currently using zoneedit for registrars DNS to vhosts on my server though I am running webmin-bind-virtualmin

CheapWebHosting.co.nz

Can someone answer me the question if i can connect my apache with an active directory 2000 server through mod_authnz_ldap
Guys i can run the script now but -testcgi but when i run it via web it tells me to download the file i have tried

then expand your regex in the setenvifnocase
the thing in the double quotes is the regex

ok i got another plan... anyway better for future problems....

nope, just if you don't know what's left out and what's not, I have no idea what's suddenly happening.. if You had said http://svn/ everything would've been perfectly fine
my problem is, that I don't get why the redirect is happening.

as i said...
if you access http://host/dir it 301 to http://host/dir/

ex. SetEnvIfNoCase blah_blah_dn "(cn=blah, email=blah|cn=foo, email=foo)" valid_user=1 ... blah blah blah

and the svn client is stupid, he drops the last / i gave him, no idea why
and then drops on the 301
totally nuts

I owe you another bear

you might want to anchor that regex... "^(... | ...)$" ...

err, Beer
over the last year I think that's a 6pack now

eh?

hmmmmmmmm

is your repository user windows? so tortoise?

let's try this:

oook...

AcceptPathINfo

AcceptPathINfo is http://httpd.apache.org/docs/2.2/mod/core.html#acceptpathinfo

aehm....
On or Off?

I can checkout svn co https://svn.brainsware.org/ocb but I cannot check out like this: http://svn.brainsware.org/ocb/ [there's a redirect to https]
on

no result
on both

hm?

Off and On brings the same result
the result we dont want
very funny that you can do this....
i just thought about that it works in https but not http
would be funny

Yes.. it's hardly bearable how much I rule...

lol
ok i do now the other setup.....
location block for any svn with macro
i anyway need dedicated user rights....
and i guess i go home now, too much work makes getty evil

hrm... what would really be cool is if you could just do SSLVerifyClient require Subject="CN=foo, email=foo, ..."
something like that...

yeah
was hoping on such but your suggestions should get really damn close

patches welcome...

Fix it yourself, loser!

can anyone point me to documentation about running apache from an NFS share?

what documentation?

The documentation is your friend. Read the documentation. It will make you happy. (http://httpd.apache.org/docs/) or http://httpd.apache.org/docs-2.0/

any documentation

apache doesn't care if it's on an nfs share or not
it's just files in a file system

as in, does it work, is there issues? things to consider?
thats not the case with everything

wouldn't it be the same as running it on the local filesystem? depending on your perms and stuff of course...
nfs

lock files and whatnot can be an issue

nfs is sucks sweaty monkey balls or see EnableSendfile and EnableMMAP

EnableSendfile

enablesendfile

a way to disable sendfile() for when you serve off a NFS/SMB share. More info at http://httpd.apache.org/docs/2.0/mod/core.html#enablesendfile. Works badly with some peoples windows systems: when in doubt, disable it. Broken with IPv6 on Linux in most configurations.
a way to disable sendfile() for when you serve off a NFS/SMB share. More info at a href="http://httpd.apache.org/docs/2.0/mod/core.html#enablesendfile"http://httpd.apache.org/docs/2.0/mod/core.html#enablesendfile/a. Works badly with some peoples windows systems: when in doubt, disable it. Broken with IPv6 on Linux in most configurations.

enablemmap

enablemmap is http://httpd.apache.org/docs-2.0/mod/core.html#enablemmap

those are the only issues i know of... and that's serving content on nfs shares...

serving content and config files and logs is what i would want to do
although i suppose logs wont work
grrr

is there a possibility to get all files from a directory that i cannot acces by http://whatever/dir/ because of error 403, but i am able to acces http://whatever/dir/special.file
it would be a little bit tricky to try http://whatever/dir/[a-z]{0-9}.suffix

files

files is http://httpd.apache.org/docs-2.2/mod/core.html#files

havocologe what are you asking?

i am trying to find a method to get files out of a directory that is not allowing to be indexed

what do you mean "get files out"

for dirname/ i get an error 403 but for dirname/filename i get no error but can acces the file

fajita directoryindex?

directoryindex is http://httpd.apache.org/docs/2.0/mod/mod_dir.html#directoryindex or http://httpd.apache.org/docs/mod/mod_dir.html#directoryindex

my guess it's an Options Indexes being off...
and...
step one

whatever the problem, step one is to look in the error log (and any other logs that may apply, such as suexec, mod_rewrite, or mod_security).

considering there's 2 types of 403 causes...

i am not on the server side. how to set my server hosting up to work like i described is a simple thing. i hoped there would be something to work arround as client

there isn't

deliberately, in fact.

well you could just allow indexes to be shown...

if you choose not to, that implies you'd want clients not to be able to list files

normaly i could download every file with wget -r http:/whatever/ but this time that does not work wired..

