Re: MD5 and URL validation (continue to other very old thread)

On 23/11/2012 4:20 p.m., Eliezer Croitoru wrote:
> The last post I wrote with the very very long diff seems to be the one.

You referring to the email of 24th Oct? You said that was not quite
ready and I can see a lot of formatting bits with missing lines which

> But since there was a change in the helpers and other stuff in the
> code I now so to speak have a working MODEL for in memory cache well
> tested.
>
> It has been like that for a about one month.
> When this TCP_SWAP_FAIL stroked me I had to read a lot of code again,
> then patch etc.
>
> What I hope for now is to first make the FAKE helper code in place
> with all the config stuff in place so I wouldnt have to work on it
> again every time there is a change in the code.
> (I rewrote *some* code about 5 times to match updates)

I'm not very keen the idea of having the visible parts of it present but
not working. Whatever goes into trunk needs to perform as documented -
omitting a bundled helper and marking some store types as not-tested-yet
is fine. But the Squid portion needs to at least be usable when it goes
public.

We can break the whole project changes down into a few individual steps.

1) a patch that only does all those url() --> originalUrl() changes.
That will shrink down the amount of work a lot and I think can be done
right now, yes?

2) review of the feature patch to see if there are any other bits like
above, or if the whole thing is ready.

3) release notes and config documentation.

Amos
Received on Fri Nov 23 2012 - 05:29:34 MST