Re: cvs commit: httpd-test/flood config.h.in flood_round_robin.c
On Sun, 7 Sep 2003 [EMAIL PROTECTED] wrote:
+if (apr_procattr_create(procattr, rp-pool) != APR_SUCCESS) {
+apr_file_printf(local_stderr,
+apr_procattr_create failed for '%s'\n,
+rp-url[rp-current_url].responsescript);
+return APR_EGENERAL;
+}
Why do all of these return APR_EGENERAL rather than catching the
apr_status_t from the function that was called (apr_procattr_create in
this case) and returning that? Is this just a flood thing I don't know
about?
--Cliff
flood 1.1 TR
I think we're *really* ready to release flood 1.1. I would like to tag the tree in a day or two. We don't have that many commits daily, so I think we can simply test HEAD, without any temporary candidate tag (like those on httpd repo). I would also like to be RM for this release, unless there's somebody more experienced to take this job. Anyway I'll need some help. First, my public key schould be added to KEYS file (see attached patch), so I can sign the tarballs. Second, in order to obey new mirroring rules, I need to have +w to www.apache.org/dist/httpd/flood, or somebody has to put release tarballs for me. In addition, I would like to put flood manual on-line. This requires also some privileges (only for manual subdir), or a person that would put them on-line for me. regards, Jacek Prucia P.S. Once 1.1 is released, we might consider moving 0.4 to archive, since we don't want anybody to use when 1.1 is out. --- KEYS.orig 2003-09-05 14:27:37.0 +0200 +++ KEYS2003-09-05 14:30:02.0 +0200 @@ -2539,3 +2539,38 @@ 7JDHKkJmLFKQAJ4/oORquEdlf3G5KxtLSYuyGEqemg== =4lbY -END PGP PUBLIC KEY BLOCK- +pub 1024D/7EFFAEB0 2003-09-05 Jacek Prucia [EMAIL PROTECTED] + Key fingerprint = 07FE 0F5B 4DA5 D43A 7765 BCFA F694 FBF1 7EFF AEB0 +sig 3 7EFFAEB0 2003-09-05 Jacek Prucia [EMAIL PROTECTED] +uidJacek Prucia [EMAIL PROTECTED] +sig 3 7EFFAEB0 2003-09-05 Jacek Prucia [EMAIL PROTECTED] +sub 1024g/7DFCDBC1 2003-09-05 +sig 7EFFAEB0 2003-09-05 Jacek Prucia [EMAIL PROTECTED] + +-BEGIN PGP PUBLIC KEY BLOCK- +Version: GnuPG v1.2.2 (GNU/Linux) + +mQGiBD9YaeARBACCzmLnOVl9OJma0QE6R4FUqKb/aPZbZm4kmsLbGC13kJm+QGp8 +47eL9bLVTfF4wBLthtluCUzz3/f+OLzhwzPX4sEobrqxonKCkCjbNYVjzne09oZe +Ww/uWJffc+sGMG3Cw012KnzwzcjsPMzppWKQEq8lS45N7qKePJJaXjJ8cwCg3zsR +cYtVrGVqouQS/IhBEnazdTED/22Pve0Bc43DzF4kdEcHdVE1GRLcjPCtcJLqjhLu +xs4NuoNAUU1ITIj574r0L3CnVreDqjoEJxiq2qns363wVuUAWzRTpfQ/DIojOC0J +nkYPtHhzCLVPFM4ElTXFDQ1P+QRf5NscKrhYooaa6WRMkXdoD+4+nmfa7ky9YEP5 +oAY/A/9bOPaxY01sEKslP+IZD5kp0NlNH/49sT9/1GXWmEXpkZvBPsjMi/eTOsjr +dQ7ehaoh44t1dSyLBt8NrAhN1Boy5da3AZdxEICxlY/buDBASBuq4cgeTHOI4nXK +xPc5PeJnEQ/jV2jZccxx5gQ6u2UdXe3lGbADDsiX4kPk2H8JSrQgSmFjZWsgUHJ1 +Y2lhIDxqYWNla3BAYXBhY2hlLm9yZz6IWwQTEQIAGwUCP1hp4AYLCQgHAwIDFQID +AxYCAQIeAQIXgAAKCRD2lPvxfv+usFGMAJ47s0rDI83SPOOMLxhiyXB514kHQgCf +UnNRyAk6b7of3rv1VkkLZY4Et9q0JkphY2VrIFBydWNpYSA8amFjZWsucHJ1Y2lh +QGFjbi53YXcucGw+iF4EExECAB4FAj9Ya4UCGwMGCwkIBwMCAxUCAwMWAgECHgEC +F4AACgkQ9pT78X7/rrA+CQCeMERG3+9VgMGQXqjdL3pduz3MpMgAn0ICgt7f2VvG +4pEOYu+DcOWbmc3CuQENBD9YaeQQBACXmGQrEbtPO0A8bA02YF4u9o57EcG/IjBC +WxUIdIYFi/4W5xUVjK0cKrzu9WLznmizf0kisCCm6mg6hZgu1j+D89cC2APoAf4D +hqEsh+QLVa7k53Zz9D8oT0XMyXIfszLJgUaURZsyWs55UdGEaknI5qwnZiZyQh/+ +ulL5ZVraswADBQP+IHk7S+agvzP3jCnko4RbbDBFPINkEnLvxoooBJgQgjTKP3Tr +/D7+dYZy7DDsM+aJwqMDcMQD+XWkdiezps7mtDo5+Xn+4+Qmx7ttcJqyqFb8/Qq6 ++26Vug8Bau1l9Wgyb4I+UE67w9Mwdh8cGGBpQ0dwpsoLBVbHowttCRa56j2IRgQY +EQIABgUCP1hp5AAKCRD2lPvxfv+usGtfAJ9ua/4DCUBvaezsV6SB5XJjUSEHrgCg +ztzUT6KsFJmyBHwo1PeP9qBgjGg= +=XInH +-END PGP PUBLIC KEY BLOCK-
RE: cvs commit: httpd-test/flood config.h.in flood_round_robin.c
From: Jacek Prucia [mailto:[EMAIL PROTECTED]
Sent: Monday, September 08, 2003 2:41 AM
On Sun, 7 Sep 2003 20:06:09 -0400 (EDT)
Cliff Woolley [EMAIL PROTECTED] wrote:
On Sun, 7 Sep 2003 [EMAIL PROTECTED] wrote:
+if (apr_procattr_create(procattr, rp-pool) != APR_SUCCESS) {
+apr_file_printf(local_stderr,
+apr_procattr_create failed for '%s'\n,
+rp-url[rp-current_url].responsescript);
+return APR_EGENERAL;
+}
Why do all of these return APR_EGENERAL rather than catching the
apr_status_t from the function that was called (apr_procattr_create in
this case) and returning that? Is this just a flood thing I don't know
about?
Nope. I'm not so familliar with APR, so a message to stderr plus APR_EGENERAL
is a safe bet. If this looks really obscure considering APR concepts, then
please feel free to commit a fix.
Basically do something like this:
apr_status_t rv;
...
rv = apr_procattr_create(procattr, rp-pool);
if (rv) { /* Or in full: if (rv != APR_SUCCESS) */
apr_file_printf(local_stderr,
apr_procattr_create failed for '%s'\n,
rp-url[rp-current_url].responsescript);
return rv;
}
IOW, preserve the error and pass it down back the calling chain.