that requires that the file listing be available
if the admin has disabled file indexing, it means they don't want you to be able to do that.
take it up with the admin

in that case...

mh. so i will have to get wget up to try every possible file name ..
long but should work

that sounds like brute force

plus you couldn't get any newly created files there either...
*shrugs*

trying every possible filename with about 8 letters.
it is only a dir with *.js files
so that should be possible

ah... brute force dictionary attack maybe?
long... painful... and still no guarantee you get every file in there...
*shrugs*

hi, how do i rewrite a single url?

redirect

redirect is http://httpd.apache.org/docs-2.0/mod/mod_alias.html#redirect or http://httpd.apache.org/docs/mod/mod_alias.html#redirect or see 'Redirect Context'

does apache support virtual ssl domains each with its own ssl certificates hosting key?

(split into 3 lines) RewriteEngine on RewriteCond %{REMOTE_ADDR} (.)* RewriteRule ^/split\.html$ /allowed.html [R] -- am I doing anything wrong (note the .* is for testing)
I am still able to open up split, when I should be forwarded to allowed

I have apache set up with several domains and each has its own SSLCertificateFile and SSLCertificateKeyFile directives but it still serves up the same ssl key for all domains

ssl vhosts

When using SSL, each virtual host must have either its own IP address or its own port. Or both. or for details see http://httpd.apache.org/docs-2.0/ssl/ssl_faq.html#vhosts2 or http://www.onlamp.com/pub/a/apache/2005/02/17/apacheckbk.html or see SNI or consider wildcard certificates or http://wiki.cacert.org/wiki/VhostTaskForce or see ip based vhosts

this causes weird prompts in ie and firefox

how can I serve a /webdav directory to public (Allow from all) but allow dav access to the same directory only for authenticated users?

not wierd... expected prompts
limitexcept

limitexcept is http://httpd.apache.org/docs/2.2/mod/core.html#limitexcept or http://httpd.apache.org/docs/mod/core.html#limitexcept

did you include RewriteEngine On ?

yeah, I did

cool - thanks!

or i guess...
limit

limit is http://httpd.apache.org/docs/2.2/mod/core.html#limit

This can't be right. it seems absurd. What is the purpose of having virtual domains if you can't handle them separately

no... i think limitexcept is what you want...
it's the ssl protocol
and it is right
problem is the host information happens to be encrypted...

privet, Apollo78.

there are some fixes in apache2.2 which will allow the use of tls, but that's subject to browser compatibility...

no sorry

but ssl, doesn't deal with this... never has...

arggggh

and there is ip address host based hosting which is how right now you'd do multiple hosts with different certs...

virtual domains are kind of a hack, anyway

i tried to install ocsinventory and after installation i am trying to restart apache and getting the following error synty

this is so irritating

configuration or is it sufficient to put it only in one vhost?

should be sufficient in the http vhost

life's like that sometimes

well tls is the way to go for that...
but older browsers don't support tls...

with some caveats

so basically there is no way to do name-based ssl certificates hosting virtual hosts?

and i'm not so sure that the tls actually works in mod_ssl... .

it means the request is not encrypted
you can do it if they all share a cert

well there is... use different ports
:P

syntax error on line 16 of /etc/apache2/conf.d/ocsinventory.conf invalid command perlsetenv! Any idea, please?

or wild card certs...

please link to wild card cert docs

that works fine if all your domain host names are {something}.you.com

yeah
that's the only caveat

yep, but it works fine. we do that.

what if my domains are all different?

you're SOL

then you need separate ip numbers

hmmmmm

if you think about it, it makes sense

or I tell my customers that is just the way apache is

the ip connection doesn't know what domain name you used.

they all have to accept the error

it's not apache... it's the ssl protocol
it's not apache... it's the ssl host protocol

in order to get that into the request, you have to send it as part of the request

how come I have never seen this with IIS

however, ssl provides an encrypted connection
you need the cert to construct the connection
but the domain name is part of the request, so it's not available when you're constructing a connection

are there any plans to fix this with newer versions of ssl?

not in ssl, that i'm aware of... again, tls is supposed to do this

there's a proposal to use STARTTLS in apache
that allows the initial part of the request to be sent unencrypted and then upgraded to an encrypted connection

do all the major browsers support tls???

yeah, smtp allows that, and ftp

yes
but it's still a proposal

hmmm
anyone know how long I have to wait?

so does ldap, for that matter

well, newer versions of the major browsers

yup

new features

new features is http://httpd.apache.org/docs-2.0/new_features_2_0.html or http://www.onlamp.com/pub/a/apache/2004/06/17/apacheckbk.html or for ssl http://httpd.apache.org/docs-2.0/ssl/ssl_compat.html http://httpd.apache.org/docs/2.2/new_features_2_0.html http://httpd.apache.org/docs/2.2/new_features_2_2.html

but the difference is that http is a stateless protocol

mod_ssl

mod_ssl is http://httpd.apache.org/docs-2.0/ssl/ or http://www.modssl.org/ (for 1.3) or http://httpd.apache.org/docs-2.0/mod/mod_ssl.html or http://cvs.apache.org/~mads/ac2004/MO18mod_ssl.pdf or see mod_gnutls for an alternative

