[HCoop-Help] diagnosing a 400 Bad Request

[Peter Gammie]

On 15/05/2008, at 12:27 AM, Adam Chlipala wrote:

> Peter Gammie wrote:
>> I'm having trouble translating a .htaccess file into domtool-speak.
>>
>> You can find my attempt in ~peteg/.domtool/peteg.org for the domain
>> hope.peteg.org.
>>
>> According to domtool-admin the rewrites are the same as in my
>> reference .htaccess file, with the sole addition of double quotes.  
>> The
>> other flags seem similar enough too.
>
> I can't help you figure out which Apache configuration you want in the
> end, but, if you can suggest changes to Domtool or kinds of Apache
> configuration you wish you could generate but currently can't, I can
> implement your suggestions.  To get to that point, if the silence on
> this list for your thread continues, you might look for help in Apache
> forums.

Thanks. I was fishing for some explanation of how to diagnose a 400  
Bad Request error. Apparently nothing gets logged in error.log (see  
below) and the rewrite log is quite out of date.

I don't expect you or anyone in particular to provide this sort of  
help, I just thought there might be a HCoop admin who is strong in  
Apache fu.

>> One problem I'm having here is the delay between making a request and
>> seeing it in the error/access log. I wonder if there is a way you can
>> give us timelier access to the logs for debugging purposes?
>
> This is on my to-do list, though I certainly wouldn't mind if someone
> else did it.  Our wiki pages should at least make clear that log files
> are updated infrequently, I hope, so that it isn't surprising, just
> annoying, how things are working.

Yes, I am unsurprised. For my usual purposes the logs are largely  
irrelevant, but while debugging domtool setups, timely feedback is  
very helpful.

So, perhaps a bug: my rewrite.log only shows data for two days ago. So  
either the 400 Bad Request means mod_rewrite doesn't log anything or  
HCoop's logging is not working. I'll log a bug in the tracker when I  
have a better idea of what's going on there.

The problem is most certainly in those rules. The aim is simply to  
remove the script name from the URL. With the rules omitted, this works:

http://hope.peteg.org/hope.cgi/

> If someone else implements a log-tailer that works securely in our
> set-up, and if I can audit the source code, I will install it.

If I had a few more spare lives, I would be that person. Sorry!

cheers
peter


_______________________________________________
HCoop-Help mailing list
HCoop-Help at lists.hcoop.net
https://lists.hcoop.net/listinfo/hcoop-help