develooper Front page | perl.perl5.porters | Postings from December 2008

Re: Perl.git "So, Tomorrow?" Release

Thread Previous | Thread Next
From:
Nicholas Clark
Date:
December 4, 2008 07:22
Subject:
Re: Perl.git "So, Tomorrow?" Release
Message ID:
20081204152221.GX49335@plum.flirble.org
On Tue, Nov 18, 2008 at 06:20:01PM +1300, Sam Vilain wrote:

> * Change 11243:
> 
>   - Change 11243 contained files which had a file type of XXX - "mac"
>     Sorry, I didn't do any .gitattributes conversion for these either.
>     It would probably need more infrastructure - currently the
>     'smudge' and 'clean' filters are just that - stream filters - and
>     there is no infrastructure for unpacking single repository files
>     to multiple working copy files (or vice versa, a possible useful
>     feature for storing collections of, eg tar archives).
> 
>    => WONTFIX

On Fri, Nov 21, 2008 at 10:35:34PM +0000, Nicholas Clark wrote:

[snip questions about revisions 42 and 83]

> I've not tested the macperl branches yet.

Right. Having read what perforce does with file type "apple", I think that
"WONTFIX" is actually the right thing.

The git conversion has the data and resource forks of mac files encoded
together using "AppleDouble" encoding*

The current OS X client will sync files of type "apple" as files with regular
data and a real resource fork. This is nice, and clean, but isn't useful for
validating the conversion.

The current *nix clients will sync files of type "apple" into two client
files - the data fork is in the file with the original name, say 'clash', and
the resource fork is written to a file named '%clash' in the same directory.
[Oh the irony. meta bot gave me 'clash', and "potential file name clash" is
a flaw I see in this]

'%clash' is stored in AppleDouble encoding.

It happens that all five files of type "apple" have an empty data fork.
However, I think I updated my conversion validator (attached) such that it
will cope with a non-empty data fork.

Anyway, I updated the conversion validator to take a git checkout, split the
single file in AppleDouble into two files, one for data, one for the
resources, redone as AppleSingle. (They're almost the same, but of course
all the internal file offsets change because there is no 12 byte header for
the data fork). And the conversion of the 3 macperl branches is byte-for-byte
perfect.

Thanks for all your hard work.

Now there's only the small matter of 5.8.9-RC2, sitting hoping that no-one
finds anything for a week (but in turn hoping that this is because there is
nothing to find, not because no-one is looking), release, and then
stop-the-world;sync from perforce;declare git master;restart-the-world.

Nicholas Clark

syncer program attached.

* This was useful: http://users.phg-online.de/tk/netatalk/doc/Apple/v1/


Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About