Sander
RE: cvs commit: httpd-test/flood config.h.in flood_round_robin.c
is a safe bet. If this looks really obscure considering APR concepts, then please feel free to commit a fix. Fix commited. Please test, since I didn't. :) PS: I also fixed a buglet or two. --Cliff
Re: cvs commit: httpd-test/flood config.h.in flood_round_robin.c
On Sun, 7 Sep 2003 22:22:44 -0400 (EDT) Cliff Woolley [EMAIL PROTECTED] wrote: is a safe bet. If this looks really obscure considering APR concepts, then please feel free to commit a fix. Fix commited. Thanks for the fix! Please test, since I didn't. :) Fix seems to be OK and error messages are more descriptive than 'Error string not specified yet'. :) regards, Jacek Prucia
Re: flood 1.1 TR
On Mon, 08 Sep 2003 00:15:27 -0700 Justin Erenkrantz [EMAIL PROTECTED] wrote: [...] My only concern is that we make sure flood 1.1 compiles against APR 1.0 (which should have a 1.0 RC out this week). Try checking out the HEAD of apr and apr-util and ensuring that flood still compiles. All of the deprecated functions in APR are now removed in HEAD of those repositories. If you look at the CHANGES in apr, you'll see the mappings of old to new functions. You were right. Compilation barfed on socket API changes. However it turned out to be a trival to fix (adding one extra argument and renaming few functions). I have patch ready and I'll be commiting it later this day. [...] Yup, you should be RM. ;-) :) I would first say that you should create 'candidate' tarballs on your own. Then, we'll have to get 3 +1s before calling it 1.1, so it's fine if there is a little latency between you creating the tarball and posting it on the site. We'll have to wait for the 3 +1s anyway. Yeah, dev/release.html rules apply. Once you have the 3 +1s, you can then tag it 'officially' in CVS (which you can already do). Yes. I've just decided to use intermediate tag (FLOOD_1_1_RC), see below. At that point, I can ensure that you have the right permissions and stuff to upload and distribute it. But, I'd recommend getting the tarball right and having a 'solid' release, then we can figure out the other stuff first. Make sense? -- justin Yes. This is my current release plan (please yell if it violates something): 1. tag repo with FLOOD_1_1_RC (probably later this day), 2. prepare RC tarball and announce availability here, 3. test and see if eventual bugs must be fixed, 4. move FLOOD_1_1_RC tag on files containing fixes, 5. create new RC tarball and start all over again (3). When we have 3 +1s for GA, then: 1. create FLOOD_1_1 tag using existing FLOOD_1_1_RC, 2. remove old FLOOD_1_1_RC tag, 3. sign the tarball and move it to dist/httpd/flood, 4. wait 48 hours ;) 5. send announcement. So like you said: we can postpone most of the stuff until we get stable RC tarball. regards, Jacek Prucia
RE: flood 1.1 TR
From: Jacek Prucia [mailto:[EMAIL PROTECTED] Sent: Monday, September 08, 2003 3:28 PM Yes. This is my current release plan (please yell if it violates something): 1. tag repo with FLOOD_1_1_RC (probably later this day), 2. prepare RC tarball and announce availability here, 3. test and see if eventual bugs must be fixed, 4. move FLOOD_1_1_RC tag on files containing fixes, 5. create new RC tarball and start all over again (3). Please do not move tags. Instead create RC1, RC2, etc. After the release is done, remove all RCx tags. When we have 3 +1s for GA, then: 1. create FLOOD_1_1 tag using existing FLOOD_1_1_RC, 2. remove old FLOOD_1_1_RC tag, 3. sign the tarball and move it to dist/httpd/flood, 4. wait 48 hours ;) 5. send announcement. So like you said: we can postpone most of the stuff until we get stable RC tarball. Sander
example configs
I have found few minor issues in current example config files. 1. round-robin-ssl.xml We have only one url working -- modssl.org SSL connection test is working again (although HTML is broken). The other URL is dead for a lng time, so we might consider removing it at all. I've found another SSL check: https://www.freessl.com/test/freessl.html If nobody has any problems with commercial site in example URLs, I can contact them and see if such inclusion is possible. 2. round-robin-keepalive.xml This just looks like a great place to use baseurl. 3. round-robin-dynamic.xml apachelabs.org is a redirect to apache.org, so first three urls are sure to fail. Google stuff is commented out, so we need to exchange apachelabs.org with something else. Something like retrieving some static page, and using a first link found. Those fixes are ok for purpose of this release, but this whole issue looks like a good showstopper for 1.2. We schould provide a bunch of CGI/PHP/JSP scripts to simulate some functionality and drop remote servers for good. All example configs could be rewritten to have baseurlhttp://www.example.com/baseurl, so they are harmless by default.
Patch to Apache::TestConfig for Mandrake
I've created the following patch to help TestConfig.pm find the proper
mod_perl httpd on Mandrake machines. Mandrake installs both a
/usr/sbin/httpd and /usr/sbin/httpd-perl, and includes configuration for
mod_perl acceleration with a standard (e.g. non-mod_perl) httpd in front of
it.
However, when Apache::Test is used, it finds /usr/sbin/httpd instead of the
-perl one. I've fixed this for my site, but so I don't run into this later
(nor for others), I'm sending this patch to the list.
If you want to reply, please CC me, since I'm not on the list.
