New update-website.mk script commited
jhuntwork at linuxfromscratch.org
Sat Aug 7 20:45:47 PDT 2004
On Sat, 2004-08-07 at 23:19 -0400, Anderson Lizardo wrote:
> Hi Jeremy,
> The makefile script is commited. I didn't test the rsync'ing yet; that's why
> it's commented out. Can I test it with your server now? Just asking in case
> you're doing some experiments with it :)
Anderson, go ahead, whenever you're ready.
> After we confirm the rsyncing is working, we can enable the post-commit hook
> (but still using the TARGETDIR on test.linuxfromscratch.org). If it's working
> OK, we backup the current site and copy the tree from
> test.linuxfromscratch.org to the definitive location , so the migration will
> be almostly atomic to the visitors.
Wait a sec... I'm confused here. I thought we were focusing on getting
the current site rsynced - I didn't think you were working on the
Care to clarify for me? What's this about moving
> Still on TODO:
> - add code to update the mirrors list automatically, disabling mirrors not
> responding to the rsync push but that were supposedly enabled to accept push.
> - See the comments marked as TODO on the script. Although the script works
> without fixing them, it will run a lot faster if we move the update-patches
> and update-hints targets to the post-commit hook scripts of the respective
> - Possibly move the HTML validation code to a pre-commit hook script, although
> I'm not sure if that's possible at that stage (e.g. how do we check the
> content that is being commited, if it's not on the repository yet?). Keeping
> it on the post-commit script is useless as there is no way to avoid the buggy
> HTML code from being commited because it was already commited. Confusing, no?
I'm ready to help in any way you need, but I think I need to better
understand the path we're taking here. ;)
More information about the website