Added a support for RFC 2817, which allows connections to upgrade from clear text to TLS encryption.
that's from the new features for 2.2

all those other ones involve a persistent verified connection
anyway, you can try it with apache 2.2

i actually haven't played with it correctly... i'll need to do that...
but i can't seem to test it with openssl s_client....

but using separate ip numbers and a connection encrypted from the start is probably more secure

although, according to the browser... TLS v1.0 128 bit ARC4 (1024 bit RSA/SHA)
*shrugs*

i don't think s_client handles rfc 2817
i could be wrong

i dunno... there's the --tls1 flag
seems odd to have that flag if it didn't support it...
i could be mistaken on it's exact purpose though...

rfc2817 is how you get http to switch to tls in the middle of a connection

hrm...

--tls1 just uses tls 1.0 instead of ssl 3.0 to make a connection
rfc2817 assumes you start with a cleartext connection, then send a StartTls header, or some such,
and then upgrade the connection.

so doing something like openssl s_client -connect blah.com:80 --tls1 still wouldn't work... hrm...

i don't remember all the details

okey dokey then...

this has the potential for leaking secure information in the initial request
so i've never used it.
maybe i was just being too paranoid.

heh
never a bad thing really

anyway, we can use a wildcard cert, so we do.

although, it depends on what's being communicated...
if the only thing sent is the host info to use, i don't see it being that insecure of a protocol to use...

the file ownership on /var/www is owner: www-data group: www-data. But when using a ftpserver to load files directly into my /var/www the ftp server needs read/write permissions, should I just add ftpuser to the www-data group?

no
you shouldn't even have done that to your docroot

or chown the whole /var/www directory to owner: ftpuser group: ftpgroup?

you should've just changed the owner of the directory where the uploads were to take place
and you need mod_umask
mod_umask

mod_umask is http://www.outoforder.cc/projects/apache/mod_umask/ and is loaded by LoadModule umask_module modules/mod_umask.so

ok to recap if for example i have /var/www/somewebsite/ i chown the somewebsite folder?

i personally chgrp it
but that's just me...
chown should work fine without having to use mod_umask...
oh wait... you were talking about ftp to that dir, huh....

ok will give this a go... yes was talking about ftp into /var/www/somewebsite...

but that's some client ftping directly to the dir, right?
or are you using something like mod_ftp?

yes some client ftping into directory... but I have setup proftp to use mysql so the users should be ftpuser ftpgroup...

mod_ftp would require some kind of chown (if it supported write), regular ftp would require probably some group change on the dir as well as adding the user to that group... unless you wanted to add everyone to group www-data
then i would chgrp the dir to ftpgroup

what info is at risk in your assessment? Is it really any more info than wireshark would provide? As long as data transmission is protected, source and destination don't seem that valuable

and leave ownership to www-data?

only if you wanted apache to write to that dir
you might also want to set the group sticky bit on that dir so all new files created will be owned by ftpgroup

doesn't apache need write access to use php?

hi folks,

hi folks, is there a good apache gui

i have a problem with apache2 and webdav
heres my site-config:
http://apache.pastebin.ca/562351

only if your php requires writing there...

ok.

is your ftp client some web based thing?

thanks all who replied to my ssl/tls question

the webdav dir is writeable from www-data and files are displayed correctly on client side

problem is lots of ways to go with that ftp thing... depends on what you're using...

but I'can't create and delete files with an logged in user, via webdav
what am I doing wrong?

im not sure if i should ask this question here... Here is the problem: my cgi scripts are not executed..(rather im able to read it) but my php scripts work fine

step one

whatever the problem, step one is to look in the error log (and any other logs that may apply, such as suexec, mod_rewrite, or mod_security).

i have mod_perl installed

so if /var/www is owned by www-data group www-data but somewebsite directory is owned by ftpuser:ftpgroup apache will still serv?

define serve... you mean read access? you'd have to set read/exec access to others for apache to serve
and having /var/www owned by the apache user is just plain bad

think i got it...
will go and tinker ...
Thx for info, ditto fajita...

so this isnt a apche problem?
s/apche/apache/

i don't use mod_perl, so i couldn't tell you... apparently no one else here seems to either...

http://apache.pastebin.ca/562370

o.O

i guess i should try at #perl

http://www.google.com/search?hl=en&client=opera&rls=en&hs=3wB&q=apache+Could+not+open+property+database.&btnG=Search
1st hit says check the owner of the dav lock file...
and i guess change the owner to the user apache is running as...

ok, that would be www-data probably

oh and if you're using selinux, try disabling it temporarily...

meandtheshell, nope it's a fresh debian installation, just can't find the lockfile