RCS file:
/home/cvspublic/httpd-test/perl-framework/Apache-Test/lib/Apache/TestConfig.pm,v
retrieving revision 1.171
diff -u -3 -p -r1.171 TestConfig.pm
--- TestConfig.pm 13 Aug 2003 19:02:51 - 1.171
+++ TestConfig.pm 8 Sep 2003 21:14:36 -
@@ -310,7 +310,13 @@ sub configure_httpd {
for my $dir (map { $vars-{$_} } qw(sbindir bindir)) {
next unless defined $dir;
my $httpd = catfile $dir, $vars-{target};
-next unless -x $httpd;
+unless (-x $httpd) {
+if (-x $httpd . '-perl') {
+$vars-{httpd} = $httpd . '-perl';
+} else {
+next;
+}
+}
$vars-{httpd} = $httpd;
last;
}
--
/* Michael A. Nachbaur [EMAIL PROTECTED]
* http://nachbaur.com/pgpkey.asc
*/
Rome wasn't burned in a day.
Re: Patch to Apache::TestConfig for Mandrake
Just after I sent the message, I saw the bug. Just change:
$vars-{httpd} = $httpd . -perl
to say instead:
$httpd = $httpd . -perl
On Monday 08 September 2003 02:53 pm, Michael A Nachbaur wrote:
I've created the following patch to help TestConfig.pm find the proper
mod_perl httpd on Mandrake machines. Mandrake installs both a
/usr/sbin/httpd and /usr/sbin/httpd-perl, and includes configuration for
mod_perl acceleration with a standard (e.g. non-mod_perl) httpd in front of
it.
However, when Apache::Test is used, it finds /usr/sbin/httpd instead of the
-perl one. I've fixed this for my site, but so I don't run into this later
(nor for others), I'm sending this patch to the list.
If you want to reply, please CC me, since I'm not on the list.
RCS file:
/home/cvspublic/httpd-test/perl-framework/Apache-Test/lib/Apache/TestConfig
.pm,v retrieving revision 1.171
diff -u -3 -p -r1.171 TestConfig.pm
--- TestConfig.pm 13 Aug 2003 19:02:51 - 1.171
+++ TestConfig.pm 8 Sep 2003 21:14:36 -
@@ -310,7 +310,13 @@ sub configure_httpd {
for my $dir (map { $vars-{$_} } qw(sbindir bindir)) {
next unless defined $dir;
my $httpd = catfile $dir, $vars-{target};
-next unless -x $httpd;
+unless (-x $httpd) {
+if (-x $httpd . '-perl') {
+$vars-{httpd} = $httpd . '-perl';
+} else {
+next;
+}
+}
$vars-{httpd} = $httpd;
last;
}
--
/* Michael A. Nachbaur [EMAIL PROTECTED]
* http://nachbaur.com/pgpkey.asc
*/
Ford, he said, how many escape capsules are there?
None, said Ford.
Zaphod gibbered.
Did you count them? he yelled.
Twice, said Ford
Re: ScriptLog
On Sun, Sep 07, 2003 at 05:38:56PM -0400, Cliff Woolley wrote: On Sun, 7 Sep 2003, Manoj Kasichainula wrote: If it's only for debugging, can't CGI writers just add a line to their code to rebind stderr to a file? Only if the error is output from the script as opposed to a compilation failure or other interpreter weirdness. Good point. The other alternative I've pondered is for people debugging their own CGIs to use a wrapper that rebinds stderr and execs the original CGI, but I suspect I'll get shouted down at this point :)
getting MPM model of apache web server
Hi All, I have an application in which I need to check whether the apache server is prefork based or worker MPM based. How can I get the model info from web server in my application. Normal commad line way to check this is by simply giving the command apachectl -V. Is there any apache specific api which gives info regading the MPM model of apache web server server. Regards, Gagan
Re: apr_global_mutex_child_init mod_watch
Anthony Howe wrote: flock is used on FreeBSD by default... flock requires that chmod be done on the lock in the child init As an aside, why is flock() used on FreeBSD? It has SysV like mutexes and POSIX ones too (at least in 4.8) - I may sound unsure cause I've only recently switched platforms from Linux to FreeBSD just to be different. Support for Posix sems was a somewhat new addition. SysV sems were also not guarateed to be present (this is all old stuff by the way, and inherited from 1.3)... at Apache 2.0.48-dev, you have to do the chmod() on the lock file, like mod_ssl does in that branch OK. Is there an API that returns what the APR_LOCK_DEFAULT mutex will be? Check out apr_proc_mutex_name() and apr_proc_mutex_defname() -- === Jim Jagielski [|] [EMAIL PROTECTED] [|] http://www.jaguNET.com/ A society that will trade a little liberty for a little order will lose both and deserve neither - T.Jefferson
Re: getting MPM model of apache web server
Gagan Puri wrote: Hi All, I have an application in which I need to check whether the apache server is prefork based or worker MPM based. How can I get the model info from web server in my application. Normal commad line way to check this is by simply giving the command apachectl -V. Is there any apache specific api which gives info regading the MPM model of apache web server server. see ap_mpm_query. for an example of its use, try http://marc.theaimsgroup.com/?l=apache-httpd-devm=106086674806449w=2 you can also read through this thread http://marc.theaimsgroup.com/?t=10568856021r=1w=2 for ideas on why it's probably better to test for the existence of threads as a server property, versus a specific mpm model like worker. HTH --Geoff
Re: apr_global_mutex_child_init mod_watch
flock is used on FreeBSD by default... flock requires that chmod be
done on the lock in the child init
if you look at Apache 2.1-dev, unixd_set_global_mutex_perms() has been
changed to handle the chmod() for you
You mean chown() right? Or do you mean both chmod() and chown()?
