Re: [PATCH] A Perforce importer for git.

From: Alex Riesen <raa.lkml@gmail.com>
Date: Jun 2, 2006, 13:09 UTC
Message-ID: <81b0412b0606020609o24ee34b4p7b1bcd013136d49a@mail.gmail.com>
In-Reply-To: <BAYC1-PASMTP117796558F3B42B2C18DBCAE980@CEZ.ICE>

On 5/25/06, Sean <seanlkml@sympatico.ca> wrote:

Show 6 quoted lines

>
> Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
> ---
>  Documentation/git-p4import.txt |  165 ++++++++++++++++++
>  git-p4import.py                |  357 ++++++++++++++++++++++++++++++++++++++++
>  2 files changed, 522 insertions(+), 0 deletions(-)

BTW, can I suggest to import the _currently_ synced state?

The reason is that because of the way how Perforce is done its working directories (views, aka clients) are often a horrible mix of occasionally synced files to some random versions. No one actually uses "p4 sync" for whole project here where I work, because it is absolutely useless at this level (updated files have abolutely no relevance at head revision, which is what you get by syncing without strictly specifying a revision). So a working state is stitched together from a lot of "mappings": perforce path to local path -> revision. That state can be actually worked on (up until you have to commit something, that is not possible except on head).

That is why I actually can't sync before importing, but have to use the method Junio described for CVS: p4 sync files...#revisions git add new-files... git commit -a (simplified. There are problems with case-sensitivity and that stupid read-only attribute perfoce uses as sync marker).