amend: reintroduce prefetching

Summary:
Add back prefetch optimization with a manual prefetch before the loop over `samefile()`.

Previously I had built this into `base.status(wctx)`, but I decided I didn't like that since it incurs a "real" diff operation that otherwise might not be needed. If you are amending to a large commit, you shouldn't need to diff the commit's changes.

Reviewed By: zzl0

Differential Revision: D66302022

fbshipit-source-id: 7d83e89f29ae627243133ba2e3d1e0dbc00574e6

eden/scm/sapling/cmdutil.py

@@ -3919,10 +3919,18 @@ def _amend(ui, repo, wctx, old, extra, opts, matcher):
             # deleted), old X must be preserved.
             with perftrace.trace("Prune files reverted by amend"):
                 statusmanifest = wctx.buildstatusmanifest(status)
+
+                if "remotefilelog" in repo.requirements:
+                    # Prefetch files in `base` to avoid serial lookups.
+                    fileids = base.manifest().walkfiles(
+                        matchmod.exact("", "", status.modified + status.added)
+                    )
+                    repo.fileslog.filestore.prefetch(fileids)
+                    repo.fileslog.metadatastore.prefetch(fileids, length=1)
+
                 for f in sorted(filestoamend):
-                    if (
-                        not samefile(f, wctx, base, m1=statusmanifest)
-                        or f in status.deleted
+                    if f in status.deleted or not samefile(
+                        f, wctx, base, m1=statusmanifest
                     ):
                         files.add(f)
                     else:

eden/scm/sapling/ext/histedit.py

@@ -247,6 +247,7 @@
     extensions,
     hg,
     lock,
+    match as matchmod,
     merge as mergemod,
     mergeutil,
     mutation,
@@ -262,7 +263,6 @@
 from sapling.i18n import _
 from sapling.pycompat import range
 
-
 # pyre-fixme[11]: Annotation `pickle` is not defined as a type.
 pickle = util.pickle
 release = lock.release
@@ -680,6 +680,12 @@ def collapse(repo, first, commitopts, skipprompt=False):
     # Recompute copies (avoid recording a -> b -> a)
     copied = copies.pathcopies(base, last)
 
+    if "remotefilelog" in repo.requirements:
+        # Prefetch files in `base` to avoid serial lookups.
+        fileids = base.manifest().walkfiles(matchmod.exact("", "", files))
+        repo.fileslog.filestore.prefetch(fileids)
+        repo.fileslog.metadatastore.prefetch(fileids, length=1)
+
     # prune files which were reverted by the updates
     files = [f for f in files if not cmdutil.samefile(f, last, base)]
     # commit version of these files as defined by head

eden/scm/tests/test-amend-file-fetches.t

@@ -24,9 +24,9 @@ Make sure we minimize content fetches:
   TRACE tree_fetches: attrs=["content"] keys=["@eaa109a0"]
   TRACE file_fetches: attrs=["content"] keys=[".hgdirsync"]
   TRACE tree_fetches: attrs=["content"] keys=["a_dir@4f20beec", "b_dir@79a5124d", "c_dir@bc250767"]
+  TRACE file_fetches: attrs=["content", "header", "aux"] keys=["b_dir/b", "c_dir/c"]
+  TRACE file_fetches: attrs=["history"] length=Some(1) keys=["b_dir/b", "c_dir/c"]
   TRACE file_fetches: attrs=["content", "header"] keys=["b_dir/b"]
   TRACE file_fetches: attrs=["content", "header"] keys=["c_dir/c"]
   TRACE file_fetches: attrs=["content"] keys=[".hgdirsync"] (?)
   TRACE file_fetches: attrs=["content", "header"] keys=[".hgdirsync"]
-  TRACE file_fetches: attrs=["history"] length=Some(1) keys=["b_dir/b"]
-  TRACE file_fetches: attrs=["history"] length=Some(1) keys=["c_dir/c"]