On 18 Jan 2012, at 4:56 AM, Noel Butler wrote:

> This never was a problem in 2.2, if one disabled dav it was disabled, as it 
> should be disabled, fully, not only in parts, here and there, granted it's 
> now changed because modules are no longer defaulted to statically built, but 
> plenty of admins will use this method, I also favour it (heh, obviously since 
> I found it), so if the option to revert to old method is offered as it is via 
> --enable-mods-static=all, then if someone disables a specific mod like dav 
> (which some security type scanners suggest), it should be fully disabled and 
> the build should succeed.

This was and still is a problem in v2.2 for mod_proxy, and a number of other 
modules. If you turn mod_proxy off, you'll have the build fail caused by 
mod_proxy_http and friends. I looked at fixing this recently and discovered 
this was a massive job, one for the timescales given by v2.6, not v2.4.

What modules should be doing is using the optional functions API, and changing 
this requires an MMN bump, so cannot be done during the lifetime of v2.4. While 
unfortunate, this isn't a showstopper.

> Without going over all docs as Tim (IIRC) and William appeared to do, the 
> INSTALL is the most important I think for users, the file has references to 
> 2.3, this is 2.4, also, it does not mention the need for the deps or apr is 
> no longer included in the base package, these are fundamental issues that 
> need to be addressed, especially the later, remember, your average apache 
> user is NOT a coder, but a system admin, and often will little experience on 
> doing anything else but copy and pasting a config and make and install from a 
> cheat sheet.

The average apache user will be installing binaries, either provided by their 
operating system, or provided by us (such as the Windows binaries). The more 
specialised users will be building Apache from source against system provided 
copies of APR. Only users with very specific requirements will be attempting to 
use the deps package, and these users will be of the level that seeing "2.3" 
instead of "2.4" won't be an insurmountable barrier. Again, this is 
unfortunate, but not a showstopper.

Regards,
Graham
--

Attachment: smime.p7s
Description: S/MIME cryptographic signature

Reply via email to