[Python-Dev] hg verify warnings

[Tim Peters]

> ...
> $ hg verify
> repository uses revlog format 1
> checking changesets
> checking manifests
> crosschecking files in changesets and manifests
> checking files
> warning: copy source of 'Modules/_threadmodule.c' not in parents of 60ad83716733
> warning: copy source of 'Objects/bytesobject.c' not in parents of 64bb1d258322
> warning: copy source of 'Objects/stringobject.c' not in parents of 357e268e7c5f
> 9799 files, 79660 changesets, 176851 total revisions
> 3 warnings encountered!
>
>
> $ hg --version
> Mercurial Distributed SCM (version 2.3.2)
> (see http://mercurial.selenic.com for more information)

FYI, I found this kind of warning in my own (non-Python) repository,
created from scratch just a few weeks ago, so it's NOT necessarily the
case that this has something to do with using ancient hg releases.  I
was teaching myself hg at the time, and suspect I did something hg
didn't expect <0.5 wink>.  Here's a tiny repository that displays the
same kind of thing:

Make a new repository (this is on Windows - does it matter? doubt it):

C:\>hg init HHH

C:\>cd HHH

Hg is up to date (2.7):

C:\HHH>hg version
Mercurial Distributed SCM (version 2.7)
(see http://mercurial.selenic.com for more information)
...

Make a subdirectory and add a file:

C:\HHH>mkdir sub

C:\HHH>cd sub

C:\HHH\sub>echo a > a.txt

C:\HHH\sub>hg add
adding a.txt

C:\HHH\sub>hg ci -m initial

Move the file up a level:

C:\HHH\sub>hg move a.txt ..

Now here's the funky part!  Unsure about what was going on in my own
repository, so doing as little as possible per step, I committed the
move in _2_ steps (do a plain "hg ci" at this point and nothing goes
wrong).  So first I recorded that the subdirectory a.txt is gone:

C:\HHH\sub>hg ci a.txt -m moving

Then up to the parent directory, and commit the new location of a.txt:

C:\HHH\sub>cd ..

C:\HHH>hg st
A a.txt

C:\HHH>hg ci -m moving

Now verify -v complains:

C:\HHH>hg verify -v
repository uses revlog format 1
checking changesets
checking manifests
crosschecking files in changesets and manifests
checking files
warning: copy source of 'a.txt' not in parents of 9c2205c187bf
2 files, 3 changesets, 2 total revisions
1 warnings encountered!

What the warning says seems to me to be true:

C:\HHH>hg log
changeset:   2:9c2205c187bf
tag:         tip
user:        Tim Peters <tim at python.org>
date:        Mon Aug 19 12:24:43 2013 -0500
summary:     moving

changeset:   1:60fffa9b0194
user:        Tim Peters <tim at python.org>
date:        Mon Aug 19 12:24:26 2013 -0500
summary:     moving

changeset:   0:0193842498ab
user:        Tim Peters <tim at python.org>
date:        Mon Aug 19 12:24:05 2013 -0500
summary:     initial

The parent of 2 (9c2205c187bf) is 1 (60fffa9b0194), and indeed the
copy _source_ (HHH\sub\a.txt) was removed by changeset 1.  Why that
could be "bad" escapes me, though.

Regardless, I suspect Python's warnings came from similarly overly
elaborate learning-curve workflow, and are harmless.