at Apache 2.0.48-dev, you have to do the chmod() on the lock file, like
mod_ssl does in that branch
Actually I did chown(), not a chmod() and everything worked on FreeBSD.
The lockfile was create with 0600 permissions already:
#if defined(__unix__)
/* Set the lockfile ownership, if it exists. This will eventually be
* handled by unixd_set_global_mutex_perms() in Apache 2.1 and maybe
* 2.0.48 (but I won't count on it).
*/
if (chown(tp-lockfile, unixd_config.user_id,
unixd_config.group_id) errno != 0 errno != ENOENT) {
ap_log_error(...);
unixd_set_global_mutex_perms((apr_global_mutex_t *) tp-mutex);
#endif
Next Linux testing...
--
Anthony C Howe +33 6 11 89 73 78
http://www.snert.com/ ICQ: 7116561 AIM: Sir Wumpus
- Zen thought
Re: apr_global_mutex_child_init mod_watch
this is a situation where mod_rewrite wasn't a great place to look since mod_rewrite never handled this situation correctly in 2.0, and I fixed mod_rewrite in 2.1-dev by teaching unixd_set_global_mutex_perms() to be smarter... unfortunately, the change has not yet been merged into the stable tree Looking at the SSL code for this aspect, I noticed that mod_ssl will pass NULL with APR_LOCK_DEFAULT, which means that it never sets mc-ChownMutexFile to TRUE and relies on either: apr_global_mutex_create() unixd_set_global_mutex_perms() to chown the lockfile. I've not looked at the code for those two functions, but empirical testing yesterday showed that passing NULL for the lock file and APR_LOCK_DEFAULT to apr_global_mutex_create() will cause an EPERM error later in apr_global_mutex_child_init(). So it looks to me that mod_ssl's solution is NOT 100% reliable in the current version of Apache when APR_LOCK_DEFAULT is used. By the way mod_digest will also have similar problems with mutexes as mod_rewrite. -- Anthony C Howe +33 6 11 89 73 78 http://www.snert.com/ ICQ: 7116561 AIM: Sir Wumpus - Zen thought
Re: apr_global_mutex_child_init mod_watch
flock is used on FreeBSD by default... flock requires that chmod be done on the lock in the child init As an aside, why is flock() used on FreeBSD? It has SysV like mutexes and POSIX ones too (at least in 4.8) - I may sound unsure cause I've only recently switched platforms from Linux to FreeBSD just to be different. at Apache 2.0.48-dev, you have to do the chmod() on the lock file, like mod_ssl does in that branch OK. Is there an API that returns what the APR_LOCK_DEFAULT mutex will be? so you can figure out what kind of mutex it is and, if it is fcntl, do the chmod() yourself, or hope that 2.0 is changed to handle this issue for you That would be the ideal solution, but I'm not sure I can wait until 2.0.48. Plus not everyone upgrades when they should, so I need a work around. BTW are there any version macros in Apache 2 or the APR like: APACHE_VERSION_MAJOR APACHE_VERSION_MINOR APR_VERSION_MAJOR APR_VERSION_MINOR that can be tested at compile time to handle variations? I've copied [EMAIL PROTECTED] to nag about getting the unixd_set_global_mutex_perms() enhancement (doing the right thing with flock mutexes) merged into stable finally. -- Anthony C Howe +33 6 11 89 73 78 http://www.snert.com/ ICQ: 7116561 AIM: Sir Wumpus - Zen thought
Re: cvs commit: httpd-dist KEYS
[EMAIL PROTECTED] wrote: martin 2003/09/08 04:03:40 Modified:.KEYS Log: Add my 768/FDE534D1 key That's a little weak, isn't it? -- http://www.apache-ssl.org/ben.html http://www.thebunker.net/ There is no limit to what a man can do or how far he can go if he doesn't mind who gets the credit. - Robert Woodruff
Re: getting MPM model of apache web server
Gagan Puri wrote: I have an application in which I need to check whether the apache server is prefork based or worker MPM based. How can I get the model info from web server in my application. Normal commad line way to check this is by simply giving the command apachectl -V. Is there any apache specific api which gives info regading the MPM model of apache web server server. see ap_mpm_query() and AP_MPMQ_foo constants in ap_mpm.h you can't find out whether it is prefork or worker specifically, but you can find out whether the server forks (both prefork and worker) and whether the server is threaded (worker but not prefork)...
