Re: PRE-RELEASE: lighttpd-1.3.1-20040922-0900.tar.gz

On Wed, Sep 22, 2004 at 10:11:30AM +0100, robin-lists@xxxxxxxxxxxxxx wrote:
> > -----Original Message-----
> > From: Jan Kneschke [mailto:jan@xxxxxxxxxxx] 
> > Sent: 22 September 2004 07:04
> > To: lighttpd@xxxxxxxxxxxxxxxxx
> > Subject: PRE-RELEASE: lighttpd-1.3.1-20040922-0900.tar.gz
> > 
> > 
> > Dear users,
> > 
> > I've packaged all bugfixes into a new release and hope that 
> > it works as expected.
> 
> Jan,
> 
> Are you the maintainer of the cygwin lighttpd package? Are you updating the
> cygwin package at the same time as making new releases? (It may just be that 
> the
> 1.3.1 update hasn't made it to my cygwin mirror.)

Yes, all platforms are maintained from the same source. 1.3.1 is not
released yet. It is just a pre-release which is  used to check if it is
everything is fixed that is supposed to be fixed. After collecting some
feedback the real release is made.
 
> Also, sort of related to cygwin, I'd like to suggest some minor changes to 
> your
> default lighttpd.conf.
> 
> Can you change the log file locations to:
> 
> server.errorlog            = "/var/log/lighttpd/error.log"
> accesslog.filename         = "/var/log/lighttpd/access.log"
> 
> Or something similar? i.e. follow the FHS http://www.pathname.com/fhs/
> 
> Actually, looking at that, the web content should go under /srv, i.e.
> /srv/www/lighttpd or whatever.
> 
> Also, I think that the distribution should as far as possible work out of the
> box, i.e. specify sensible default values, ensure that any directories/log 
> file
> locations are created on install if they don't exist
> 
> Related to my last point, it would be a good idea to distribute some sample
> content in whatever location specified as default server.document-root, sort 
> of
> like apache do. You could even include a link to the local lighttpd
> documentation!

Just prepare something nice and send me a patch against the latest
release. I'm happy the include it.
 
> One other thing that didn't work for me was the spawn-php.sh script. 
> spawn-fcgi
> works fine for me:
> 
> $ spawn-fcgi.exe -f /usr/local/bin/php -p 1026 -C 5
> spawn-fcgi.c.150: child spawned successfully: PID: 4676
> 
> $ ps -ef
>      UID     PID    PPID TTY     STIME COMMAND
>    robin    4804       1 con  09:27:54 /usr/bin/bash
>    robin    4676       1   ?  09:48:07 /usr/local/bin/php
>    robin    4252    4676   ?  09:48:07 /usr/local/bin/php
>    robin    5016    4676   ?  09:48:07 /usr/local/bin/php
>    robin    2560    4676   ?  09:48:07 /usr/local/bin/php
>    robin    5380    4676   ?  09:48:07 /usr/local/bin/php
>    robin    4712    4676   ?  09:48:07 /usr/local/bin/php
>    robin    2340    4804 con  09:49:23 /usr/bin/ps
> 
> But (having killed all the process from the previous command) spawn-php.sh
> returns this:
> 
> $ spawn-php.sh
> UID=1005
> SPAWNFCGI=/usr/local/sbin/spawn-fcgi
> FCGIPORT=1026
> PHP_FCGI_CHILDREN=5
> spawn-fcgi.c.71
> 
> (I've added the 4 debuging lines to check the values of the vars used on the
> spawn-fcgi command line)
> 
> Oh and one last sanity check:
> 
> $ which spawn-fcgi
> /usr/local/sbin/spawn-fcgi
> 
> Looking at the source, line 71 of spawn-fcgi.c appears to be related to
> attempting to open a socket:
> 
>      69    if (-1 == (fcgi_fd = socket(socket_type, SOCK_STREAM, 0))) {
>      70       fprintf(stderr, "%s.%d\n",
>      71          __FILE__, __LINE__);
>      72       return -1;
>      73    }
> 
> I'm not sure why this would work when spawn-fcgi is called directly but not 
> when
> it is called from the shell script wrapper.

Please print out strerror(errno) to see what is the actual complaint.
Actually it might just be a 'adress already in use'. In that case just
wait until the socket is finally release by the OS.
 
> Oh, another minor change - on line 200 of spawn-fcgi.c, the plural of "child" 
> is
> "children".

Ok, will fix that.
  
> Anyway, these are all minor points - I'll be giving lighttpd a go on my WinXP
> box under cygwin over the next few weeks. No doubt I'll have more to say when 
> I
> don't understand something!
> 
> Oh, and by the way, "Thank you Jan" :o)

Please spread the thanks on the full list as there are many
contributiors.
 
> R.
> --
> http://robinbowes.com  
> 
    Jan
  
-- 
Jan Kneschke                                     http://jan.kneschke.de/
Perhaps you want to say 'thank you, jan':    http://jk.123.org/wishlist/