Rst wrote:
>Re: your patch for the AllowOverride problem --- I'm not sure I understand it.
>In particular, it isn't clear why you're just merging the override portion
>of the current directory's config, rather than the whole thing (is it to
>continue to give whatever else is there the ability to override any .htaccess
>file, or is there some other reason?).
Yes. I could do
new_config = merge (previous config, access.conf directory config)
if (new_config->allowed_override)
{
htaccess_config = parse_htaccess()
new_config = merge(previous config, htaccess_config)
new_config = merge(new_config, access.conf directory config)
}
As you say, it's so <Directory> directives override .htaccess files.
However, I thought that merging the entire configuration lists just
to get at the new value of AllowOverride was inefficient.
>[. I'm also not sure of the purpose of some of the changes you made to
> directory_walk --- for instance, the first chunk of your patch shuffles
> around the assignment to the variable core_dir, but it *looks* like it
> winds up with the same value as it had in the base code. Again, I have
> the awkward feeling I'm missing something... ]
Sorry about that; I rearranged a few lines when I was trying to understand
the code; I moved a few assigments to be closer to where the values
were used.
David.
P.S. I think this bug is important to not go unfixed in the next release.
>Re: your patch for the AllowOverride problem --- I'm not sure I understand it.
>In particular, it isn't clear why you're just merging the override portion
>of the current directory's config, rather than the whole thing (is it to
>continue to give whatever else is there the ability to override any .htaccess
>file, or is there some other reason?).
Yes. I could do
new_config = merge (previous config, access.conf directory config)
if (new_config->allowed_override)
{
htaccess_config = parse_htaccess()
new_config = merge(previous config, htaccess_config)
new_config = merge(new_config, access.conf directory config)
}
As you say, it's so <Directory> directives override .htaccess files.
However, I thought that merging the entire configuration lists just
to get at the new value of AllowOverride was inefficient.
>[. I'm also not sure of the purpose of some of the changes you made to
> directory_walk --- for instance, the first chunk of your patch shuffles
> around the assignment to the variable core_dir, but it *looks* like it
> winds up with the same value as it had in the base code. Again, I have
> the awkward feeling I'm missing something... ]
Sorry about that; I rearranged a few lines when I was trying to understand
the code; I moved a few assigments to be closer to where the values
were used.
David.
P.S. I think this bug is important to not go unfixed in the next release.