Re: Tagged the trees
On Sat, 6 Sep 2003 09:49:54 +0200, Sander Striker wrote: I tagged the trees today, as STRIKER_2_0_48_PRE1 and STRIKER_2_1_0_PRE1 respectively. I'll try and get some tarballs up for testing, but for now, please test the tag. Both are currently broken on OS/2 from the recent max mem free stuff. Looks like I have to turn on AP_MPM_WANT_SET_MAX_MEM_FREE for it to compile without error but I don't know exactly what it implies. Is there any reason why I wouldn't want to do the change below? Index: mpm.h === RCS file: /home/cvs/httpd-2.0/server/mpm/mpmt_os2/mpm.h,v retrieving revision 1.4.2.1 diff -u -r1.4.2.1 mpm.h --- mpm.h 3 Feb 2003 17:32:05 - 1.4.2.1 +++ mpm.h 8 Sep 2003 13:08:47 - @@ -71,5 +71,6 @@ #define AP_MPM_WANT_SET_PIDFILE #define AP_MPM_WANT_SET_MAX_REQUESTS #define AP_MPM_DISABLE_NAGLE_ACCEPTED_SOCK +#define AP_MPM_WANT_SET_MAX_MEM_FREE #endif /* APACHE_MPM_MPMT_OS2_H */ -- __ | Brian Havard | He is not the messiah! | | [EMAIL PROTECTED] | He's a very naughty boy! - Life of Brian | --
Re: Tagged the trees
Sander Striker wrote: I tagged the trees today, as STRIKER_2_0_48_PRE1 and STRIKER_2_1_0_PRE1 respectively. I'll try and get some tarballs up for testing, but for now, please test the tag. with recent apr resolver changes, Apache on AIX has picked up the old assert failure that solaris-8-with-no-patches used to have... investigating further...
RE: getting MPM model of apache web server
unsubscribe -Original Message- From: Gagan Puri [mailto:[EMAIL PROTECTED] Sent: Monday, September 08, 2003 4:05 AM To: [EMAIL PROTECTED] Subject: getting MPM model of apache web server Hi All, I have an application in which I need to check whether the apache server is prefork based or worker MPM based. How can I get the model info from web server in my application. Normal commad line way to check this is by simply giving the command apachectl -V. Is there any apache specific api which gives info regading the MPM model of apache web server server. Regards, Gagan
lifetime of r-user in 2.0 tree
There is a comment in protocol.c which states that the lifetime of r-user is per connection. Yet, the code only allocates from the request pool. Tracing this back, I see that in the 1.3 tree, the allocation is in fact from the connection. Can someone explain why the user information's lifespan was changed to per connection to per request? (not to mention updating the misleading comment in the 2.0 tree). Aryeh --- Aryeh Katz Secured-Services Inc.
