Hello Andrew!
Sunday February 17 2002, [EMAIL PROTECTED] (Andrew M. Bishop) wrote:
>> when I read that WWWOFFLE 2.7 is now available, I was glad to finally
>> get rid of the Purge-bug that had required regular manual intervention.
> I don't know which bug you are refering to, but I am glad that the new
> version fixed if for you.
In the announcement, one point was "Fixed purge bug with purge min-free
option". I guess that was the one I was seeing, because I am using the
min-free option.
Here, wwwoffle -purge is called once a day in a cron job, and some day
I noticed that there was a "WWWOFFLE Purge" in the syslog, but no
"WWWOFFLE Purge finished". Redirecting the screen output of wwwoffle
to a file showed that the purge was aborting somewhere in the first half
of the process, before the age profile is displayed.
When running wwwoffle -purge manually, it sometimes went through without
any problem.
I never reported this here because I just couldn't put my finger on a
possible reason. I hadn't thought of trying -d 6, though.
But, as I said, this seems to be fixed now.
> In the WWWOFFLE code there are three places where the 304 error code
> can be generated for a reply. [...]
As I already told you off-list, I found out that it is the one in line
1729 of wwwoffles.c that is giving me these problems.
Commenting out the lines 1727 through 1734 and line 1739 has at least
allowed me to use WWWOFFLE at all. And since I don't use a large browser
cache anyway, the speed loss doesn't hit me very hard.
> I don't see any way that the reply that you get can be generated by
> WWWOFFLE for the input that you give.
Hum, maybe if I show your statement to my installation of WWWOFFLE... ;-)
No, seriously, unfortunately it /does/ happen.
> I have tried it on some example pages and I don't see the problem.
The problem seems to be system, not page related, it happened with all
pages here.
Since John and TonyC have reported the same problem, it may be helpful
to find out what our systems have in common that yours doesn't. If
there's anything I can look up or try out, just let me know.
(Argh. I thought I had it narrowed down to a configuration problem by
using the one from the distribution archive, but it seems that it was
just the request-changed and request-changed-once options that prevented
WWWOFFLE from even trying to get the pages again. What would be the
absolute minimum config I could use to try again?
My settings here are request-changed=0 and request-changed-once=no, BTW.)
> The best that I can suggest is that you run the test again, but with
> the highest level of debugging enabled 'wwwoffled -d 6 -c wwwoffle.conf'.
> This will show you the full headers of the request and reply.
I've mailed the log to you off-list, but as I said, it didn't give me any
information I hadn't at least guessed before.
>> [1] BTW, IMVHO the new web pages for editing that are unusable. It takes
>> far too many clicks to get to and change the setting I want. I found
>> the old page far easier to use.
> If you prefered the old page then you are probably somebody who could
> edit the configuration file manually.
That's what I'm resorting to now. But that is far more troublesome for
me. Before, I could just get the Config Edit Page from my bookmarks,
scroll to the desired section (easy to find because comments and entry
fields look different), make the change, click Update, click Reread
Config and be done.
Now I have to log into my Linux box, become root, open wwwoffle.conf in
an editor, locate the section (more difficult because everything looks
the same[1]), make the change, save the file, call wwwoffle -config, exit
the root shell and log out. This takes far longer for me.
[1] Yes, I know this sounds stupid. I am a great fan of textmode editors
myself, but in this case I have a real problem.
> The new web pages have advantages:
I do believe you that the new pages are a great improvement for people
who feel uncomfortable editing config files directly or for new users
who don't know the config keywords. And they look like you put quite a
deal of work into them.
But what I don't understand is why you took out the old page. It was
perfect for me, certainly for others too, and I don't believe it would
have clashed with the new pages in any way.
--
|_|\ Fido: 2:244/1512
| |/ PGP: 2047/962CC1B5 1996/04/05
GCS d- s:+>:- a- C++(+++) US P+ L+ E- W++(-) N++(*) K-- w-- O++ M- V?
PS+ PE Y+ PGP++ t+ 5(+) X(+) R tv++ b+ DI- D+ G+ e h! r++(---) y*
