[perl #39921] [CAGE] generated by tools/dev/mk_manifest_and_skip.pl
Will Coleda
# Please include the string: [perl #39921]
# in the subject line of all future correspondence about this issue.
# <URL: http://rt.perl.org/rt3/Ticket/Display.html?id=39921 >
in r13493, a new MANIFEST was put in place that was generated by
tools/dev/mk_manifest_and_skip.pl
This MANIFEST had a *lot* of churn; most of it seems to be due to re-
ordering!
The current MANIFEST (r13493) starts with:
RELEASE_INSTRUCTIONS []
DEPRECATED.pod []
VERSION []
CREDITS []
vtable.tbl [devel]
While the old manifest (r13492) starts with:
ABI_CHANGES []
ChangeLog []
Configure.pl []
CREDITS []
DEPRECATED.pod []
DEVELOPING []
MANIFEST []
Looking at the code, it seems that the order in the manifest is being
driven by "File::Find". So it's not a surprise that we get churn
depending on who is running the tool.
This list should be collated in some predictable fashion to avoid
this svn-churn, and to make the file easier to read/update by hand
when necessary.
Regards.
--
Will "Coke" Coleda
wi...@coleda.com
Will Coleda via RT
On Fri Jul 28 03:01:37 2006, guest wrote:
> On Mon Jul 24 06:01:38 2006, coke wrote:
> > Looking at the code, it seems that the order in the manifest is being
> > driven by "File::Find". So it's not a surprise that we get churn
> > depending on who is running the tool.
> >
> > This list should be collated in some predictable fashion to avoid
> > this svn-churn, and to make the file easier to read/update by hand
> > when necessary.
> With the attached patch, the manifest files are pushed onto an array
> during the File::Find search, and then sorted before being printed to
> MANIFEST. Also, a one for loop is not sorted which didn't used to be,
> but it might not make as much of a difference.
>
> diffstat:
> mk_manifest_and_skip.pl | 6 ++++--
> 1 files changed, 4 insertions(+), 2 deletions(-)