/var somewhere??? not sure...
but my only experience with dav is svn...

are there any web stats programs that do ip-geolocation stuff automatically and compile city/state/country stats?

maybe look in /var/www?

yep
i looked there but there's only the apache2-default site dir

hrm... maybe /var/apache2 or /var/log/apache2?
or heck, maybe in /etc/apache2?
or just use locate... :P
well, updatedb first...

i removed dir /changed owner of the empty /var/cache/apache2
sorry /var/lock/apache2
hmm, restarted and still the same message
Could not open property database.
so the file must be missing

well, /var/lock/apache2 i would imagine would only have apache's running pid...
do you have mod_dav and mod_dav_fs enabled?
ah, you have to specify the davlock
ex. DavLockDB /usr/local/apache2/var/DavLock
The directory containing the lock database file must be writable by the User and Group under which Apache is running.

yep, found out in that moment, that pointet to somewhere...
...wrong

http://httpd.apache.org/docs/2.2/mod/mod_dav.html

megaspaz, thx :-)
it's a bit slow but...
it's working :-)

good job
motd

I could calculate your chance of survival, but you won't like it...

0 - this means that this apache2 will never accept requests from other machines than 127.0.0.1,

help! i can't seem to have more than 1 proxy on a domain in apache

Whoops! That's not the name of a module.

fasta, correct

great, thanks

elaborate please

ok, thanks...just a sec
I got ProxyPass and ProxyPassReverse lines that work fine for /financial/ extension...when I tried to add another 2 lines for /wiki/ it didn't take

could you put what you have up for us to see please?
apc

Try using http://apache.pastebin.ca - It's a good pastebin, and is even set up to highlight Apache 'stuff'.

http://apache.pastebin.ca/562465

hello world
to make apache only listen for local network connections is it just "Listen 192.168.0.0/24:80"?

the /financial/ works but not /wiki/

Eclipse75, I wouldn't use a networkaddress

how the hell could you make apache on one machine listen on port 80 from another machine?

well im just setting up a http remote managed local file server

what's it doing?

i get server not found when i try to go to /wiki .....however

eh yeah i see what youre talking about now

on server i can do "lynx localhost:8200" and get it

did you restart apache after the change?

im just trying to figure out how to only allow local addresses to connect

also...if i go to /wiki/ --- with 2nd slash it works
yes

well yeah...the way you wrote the rule you need the /

you use access controls like allow, deny, order
order

order is http://httpd.apache.org/docs-2.0/mod/mod_access.html#order or http://httpd.apache.org/docs/mod/mod_access.html#order also see: http://httpd.apache.org/docs/2.2/mod/mod_authz_host.html#order

allow

allow is http://httpd.apache.org/docs-2.0/mod/mod_access.html#allow or http://httpd.apache.org/docs/mod/mod_access.html#allow http://httpd.apache.org/docs/2.2/mod/mod_authz_host.html

deny

deny is http://httpd.apache.org/docs-2.0/mod/mod_access.html#deny or http://httpd.apache.org/docs/mod/mod_access.html#deny or http://httpd.apache.org/docs/2.2/mod/mod_authz_host.html

but why /financial doesn't need 2nd slash
?

alrighty, thank you

this is all port 80 right? so we can disregard the 443 stuff you put up

correct

does /wiki give a 403?

Not Found
The requested URL /wiki was not found on this server.
i don't see the numbers 403

I think you are hitting your default vhost and not being proxypassed
for the /financial request, does it actually appear in the logs of your server listeing on port 8100?

i assume so..let me check
what is a netstat or related command to get listening ports?

netstat -nltp

netstat

netstat prints network connections, routing tables, interface statistics, masquerade connections, and multicast memberships. Use netstat -an grep -w 80 to check if port 80 is in use. (On linux, you can use netstat -pan to see which program uses the port.) (For Windows use -ano to show the state, PID, and connection in IP not name format.)

yes 8100
weird that 8100 and /financial work fine

yu see that request in its logs?
is /var/www/family the same docroot on the server listeing on 8100 as the server listening on 80?

i think you may have found the bug
even though i don't use /var/www/family/financial i had to create it....
i forgot to create /var/www/family/wiki
arreyder:

well, we still have a problem
inthat you arent really doing what you think you are doing

oh?

you are just serving them with apache on port 80, and not really proxypassing
at least when going without the / on the end

wait.....if /financial is a web app on 8100 then apache must be proxying to 8100 right?

because you are sharing the documentroot it works
nah, /financial isnotproxied, /financial/ is

yes.../financial gets converted to /financial/home

umm, no I donna thinka so
unless there is moreto your config than you gave me

ever heard of TurboGears? it is like rails
i got turbogears apps on /financial and /wiki
each uses a different port

No, but I am intimately familiar with ProxyPass

not that i created /var/www/family/wiki for some reason apache successfully hands off /wiki to localhost:8200
now*
is there some way to debug steps apache takes?

