log: make file log slow path usable on large repos
Running "hg log <pattern or directory>" on large repos took a very, very long
time because it first read ctx.files() for every commit before even starting to
process the results.
This change makes the ctx.files() check lazy, which makes the command start
producing results immediately.
$ hg init
$ echo "added file1" > file1
$ echo "another line of text" >> file1
$ echo "added file2" > file2
$ hg add file1 file2
$ hg commit -m "added file1 and file2"
$ echo "changed file1" >> file1
$ hg commit -m "changed file1"
$ hg -q log
1:dfab7f3c2efb
0:c3fa057dd86f
$ hg id
dfab7f3c2efb tip
$ hg update -C 0
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg id
c3fa057dd86f
$ echo "changed file1" >> file1
$ hg id
c3fa057dd86f+
$ hg revert --no-backup --all
reverting file1
$ hg diff
$ hg status
$ hg id
c3fa057dd86f
$ hg update
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg diff
$ hg status
$ hg id
dfab7f3c2efb tip
$ hg update -C 0
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ echo "changed file1 different" >> file1
$ hg update
merging file1
warning: conflicts during merge.
merging file1 incomplete! (edit conflicts, then use 'hg resolve --mark')
0 files updated, 0 files merged, 0 files removed, 1 files unresolved
use 'hg resolve' to retry unresolved file merges
[1]
$ hg diff --nodates
diff -r dfab7f3c2efb file1
--- a/file1
+++ b/file1
@@ -1,3 +1,7 @@
added file1
another line of text
+<<<<<<< local
+changed file1 different
+=======
changed file1
+>>>>>>> other
$ hg status
M file1
? file1.orig
$ hg id
dfab7f3c2efb+ tip
$ hg revert --no-backup --all
reverting file1
$ hg diff
$ hg status
? file1.orig
$ hg id
dfab7f3c2efb tip
$ hg revert -r tip --no-backup --all
$ hg diff
$ hg status
? file1.orig
$ hg id
dfab7f3c2efb tip
$ hg update -C
0 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ hg diff
$ hg status
? file1.orig
$ hg id
dfab7f3c2efb tip