Re: lifetime of r-user in 2.0 tree
Aryeh Katz wrote: There is a comment in protocol.c which states that the lifetime of r-user is per connection. Yet, the code only allocates from the request pool. Tracing this back, I see that in the 1.3 tree, the allocation is in fact from the connection. Can someone explain why the user information's lifespan was changed to per connection to per request? (not to mention updating the misleading comment in the 2.0 tree). The request is independent from the connection. I'd count the storage of the request user in r-connection in 1.3 as wrongly designed. The second is a technical reason. In 2.x connection and request are *really* separated, so storing the user in the connection structure would be even more problematic. nd
Re: lifetime of r-user in 2.0 tree
From: André Malo [EMAIL PROTECTED] The request is independent from the connection. I'd count the storage of the request user in r-connection in 1.3 as wrongly designed. The second is a technical reason. In 2.x connection and request are *really* separated, so storing the user in the connection structure would be even more problematic. patch to remove erroneous comment supplied --- protocol.c Thu Sep 4 10:03:11 2003 +++ protocol-2.0.47.c Thu Sep 4 10:02:59 2003 @@ -1142,6 +1142,10 @@ } t = ap_pbase64decode(r-pool, auth_line); +/* Note that this allocation has to be made from r-connection-pool + * because it has the lifetime of the connection. The other allocations + * are temporary and can be tossed away any time. + */ r-user = ap_getword_nulls (r-pool, t, ':'); r-ap_auth_type = Basic; nd --- Aryeh Katz Secured-Services Inc.
RE: ScriptLog
If we ever get the cgi-stdout-stdin-stderr read deadlocking issue fixed... I did some initial work to make mod_cgi allow 3rd party modules to register StderrAndOutput filters (techically, ScriptLog is itsself a StderrAndOutput filter and is added as the default). So you could have a mod_javascript_cgi_debugging module that emits javascript code if stderr is present. If they click the javascript alert, it opens a debugging window with your posted form vars, etc. The reason it's called StderrAndOutput filter is, depending on data in the stderr, output may conditionally be good or bad (you could force a CGI bomb to be a status 200 or a status 500, output an error page, etc.). -Original Message- From: Manoj Kasichainula [mailto:[EMAIL PROTECTED] Sent: Monday, September 08, 2003 3:08 AM To: [EMAIL PROTECTED] Subject: Re: ScriptLog On Sun, Sep 07, 2003 at 05:38:56PM -0400, Cliff Woolley wrote: On Sun, 7 Sep 2003, Manoj Kasichainula wrote: If it's only for debugging, can't CGI writers just add a line to their code to rebind stderr to a file? Only if the error is output from the script as opposed to a compilation failure or other interpreter weirdness. Good point. The other alternative I've pondered is for people debugging their own CGIs to use a wrapper that rebinds stderr and execs the original CGI, but I suspect I'll get shouted down at this point :)
ap_sub_req_method_uri and POST
Is it possible to change a GET to a POST, and add post data to the POST request by using get_brigade and pass_brigade and ap_sub_req_method_uri? --- Aryeh Katz Secured-Services Inc.
RE: Tagged the trees
From: Cliff Woolley [mailto:[EMAIL PROTECTED] Sent: Monday, September 08, 2003 8:27 PM On Mon, 8 Sep 2003, Brian Havard wrote: Both are currently broken on OS/2 from the recent max mem free stuff. Looks like I have to turn on AP_MPM_WANT_SET_MAX_MEM_FREE for it to compile without error but I don't know exactly what it implies. Is there any reason why I wouldn't want to do the change below? That is a correct change. My fault... I should have added that for you. Sorry. Depending on the arrival of the changes that are needed to make 2.0.48 and 2.1.0 happen, I'll prolly tag again tomorrow. Sander
cleaning binary distributions
Unless someone has an objection, I'm going to do the following: - Remove all but the most recent (2.0/1.3) binary distributions from each directory under http://www.apache.org/dist/httpd/binaries/ - Remove all versions prior to 1.3.26/2.0.37. (I picked that one because of the chunked encoding vulnerability.) - Delete any directories that wind up empty. Of course, these files will remain permanently on http://archive.apache.org/dist/httpd/binaries/ Joshua.
Re: cvs commit: httpd-2.0/server main.c
* [EMAIL PROTECTED] wrote: martin 2003/09/08 09:39:05 Modified:server main.c Log: Documentation says -DDUMP_VHOSTS is equivalent to setting -S eww. It should say, that -t -D DUMP_VHOSTS is equivalent. So it would be a documentation bug. Where did you read it? nd
Re: Tagged the trees
Sander Striker wrote: Hi, I tagged the trees today, as STRIKER_2_0_48_PRE1 and STRIKER_2_1_0_PRE1 respectively. I'll try and get some tarballs up for testing, but for now, please test the tag. Looks good on RedHat 9 with worker. I didn't do anything special to change thread libraries, so I assume that means it's using the new pthreads. Graceful and non-graceful restarts work fine. Greg