ok, but it is not happening for the reasons that you think it is. at least not if those rules are all that is really going on
in this case the access_log would be enough
but you can turn Loglevel Debug on in that vhost and waht mod_proxy handle or NOT handle the request
I submit that it is not handling /wiki or /financial but you may have a default page in those locations that is redirecting/rewriting to /wiki/ /financial/ and making it work

ok..i sent you a private message with URL to real virtual hosts using Now

currently using zoneedit for registrars DNS to vhosts on my server, though I am running webmin/bind/virtualmin and everything works fine... BUT cant I just bypass zoneedit and point to my box instead?

ok, giving it a look.

my /var/ww/family/* directories are empty :/

testing a theory, one minute

ok..thanks

could you paste up the tail of your access log please for the apache on port 80?
while you are doing that, I need to afk a sec, brb

nothing shows up in /var/log/apache2/access.log

does anyone have any experience with grinder
or know of a better one to use?

not possible unless you have logging off

does anyone have any experience with benchmarking server tools?

siege, jmeter, ab
siege

siege is HTTP regression testing and benchmarking utility that can be found at http://www.joedog.org/JoeDog/Siege

jmeter

jmeter is http://jakarta.apache.org/jmeter/ or http://javaboutique.internet.com/tutorials/JMeter/

apachebench

apachebench is http://httpd.apache.org/docs/2.2/programs/ab.html or better known as 'ab' or comes with apache or not very good to test with according to chipig.
apachebench is a href="http://httpd.apache.org/docs/2.2/programs/ab.html"http://httpd.apache.org/docs/2.2/programs/ab.html/a or better known as 'ab' or comes with apache or not very good to test with according to chipig.

thanks

ab's not enough characters for a factoid...

where do you toggle logging on/off?

logging

logging is http://httpd.apache.org/docs/logs.html or http://httpd.apache.org/docs-2.0/logs.html

but whats the best?

I like jmeter for testing applications, but building the scripts can be a bit tenuous

httperf

httperf is http://www.hpl.hp.com/research/linux/httperf/

that one's nice. gives stddev numbers too so you can tell spurious results

hmm wrong word, bit difficult
cool, new toy to try

I tried running grinder but its confusing and the documentation sucks

not free but cheap enough, and goes in to good detail. somewhat rails specific but not so much that it's not generally applicable

bah

humbug

noodl and his rails stuff
:P

not free not for me

for some reason I am only able to send two concurrent requests to my apache server from any given client
is there a way to increase that?
any more than two wait for one of the active two to complete before making progress
i've seen total request limits, but not per-client

check your client too

i've created a .htaccess but it doesn't seem to be doing a thing (it's just listing the directory contents of the folder .htaccess is in) - the .htaccess looks like this: http://dpaste.com/12108/ - thansk

my client is firefox
at least at this point

firfox has a limit

hmm
isn't it like 8 though...

check about:config for the exact ones

ooo
network.http.max-persistent-connections-per-server
persistent-connections...
just plain connections per is 8
that one looks like two
i wonder what the threshold is for "persistant"
thanks for the tip

no problem
btw, I am sure there are apache modules to limit this

Im trying to run jmeter on mac os x
what do i need to dl to get it to work?
jpk?
I dont know what this means
-bash: PK: command not found

block are totally ignored?

yes

thanks

Perhaps a very simple question, but how do forward the visitors of a certain site another site?

you shouldn't need anything extra. jmeter works just fone for me on the mac.. just double click the .jar file

*to another site

fajita redirect?

redirect is http://httpd.apache.org/docs-2.0/mod/mod_alias.html#redirect or http://httpd.apache.org/docs/mod/mod_alias.html#redirect or see 'Redirect Context'

noodl, ok, i'll check thanks

Directory /var/www/darcs
Options +ExecCGI +FollowSymLinks
Order Allow,Deny
Deny from All
Allow from 127.0.0.1

Order Allow,Deny is The Allow directives are evaluated before the Deny directives. Access is denied by default. Any client which does not match an Allow directive or does match a Deny directive will be denied access to the server
Deny from All is Denies access to the resource. See also Order, Allow, and Satisfy.

/Directory
It appears symlinks are _not_ being followed by a cgi script that uses this directory.

don't paste

What can be the possible reasons for this?
5 lines.
6

what error are you getting? 404?

It's darcs.cgi and I get that the repository contains zero patches.
It does show the name of the symbolic link to the repo.
I get no errors.

uh
repository?
patches?
im lost
i thought you're trying to execute a CGI

the cgi file gets executed already
it just seems that it doesn't cross the symbolic link

the CGI is a symlink?
if it gets executed then it must be following the symlink right?

no, the CGI is not a symlink

the script itself is trying to access a symlink but can't?

the file in the directory /var/www/darcs is
that's what I think is happening.

that has nothing to do with apache config...

that should have nothing to do with the webserver

it could be a permissions problem
what happens if you run your script on the command line?

