or self.args.no_snap: