Re: Re: PHP script for viewing subversion repositories

Chuck Hagenbuch <[email protected]>
Newsgroups gmane.comp.horde.chora
Message-ID <[email protected]>
Quoting "Jay Freeman (saurik)" <[email protected]>:

> Why exactly do you want to have to make it a seperate "set the repository"
> call rather than just passing it to the object's constructor (which is how
> the standard mode of operation would be were this C++ or Java)? The key
> example I'm looking at is CVSLib_Directory.

Eh, I was thinking it'd be cleaner for cases like CVSLib_File. I guess it's not
that important, as long as it's encapsulated in the CVSLib/SVNLib calls

(ie, passing $this from CVSLib::getFileObject() is fine; doing $foo =
$VC->queryDir($VC, $where) is just silly.

> And then there's CVSLib_File. Were the cache to work as "first construct an
> object, then pass a key for the object along with a reference to that object
> to some cache storage object" then the course of action would be easy: I
> construct the object with a repository reference (without actually saving
> the repository reference in the object) and then save the object to the
> cache. After saving it I set it's repository reference and start to use it.
> Later, when I want the object again, I use it's key to recover it from the
> cache (or construct it using the previous algorithm if I can't find it and I
> get back a null reference) and then set its repository pointer. However, the
> cache system is doing something that I find just downright scary, hehe... it
> works by taking a string argument which is some kind of code snippet that it
> runs to construct the object? Isn't that largely redundant? The only usage I
> could see for that is if someone later wants that object and it got purged
> from the cache and it needs to get recreated automatically, but in that
> scenario it wouldn't have any of the other objects it needs to reconstruct
> it anyway, and seemingly the person calling the cache function has to pass
> the construction code in again anyway, in which case they might as well just
> not get back an object at all, create the object, and then donate it back to
> the cache...

Because our Cache system has a driver for the Zend performance suite, and that's
how the ZPS API works.

-chuck

--
Charles Hagenbuch, <[email protected]>
They're just looking at a wall of meat.

-- 
Chora mailing list
Frequently Asked Questions: http://horde.org/faq/
To unsubscribe, mail: [email protected]
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.