* Rogue Branch - can git help here?
@ 2006-02-10 3:09 Don Imus
2006-02-10 11:55 ` Ralf Baechle
0 siblings, 1 reply; 2+ messages in thread
From: Don Imus @ 2006-02-10 3:09 UTC (permalink / raw)
To: linux-mips
I've got an old Linux 2.4.18 source tree I downloaded from a vendor who
sells devices using MIPS processors with embedded Linux running on them.
There are clearly places in the code where the vendor has made changes.
Unfotunately, the vendor used their own CVS and every tagged file
shows revision 1.1.1.1.
Can GIT help me determine on which date the vendor's tree was originally
pulled?
I thought of a script call GIT to diff the file against all revisions in
the repo, possibly creating patch files, and then I could look at the
number of changes for each as a measure of "closeness". Doing this for
several files and finding commonality in the dates would increase the
probability of finding the right one.
If I copy the vendor source tree into the local GIT tree and commit a
new branch are there any facilitites in GIT that would tell me which
older revision, prior to a given date, is the best "match" on the files
in the tree?
I just started using GIT and maybe the question is better for the GIT
mailing list but I figured I'd ask it here first and see if anyone has
already done something like this.
In case anyone's wondering the device ships with a binary-only device
driver module that will only work with a 2.4.18 kernel and that's why
I'm stuck in the past on this.
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: Rogue Branch - can git help here?
2006-02-10 3:09 Rogue Branch - can git help here? Don Imus
@ 2006-02-10 11:55 ` Ralf Baechle
0 siblings, 0 replies; 2+ messages in thread
From: Ralf Baechle @ 2006-02-10 11:55 UTC (permalink / raw)
To: Don Imus; +Cc: linux-mips
On Thu, Feb 09, 2006 at 10:09:40PM -0500, Don Imus wrote:
> I've got an old Linux 2.4.18 source tree I downloaded from a vendor who
> sells devices using MIPS processors with embedded Linux running on them.
>
> There are clearly places in the code where the vendor has made changes.
> Unfotunately, the vendor used their own CVS and every tagged file
> shows revision 1.1.1.1.
>
> Can GIT help me determine on which date the vendor's tree was originally
> pulled?
>
> I thought of a script call GIT to diff the file against all revisions in
> the repo, possibly creating patch files, and then I could look at the
> number of changes for each as a measure of "closeness". Doing this for
> several files and finding commonality in the dates would increase the
> probability of finding the right one.
>
> If I copy the vendor source tree into the local GIT tree and commit a
> new branch are there any facilitites in GIT that would tell me which
> older revision, prior to a given date, is the best "match" on the files
> in the tree?
Nothing straight of the shelf, unfortunately.
> I just started using GIT and maybe the question is better for the GIT
> mailing list but I figured I'd ask it here first and see if anyone has
> already done something like this.
>
> In case anyone's wondering the device ships with a binary-only device
> driver module that will only work with a 2.4.18 kernel and that's why
> I'm stuck in the past on this.
The crude ad-hoc method would be to write a quick shell scripts that
creates a diff between every revision between linux-2.4.18..linux-2.4.19
but that would be slow. And it's a just too common problem (I last faced
it less than a week ago ...) so may deserve a better solution in git than
just an ad-hoc script, so I suggest you indeed take this to the git
mailing list.
Ralf
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2006-02-10 11:49 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-02-10 3:09 Rogue Branch - can git help here? Don Imus
2006-02-10 11:55 ` Ralf Baechle
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.