On May 11, 2015, at 12:51 AM, René J.V. Bertin <[email protected]> wrote:

> On Saturday May 09 2015 06:47:02 Ryan Schmidt wrote:
> 
>> I don't understand... why would you build documentation in post-activate? 
>> why is this situation so unusual that normal MacPorts procedures cannot 
>> accommodate it?
> 
> The point is that if I couple this documentation too tightly to the port it 
> documents, you get a situation where it is rebuilt each time the port is 
> updated, even if that's just a revision bump to force a rebuild against some 
> changed dependency. That's overkill which carries a cost I don't want. That 
> was one of the reasons why I initially thought of making the generation 
> dependent on an env. variable. However, if documentation is not rebuilt every 
> time, you cannot have it installed through the usual destroot, because then 
> it would not be present anymore after the previous version/build has been 
> deactivated.
> 
> It's true I could create a subport that drops part of the main ports 
> versioning, for instance so that 4.14.x.y becomes 4.14 . That's something 
> that can be captured rather easily in tcl code. Still, after looking a bit 
> more at the actual generation process, I'm beginning to think I'd be using an 
> external script anyway. I don't even know yet exactly what it will take to 
> generate the monolithic .qch file I have in mind, but it could well be 
> needlessly complex to do that in inline tcl code. And from there it's just a 
> small step to install such a script (along with the main port) so that 
> interested users can invoke it directly themselves. 

I think I follow and admire your desire to not over burden the user/system with 
superfluous builds.

Regards,
Bradley Giesbrecht (pixilla)

Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail

_______________________________________________
macports-dev mailing list
[email protected]
https://lists.macosforge.org/mailman/listinfo/macports-dev

Reply via email to