Patches md5sums

Bruce Dubbs bdubbs at
Thu Nov 3 16:16:03 PST 2005

Archaic wrote:
> On Thu, Nov 03, 2005 at 05:10:40PM -0600, Bruce Dubbs wrote:
>>(2) have the render script extract the needed ones from there and put
>>them into the blfs/downloads/svn/ directory.  At the same time, it
>>could do the check for presence.
> I'm not following this part here. If (3) below is implemented, why does
> there need to be two sets of patches? As it currently stands there
> exists{lfs,blfs,hlfs} directories which are already
> mirrored. Inside those dirs are version dirs (and BLFS-5.1 exists in
> there, too, so it was used at sometime). The result is still having a
> single dir for all patches of a given book, and uniformity across the
> projects. BTW, it is Manuel's xslt scripts that populate the lfs and
> hlfs directories. That xslt script is also used to copy the released
> book's set of patches. It's all automated, and any missing patches cause
> the render script to spit out an email.

The exact directory is not that important.  Changing it would only
require a single entity change to the book. is
OK with me.  The implementation method you mention is fundamentally the
same as we are doing it now.

The issue of making sure the patches are current is a bit more difficult
in BLFS than LFS due to BLFS having more patches (118 vs 26 right now)
and because we don't have a master list on a single page like LFS does.
 I think we had it automated for a presence check before LFS (could be
wrong about that tho).   The major problem was that BLFS editors didn't
have commit privs to the patches svn.  That proplem is no longer present.

In any case, we can migrate our process to the way LFS is doing it.
Right now it's just a matter of hammering out the details and making it

>>For the mirrors, (3) I believe a script based on the extraction script
>>in the rendering could also extract the needed patches and place them in
>>the appropriate directories.

I wasn't aware that the mirrors were syncing patches from belgarath.  If
that is so, then (3) would not be necessary.

  -- Bruce

More information about the blfs-dev mailing list