rejetto forum
Software => HFS ~ HTTP File Server => Beta => Topic started by: rejetto on January 25, 2008, 03:04:30 PM
-
this bug urged to be fixed, sorry
download @ www.dovedove.it/hfs/hfs177.exe
what's new
- missing ":" in host, broken in build #176
-
Seems to me that limits due inactivity is broken. As i set the time to 30 sec, people that idle longer then 30 sec are not cut off.
anyone more with that problem ?
-
Turn Off Presistent Connection
Your issue will solve hopefully
i had the same problem too :)
Menu > Other Option > Presistent Connections
this shouldnt have a Check
-
a strange behavior tested on Indian's attack...
for some users (without the good login and pass) trying to log with a mail address,
the disconnected connection don't leave directly after they are gone.
you need to select "leave disconnected connections", unselected it, and select it again to kick off the slot.
@ character is not disconnected in the list until you leave disconnected connections the 2 time
( it can be useful to catch the offered mail and return a message)
if you can catch the pass code used in the historic list, to make a "mail hack party" it could be fun. ;D
check to log this one "hellorejetto@yahoo.com"
never disconnected from the list
it smell the hfs178.exe thank you the Indians attack for the bug check
also a "log off" function in the menu should be useful to not have to close the explorer after this test, and reset the login still in the cache.
-
Turn Off Presistent Connection
Your issue will solve hopefully
i had the same problem too :)
Menu > Other Option > Presistent Connections
this shouldnt have a Check
Yes that was the solution, how dumb of me should have known that.
but thnx anyway
-
When using a mail address as username, HFS gets stuck on 'replying'. For that reason the connection is not dropped. Even if it eventually disconnects it is not removed from the list.
-
Caché building template in HFS?
There is a small problem when hfs builds the page using external language files. In my mainpage visitors can choose form different languages, and loading the language-files works well. But switching from one language to another, sometimes (1 out of 3 or 4), in some sections the variables (from special:strings) are not changed. So for example, if i switch from german to english, 1 or 2 sections remain in german, until i force a second refresh to reload the page.
It's not question of the browser, because the html generated contains that mix of languages. It seems that hfs does caching of some sections (in my case [login-link] and [footer] ) that were already built before. Is this a bug or is this done to save build-time?
If it's not a bug, could we set could we set in [special:begin] some sentence to be recognized by hfs like:
{{if | {{=|%folder%|/|}}| reset}}
being reset (or someting like clearcache) a keyword for hfs so that it clears de caché and does a full rebuild of the whole template?
I prefer that this would be considered as a bug, because not rebuilding the whole template everytime would only save some microsecond or less.
-
bacter, could your problem be the same at www.rejetto.com/forum/?topic=5620 ?
-
When using a mail address as username, HFS gets stuck on 'replying'. For that reason the connection is not dropped. Even if it eventually disconnects it is not removed from the list.
fixed in next beta
(same problem reported by jerome)
-
Seems to me that limits due inactivity is broken. As i set the time to 30 sec, people that idle longer then 30 sec are not cut off.
anyone more with that problem ?
i just tested setting the limit to 5 seconds
6.00.53 127.0.0.1:61310 Connected
6.00.53 127.0.0.1:61310 Requested GET /util/
6.00.53 127.0.0.1:61310 Served 5,13 KB
6.00.59 127.0.0.1:61310 Disconnected by server: inactivity - 16791 bytes sent
all went well.
Are you sure that connections were truly inactive for 30 seconds? Any activity will reset the timer.
If you are sure, then Log menu -> Log what -> enable everything from "Connections" to "Bytes sent".
Then make a test connection, be idle for more than the time you set the limit, and see what the log says.
If the you think you were right, and there's a bug, then post your log on the forum.
Note: disabling persistant connections is a workaround to your problem, but it should be left ON when the problem is solved. When it is off, there's more network activity and delay.
-
Thanks rejetto,
the problem mentioned above with supposed caching of sections is not more present in build #178! Now it works o.k.