I am not sure how to execute that script. Just doing ./darcs.cgi does execute the script, but doesn't give HTML, like when run via Apache.
This is the script: http://pastebin.ca/562706

is there any special way in which Apache calls the cgi script?
I just do 127.0.0.1/cgi-bin/darcs.cgi in my browser.

well there's a lot of environment variables that will be missing/different when run from the command line
so if the script is designed to know whether its being executed from a browser or not it might behave different

well, that's happening.

but if it gives no output at all then it sounds like something's wrong with it

it does give output

oh
did you paste the output to the pastebin?
make a new pastebin with the command line output

bad script_name "" at ./darcs.cgi line 311.
that's what I get from the CLI.

ehhhh
i guess that's what it's designed to do
heh
it wants to be run from a browser
what's the output when run from a browser?

anyone know how to link to an Intranet path from the web? is it file://////Server/Folder/ ?

when run from a browser I get a listing containing one element of the project I am interested in.
when I click that, the project is empty.
i.e. contains zero patches.
Darcs is a SCM

i don't know how to fix that since it's not an apache problem

yes, I understand.
I think it's just some piece of crappy software.
(the script)
if it takes me longer than 5 minutes to get working, obviously it's not polished.

could you give an example? I don't understand what you want to do

he wants to link to a file on his computer
local file
i think
at any rate, it's not an apache question :P

i'm building an intranet document site and I want to include some links to shares and documents on the network

the link should be whatever appears to work in your browser
if file://blah/blah works then that's the link

perfecto
the answer is whatever the answer is
is your last name Bush?

uhh
wha?

nm that was retarded

do you know how to bring up the file in a browser?

try the link in a browser and see if it works

so in other words, when I find out what the right address is... thats the address

open file...

what OS/browser?

will probably give you a directory browser and then show you the url and the file

needs to work in windows

right click on the file, open with, internet explorer
then look in the browser and see what URL appears

but for folders

i guess that part would be easy enough to work out with the quick file test...

thats where it gets tricky

there's no difference

ok...

you just leave out the /filename part at the end and you have a folder

I want to open ctemp
file:///c:/temp/ is the URL I use

yup

right

with a sharename iirc

so that's what your link should be like

but what if its an intranet server path?
such as \\server\path

then it's file://server/path/
have you tried it?

rotate the \ clockwise 60 degrees and put file: on the front

maybe if you try it you'll find that it works :P

and if you want to open \\SERVER\folder\ you go to file://///SERVER/folder/

where did all those /'s come from?

file:/// + //SERVER/folder/

i don't know
ask bill gates

yaha wtf it didn't work before
maybe I has an extra /

you did in your initial example

just keep adding or removing /'s until it works
the end

the end is only the beginning

lol, best advice

/// is for any folder and since \\ is for server you add // ?
perfect

peeeeeerrrrrrfect. With Cheese on Top

jeeze

don't try to make sense of it

that doesn't seem right to me. but then windows often doesn't

yep

i'd go with jesse's strategy. start with a couple of slashes and then add more until it works

you'll probably find \\SERVER\folder\ works fine in IE though

i'd be surprised if you got to 5

although it will break FF

smells like a contest

file:/// is "this server" btw

right

the leading // is absolute, and the server is defaulted

how does one forward all requests to http://mysite.com to http://www.mysite.com
I've tried different combinations of server alias and name but with no effect

i would suggest a rewrite rule that looks at the HTTP_HOST

could you point me to an example of that ?

RewriteCond %{HTTP_HOST} ^mysite.com$
RewriteRule ^.*$ http://www.mysite.com/$1

I see - thanks mercutio

but i'm not an expert
the rule follows the condition
if ...RewriteCond... then ...RewriteRule... (if you see what i mean)

I do

hello, what's a 'specific error 1' when i try to netstart apache?

where do you see it

in the cmd windows at starts
i'm on windows, i'm using wampserver, and since 1hour i can't start apache
on localhost
3 2007] [notice] Parent: Child process exited
and no log error registered since there

oh
uhh

i really dunno if i done something wrong in the httpd.conf or php.ini
or if it's a bad save file encoding via notepad, for instance

windows doesn't give ya much help

eh i've no time and no money to switch to linux

linux is free
well
it can be free

i've a win modem

i c

and i've ubuntu here already
i just wanna work on localhost as all every normal people
1 2007] [notice] Parent: Received shutdown signal -- Shutting down the
3 2007] [notice] Child 2560: All worker threads have

ok sto
stop

HAMMERTIME!

that's all i've in the log
before 21.40 it worked
maybe it's better if i google a bit instead bother yourself
thanks the same

true

does apache 2 still use a pooled-process architecture?

i think so

mpm

mpm is http://httpd.apache.org/docs/2.2/mpm.html or Multi-Processing Modules or sexyhandyhandy

sexyhandyhandy?

uh yeah
i guess

i donut geddit.

