RE: RewriteRule breaks slave SVN commits for 1.6.x clients only

Designation: Non-SSA/Finmeccanica

Daniel Shahaf wrote:
> Sure, knowing exactly what build you used is relevant information.
>
> It's just the adjective "official" that caught my eye.

Thank you for clearing up that bit of minutia.

>> That's a solution much like cutting off a finger to get rid of a splinter. Yeah it works, but...
>>
> Disagree. Your description was that you added the RewriteRule to
> solve an alleged problem; I just told you that the alleged problem isn't a problem.

I didn't say multiple slashes were a (technical) problem. I said it would be a good practice to ensure the URLs didn't include them since they're superfluous. Yes, it's not a huge issue, but it's puzzling why this rule, which worked fine with the older 1.6.x server using both 1.6.x and 1.7.x clients, would give the 1.7.x server grief ONLY when an older 1.6.x client is used and ONLY when committing to the slave server, but not the master.

>>> (And, by the way, your RewriteRule is less than ideal; it causes a
>> repeated request for each doubled slash: http://foo/bar/////baz)
>>
>> How so? I thought the plus sign after the second slash will capture one or more slashes?
>> Could you provide a revised expression?
>
> You're right, but consider: http://host/foo//bar//baz

Yes...but in that case, repeating the request isn't a bug it's a feature. I'd want http://host/foo//bar//baz to become http://host/foo/bar/baz

>> I stated that this started happening "when I upgraded from the latest
>> 1.6.x server to the 1.7.1" one. Both master and slave servers were
>> upgraded simultaneously.
>
> You didn't mention this, AFAIK.
>
> (And, BTW, master's version should always be ≥ slave's version.)

When I stated that I had upgraded from "latest 1.6.x server to the 1.7.1," I felt it was implied that I meant both servers. If that was confusing for anyone, I apologize.

3.1.1001
Received on 2011-12-20 21:29:13 CET