Andreas Fuchs | 24 May 2006 11:26
Gravatar

an experimental git mirror of SBCL cvs

Hi there,

yesterday I managed to get git-cvsimport to run and produce a mirror
of the SBCL repository. That mirror contains vastly more information
than the arch one (although git-cvsimport doesn't keep track of file
renames, afaict. I have tagged the repository "beta" to reflect that):
All branches (lutex-branch and HEAD in particular) are kept in sync
automatically. Still, repository size is just 47MB, and retrieving it
all via the command line below takes 2 minutes on my 2MBit line. After
that, all operations except update can happen offline.

The repository can be retrieved via

$ cg-clone http://sbcl.boinkor.net/git/sbcl-beta.git/ sbcl

if you have cogito installed.

Also, there is a nice web interface at http://sbcl.boinkor.net/gitweb/
that shows SBCL development in all its glory.

Like the arch gateway, this is meant as an experiment using
alternative VC systems for sbcl development, with emphasis on
"experiment". I would like to get file renames working, and so the
repository URL (and worse yet, the ancestry graph) is likely to change
and become obsolete in a while; so please don't yet base important
month-long branches off it.

Hope you enjoy it,
--

-- 
Andreas Fuchs, (http://|im:asf <at> |mailto:asf <at> )boinkor.net, antifuchs
(Continue reading)

Timothy Ritchey | 24 May 2006 19:38

Win32 Sockets II

Just when you thought it was safe to go back in the water...

I've attached a patch and file that reworks the sb-bsd-sockets win32  
implementation to use sb-grovel instead of the hand-coded alien  
garbage I had written before. On the plus side, this means the  
compiler doesn't yell at us every time we touch a socket. On the down  
side, the sb-bsd-sockets.asd is a bit more convoluted to address the  
whole fd->handle wrapping that needs to be done, as well as making  
sure we load the winsock library before we do our groveling. There  
may very well be a better way to organize the win32 specific changes,  
so any suggestions would be welcome.

Attachment (sb-bsd-sockets.patch): application/octet-stream, 30 KiB
Attachment (win32-lib.lisp): text/x-lisp-source, 141 bytes
Nikodemus Siivola | 30 May 2006 14:26
Gravatar

Re: Win32 Sockets II

Timothy Ritchey <tritchey <at> paragent.com> writes:

> I've attached a patch and file that reworks the sb-bsd-sockets win32
> implementation to use sb-grovel instead of the hand-coded alien
> garbage I had written before. On the plus side, this means the

Merged as 0.9.13.10, thanks!

Cheers,

  -- Nikodemus              Schemer: "Buddha is small, clean, and serious."
                   Lispnik: "Buddha is big, has hairy armpits, and laughs."

-------------------------------------------------------
All the advantages of Linux Managed Hosting--Without the Cost and Risk!
Fully trained technicians. The highest number of Red Hat certifications in
the hosting industry. Fanatical Support. Click to learn more
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=107521&bid=248729&dat=121642
Florian Weimer | 24 May 2006 22:06
Picon

Re: an experimental git mirror of SBCL cvs

* Andreas Fuchs:

> yesterday I managed to get git-cvsimport to run and produce a mirror
> of the SBCL repository. That mirror contains vastly more information
> than the arch one (although git-cvsimport doesn't keep track of file
> renames, afaict.

Neither CVS nor GIT track renames.  The GIT merge tools have
heuristics to discover renames after the fact and deal with them.
This makes importing projects into GIT much easier than with other
systems.

Do you plan to run the GIT daemon?  Those of us who already downloaded
the entire history could use git-clone --reference for the final
download, to avoid retransmitting all the (compressed and delta'd)
fulltexts.