well years ago they used to be sexy
c.f. not having them
maybe they grew hands too?

guys, one question. What does apache don't who letter with acents? like: á ?

sorry Toerkeium i can't parse your question

I mean, I just uploaded a web site which contains á characters, and ther arent shown like that

AddDefaultCharset

AddDefaultCharset is http://httpd.apache.org/docs/2.2/mod/core.html#adddefaultcharset

thank you!

because I have á and ñ, if I add UTF8 it shows 'á', but not 'ñ', and vice versa
I guess I have to add two then

no, that can't work

only one?

only one is perl and its enabled

http can only specify one charset= per response, and besides, there's no way to indicate which parts of the response are in which charset
no only one is reply

OK, noodl.

humm.. and how can I troubleshoot this?
any idea what can be done?

if your html file uses two distinct charsets, it's broken. not much apache can do about that.

actually, what i was wondering

I don't think it uses 2 distinct charsets..

is whether apache server runs as a single process

I am probably missing the right one

or pre-forks
like 1.3 did

see the mpm docs. generally is keeps a process pool, which can be threaded if you run worker

what's the most common charset used?

k, i didn't read far enough into it, ty

I guess that should be the right one

very much

the default is 8859-1 but that's pretty useless. go with utf8 and fix your content to use it ")

but I think UTF8 doesn't support the ñ character

that's doubtful

when I force apache to use UTF8 the ñ character is not shown

then your document isn't encoded with utf-8

probably

probably is the same as 'definitely not' when dealing with important or critical applications - if you're not sure, don't let 'probably' suffice

find a text editor that doesn't lie to you

probably

probably is the same as 'definitely not' when dealing with important or critical applications - if you're not sure, don't let 'probably' suffice

lol
it's true, definitely not using UTF8 :P)
I mean, I have just downloaded this pages from a web server from the usa, where á and ñ characters are not default for that language, so they have to have a charset which can be used for this, some spanish charset I guess
ISO-8859-8 ?

iso-8859-15 perhaps. anyway it's pointless guessing.. one strategy is to funge your browser's encoding menu 'till it looks right and then stick with that
if you have documents with mutliple character sets, use mod_mime to send the correct header based on a file name extension

Do I need to do apache2ctl restart to reread the server configuration always? Isn't there some binding to the Linux kernel to some fils system notification library?
s/fils/file

that'd be silly. saving a broken config would kill your server
but restart is wrong too, use graceful

"Kill"?
it would simply not work

that doesn't kill existing connections

In a production setup I can imagine that.
ok, is there some mode s.t. it is in "develop" setup?

nope. remember that apache is maintained by C programmers who have to recompile to check ever change they make. at a guess, issuing a USR2 for them isn't a big deal :p
*every

definition but I get the same error.

heh, yes, what horror would it be to develop it.
I modified an Emacs IRC client while it was running.

Options context

server config, virtual host, directory, .htaccess (.htaccess requires AllowOverride Options in the httpd config)

just like you can do in SmallTalk or other 30 year old tech.

WAAA! lisp zealot!

block

hmm, not really.
C, the language, is good in some areas.
It's just that the notion of "interactive development" doesn't exist in the C world.

it has nothing to do with the language in which apache is written

here's the cookie.. is emacs threaded in such a way that saving config files/scripts could change the behaviour of running requests? that's why i think httpd avoids checking httpd.conf for changes

it's for efficiency

that's ridiculous
getting the kernel to notify you takes constant time.
O(1)

did that.

restart apache.

not really.. there's a bunch of stuff that can be done in advance with apache modules that you wouldn't want to do per-request

if you have a kernel which can notify on file changes and you know all the relevant files,

AllowOverride All /Directory

then script a graceful restart

did that

if all else fails, pastebin the relevant section of your config.

not all os's have that facility anyway, and apache is xplat.

sample url

Could you please give us a URL at which the behavior is demonstrated. This would help us troubleshoot the problem.

however, the efficiency i was referring to was effectively compiling parts of the config file into runtime structures

ok, google just won't give me the right answer

rereading the config file is relatively expensive

how can I do a simple redirect to make http://webmail.domain.com go to http://mail.domain.com/webmail ?

that can also work differently

redirect

redirect is http://httpd.apache.org/docs-2.0/mod/mod_alias.html#redirect or http://httpd.apache.org/docs/mod/mod_alias.html#redirect or see 'Redirect Context'

sure, there are alternatives

change your nick

thanks fasta

fasta--

+ thumbs

but the one apache chose is optimized for the case where the webserver runs for a long time without config changes, and uses the minimum resources on each request

oh, never mind
your tab completion stinks
not my fauls

http://apache.pastebin.ca/562995

er fault
I am sure that Apache also isn't optimal for that purpose.

is usually sufficient.

it may not be but it gives it a good try

I can agree with that.

I never had probems making configs in 1.3... ever since I switched to 2 everything I touch busts.

i think fasta's got a point. parsing the config, which might include reading big mime lists etc, is expensive and not something i'd want to see happen simply on file saves, but there's a case to be made for being able to change the config tree in-mempory at runtime

http://apache.pastebin.ca/562995

sure, but it's harder than it looks

which brings us back to C

no doubt

in the original prefork'ing architecture, every child has its own copy of the config

pastebin the FULL error log entries as well please

chill.. it's not fasta's fault that you mistyped the nick and you're being a little aggressive on the point
fasta++

in the threaded model, there's just one but it's immutable and therefore doesn't need to be protected with a mutex
in either case, doing a graceful restart does actually just reread the config file and use it for new connections
so that seems like it's good enough, no?
of course, there are some limitations.

http://apache.pastebin.ca/563018

so it's not ideal, certainly

thumbs, the apache docs only provide options for redirecting directories, not redirecting FROM the subdomain

no arguments here.. just there's a bunch of people who'd like to be able to configure apache via their browsers without needing plesk etc

you can redirect in the vhost.
redirect context

server config, virtual host, directory, .htaccess

could I bother you for an example?

that seems OK. There is an external element that seems to complicate matters

I found "something"

redirect 301 / http://some.server.com

it messes up the character set when the page get values from a database, which field is defined as latin1_spanish_ci
so I guess it's not an apache problem
probably a mysql or php

blame php

PHP should be presumed to be at fault until conclusively proven otherwise. And even thereafter, if it's convenient

yeah, i sympathize with that

heh

apache doesn't care much about character sets
it just sends the encoding to the browser, which needs to know.

ahhh. nice..

so you have to tell apache what character encoding is used, so it can tell the browser

thumbs++

how does sound like this to you? :

i'll look in the main config

no no.. I was going to say something stupid

go ahead. maybe you'll be immortalized.

What's wrong with this cgi configuration? http://pastebin.ca/563036

(knowing, I could already said something stupid without knowing)

you never know

heheh jpeg

I placed a file helloworld.cgi, made it executable.

gonna do a bit of more research
it's killing me

Options ExecCGI is redundant, the ScriptAlias directive takes care of that.

Never mind

what does the error log say?

what is your issue?

I looked at it and solved it already

ok.

the database is presumably returning latin1 (with spanish collation) -- i don't know what ci means

Sorry for bothering. I am a new Apache user, well, I used it years ago, but that was an EasyPHP distribution. I am now setting up a real solution for my personal needs.

disabling UTF8 from apache makes everthing works, but only data retrieved from a database is now showing characters properly
now = not

"disabling utf8" means removing the adddefaultcharset line, i presume?
i think your problem is that you're inserting latin1 characters into a utf8 template, probably

yes

you'll need to make sure that the whole web page is in just one character set
php has an interface to iconv iirc
it's nothing really to do with apache. apache just ships the bits

the problem is that this is a customer, which was hosting the webpage in some other place, and without this issue
yes, I think apache is not the problem

perhaps the pages are in a consistent character set but you're not setting the correct one
you could do that in php, too.

Ok, I got CGI working in a different virtual host. How nice

what do you mean by "consistent" ?

all the same

ohh

Consistent means not(A and not A)

ok, now I am sure it's because of the data field setting to latin_bla_bla

where A is any formula

in this case, the formula is (exists ch s.t. ch is encoded in X) for some char encoding X

buaaa

noodl-- not(chilli sauce & not chilli sauce (noodls, in fact))

yes, you have to tell databases how text is encoded, and how to collate it.

before I lear abour charaset encoding, I need to learn english

we could do this in spanish if you prefer

heheh nice!
no it's ok, I understand what you said
so I should try to find how to show that information retrieved from the database which is set to a charset 'latin_blabla' properly
try to find how to show = convert, right?

well, there are two options
either you make sure the text it is inserted into is also latin1
and then you send the whole page as latin1
or you convert the database text to utf8 (assuming that's what the template uses)
and then you send the whole page as utf8
where "send the page as" means setting the character encoding in the headers

so for example, if I let clear the charset configuration in database, then I it should be corrected right?

no, the database isn't converting the data.
the configuration is a declaration. probably
you can't change it without changing all the data.
at least in the databases i'm familiar with.

yes, realize of that just one second ago

the problem is that text doesn't have an implicit encoding.

can't change it

it's just a bunch of bits

so I have to screw-up php

was the whole installation working on "some other system" (possibly apache 1.3)?
or did you change the php
or perhaps the database?

yes, it was another apache verion, I tried copying the httpd.conf file, and said "found apache 1.3 directives" or something like that
you know, I tried all the lazzy things to do before using my brain
heh
there are surely all different version, php, apache and mysql
it was in a old system

ah
well, there is something you should know, anyway.
there are two ways of telling a browser what the character encoding is:
1) in the headers sent before the html

tag

the second one only works if the first one is not used

yes, I readed that in httpd.conf