-------------------------------------------------------
All the advantages of Linux Managed Hosting--Without the Cost and Risk!
Fully trained technicians. The highest number of Red Hat certifications in
the hosting industry. Fanatical Support. Click to learn more
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=107521&bid=248729&dat=121642
Andreas Fuchs | 25 May 2006 18:52
Gravatar

Re: an experimental git mirror of SBCL cvs

On 2006-05-24, Florian Weimer <fw <at> deneb.enyo.de> wrote:
> * Andreas Fuchs:
>
>> yesterday I managed to get git-cvsimport to run and produce a
>> mirror of the SBCL repository. That mirror contains vastly more
>> information than the arch one (although git-cvsimport doesn't keep
>> track of file renames, afaict.
>
> Neither CVS nor GIT track renames.  The GIT merge tools have
> heuristics to discover renames after the fact and deal with them.
> This makes importing projects into GIT much easier than with other
> systems.

Ah, I see. That was unclear, and I have since read up on that. Sounds
useful.

> Do you plan to run the GIT daemon?  Those of us who already
> downloaded the entire history could use git-clone --reference for
> the final download, to avoid retransmitting all the (compressed and
> delta'd) fulltexts.

I tried to get it working, but couldn't. My time and patience for
software with no error reporting is limited at the moment, so I
probably won't try again for a few weeks.

BTW, I found that the http backend (at least on 1.3.2) works with
git-clone --reference, too, so I don't think that the gitd is
necessary at all. (phew) (:

Good luck,
(Continue reading)

Florian Weimer | 26 May 2006 19:45
Picon

Re: an experimental git mirror of SBCL cvs

* Andreas Fuchs:

>> Do you plan to run the GIT daemon?  Those of us who already
>> downloaded the entire history could use git-clone --reference for
>> the final download, to avoid retransmitting all the (compressed and
>> delta'd) fulltexts.
>
> I tried to get it working, but couldn't. My time and patience for
> software with no error reporting is limited at the moment, so I
> probably won't try again for a few weeks.

Uh-oh.  I must admit I haven't tried to install it myself.

> BTW, I found that the http backend (at least on 1.3.2) works with
> git-clone --reference, too, so I don't think that the gitd is
> necessary at all. (phew) (:

You need to be very careful, then, in the way you layout the packs.
Otherwise you end up downloading a huge pack just because you need a
few commit objects in it.  (The GIT daemon generates suitable packs on
the fly, IIRC, so it's not affected by this issue.)

-------------------------------------------------------
All the advantages of Linux Managed Hosting--Without the Cost and Risk!
Fully trained technicians. The highest number of Red Hat certifications in
the hosting industry. Fanatical Support. Click to learn more
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=107521&bid=248729&dat=121642
Andreas Fuchs | 26 May 2006 22:29
Gravatar

Re: an experimental git mirror of SBCL cvs

Today, Florian Weimer <fw <at> deneb.enyo.de> wrote:
>> I tried to get it working, but couldn't. My time and patience for
>> software with no error reporting is limited at the moment, so I
>> probably won't try again for a few weeks.
>
> Uh-oh.  I must admit I haven't tried to install it myself.

I'm sure there is a way, I just haven't found it yet.

>> BTW, I found that the http backend (at least on 1.3.2) works with
>> git-clone --reference, too, so I don't think that the gitd is
>> necessary at all. (phew) (:
>
> You need to be very careful, then, in the way you layout the packs.
> Otherwise you end up downloading a huge pack just because you need a
> few commit objects in it.  (The GIT daemon generates suitable packs
> on the fly, IIRC, so it's not affected by this issue.)

Right. At the moment, there is only one pack, which leads to a huge
initial download; I hear the recommended pack size is 5MB; we'll see
how this will work out. As I said before, it's experimental, so I'll
experiment (-:

Cheers,
--

-- 
Andreas Fuchs, (http://|im:asf <at> |mailto:asf <at> )boinkor.net, antifuchs

-------------------------------------------------------
All the advantages of Linux Managed Hosting--Without the Cost and Risk!
Fully trained technicians. The highest number of Red Hat certifications in
(Continue reading)


Gmane