ANNOUNCE: Apache::OpenIndex 1.05
Apache::OpenIndex version 1.05 was uploaded to CPAN on 02 Sep 02. New features include: A UserDir directive was added to support user home directories. (Thank Noam Solomon for the spot.) A UserSite directive was added to allow multiple users to share a common GID and site directory while denying each user's access to the other user sites. (Before each user required a unique GID.) More HtmlText directives were added to provide HTML just before each of the form objects. (Thank Noam Solomon for the suggestion.) During my haste, I left the internal version number set at 1.04b. This is a unique version number, so the error should not cause any problems. However, I hope to be patient and either provide a new update because any bug fixes reported or simply upload a subversion with the correct version number included (1.05) in the near future As always (except for the last 3 days when my ISP mysteriously reconfigured the DSLAM) a demo is available at: http://www.xorgate.com/Apache/OpenIndex/demo
RE: cpan hacked or april fools?
I'mamuzed!=) At 12:49 AM 4/2/2002 +0100, Jonathan M. Hollin wrote: Anyone looked at CPAN's front page today: www.cpan.org? It is either a pretty amusing April Fools joke or a hack. Fun either way! It's gotta be an April Fools gag (great one). The hackers almost always leave their stamp... Jonathan M. Hollin - WYPUG Co-ordinator West Yorkshire Perl User Group http://wypug.pm.org/ http://wypug.digital-word.com/ ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
If mod_perl module loaded from within httpd.conf
What is the best way to check from within the httpd.conf file if a mod_perl module is loaded. For C modules, I can use: IfModule mod_perl.c ... /IfModule to conditionally load it's directives. Is there a similar conditional syntax for the mod_perl script modules? I only want a module's directives to be processed when that module is loaded. ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
Re: PREANNOUNCE Tied Request and Cookies Module
At 10:18 PM 1/7/2002 -0500, Jay Lawrence wrote: I am thinking of names such as: Apache::Request::Tied or Apache::Args::Tied and Apache::Cookies::Tied and Apache::RequestFiles::Tied or possibly: Apache::Tied::Request, etc. Any interest, comments, suggestions, Personally, I like placing the most significant part first and using a verb, so how about: Apache::Tie::... and so forth? ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
Re: [VERY OT] How to Use Apache as a FTP server
I somehow missed the first part of this email thread (from anandr). The Apache::OpenIndex module (when loaded on the Apache server) provides a way to upload and download files using a http browsers. A demo is available at: http://www.xorgate.com/Apache/OpenIndex At 03:55 PM 12/19/2001 +0200, you wrote: That's not mod-perl. That's not even Apache. Many popular web browsers have browsing features for anonymous FTP. You need an FTP server for that. Apache 2 will include an FTP server (I think), but for now, try wuftpd. Issac anandr wrote: Hi All, I want to use my Apache as a ftp service also. Can the ring help me. I am running Apache 1.3.22 at port 80. when I give http:///, I get the list of files in my htdocs. Like wise when I give ftp:///, I need a list from my server folders. Have look at this address,for what I am intending to do,..ftp://tiger.com. With Regards, Anand Ratnasabapathy, DSM Soft(P)Ltd, No:25,Nungambakkam High Road, Chennai - 600 034. http://www.dsmsoft.com ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
ANNOUNCE: Apache::OpenIndex 1.03
OpenIndex provides a file manager for an Apache modperl web site from a web browser. OpenIndex also provides the same functionality as autoindex, therefore it can be used to both manage and navigate a web site. An Apache::OpenIndex (OpenIndex-1.03.tar.gz) update was uploaded to CPAN on 25 Oct 2001. The module can also be downloaded at: http://www.xorgate.com/Apache/OpenIndex/ OpenIndex now supports Cascading Style Sheets and individual HTML tab configurable attributes. For example, you can now add TAG attributes like: TAG attributes or add TAG text, like: TAG text /TAG Two new directives were added to support this feature: IndexHtmlTagtagattributes IndexHtmlText tagtext The following tags have attribute support: head style frameset noframe body table form font input textarea th tr hr h1 h2 h3 td a p The following tags have text support: frame noframe style The FoldersFirst index option was also added. Again, a demo is available at: http://www.xorgate.com/Apache/OpenIndex/demo Enjoy:) ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
Please ignore this message (just a test)
I previously sent two messages today and they did not appear on the list. So I am sending this message as a simple test. ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
ANNOUNCE: Apache::OpenIndex 1.03
OpenIndex provides a file manager for an Apache modperl web site from a web browser. OpenIndex also provides the same functionality as autoindex, therefore it can be used to both manage and navigate a web site. An Apache::OpenIndex (OpenIndex-1.03.tar.gz) update was uploaded to CPAN on 25 Oct 2001. The module can also be downloaded at: http://www.xorgate.com/Apache/OpenIndex/ OpenIndex now supports Cascading Style Sheets and individual HTML tab configurable attributes. For example, you can now add TAG attributes like: TAG attributes or add TAG text, like: TAG text /TAG Two new directives were added to support this feature: IndexHtmlTagtagattributes IndexHtmlText tagtext The following tags have attribute support: head style frameset noframe body table form font input textarea th tr hr h1 h2 h3 td a p The following tags have text support: frame noframe style The FoldersFirst index option was also added. Again, a demo is available at: http://www.xorgate.com/Apache/OpenIndex/demo Enjoy:) ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
ANNOUNCE: Apache::OpenIndex 1.02
An Apache::OpenIndex (OpenIndex-1.02.tar.gz) update was uploaded to CPAN on 12Oct2001. The module can also be downloaded at: http://www.xorgate.com/Apache/OpenIndex/ This release fixed two bugs: 1) the directory and location Menu OpenIndex directive only maintained one copy of the array. 2) The Copy command could not copy files with spaces in their name. A filter was added, to prevent non-admin users from creating any hidden files or directories. Again, a demo is available at: http://www.xorgate.com/Apache/OpenIndex/demo ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++ ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
[OT] Off Topic?
Is there anything special about putting things, like, [OT] in the subject line, other than being informative? The reason I ask, is that I would like to send [OT] messages to the most appropriate list. Is there a reference that provides a list of lists for apache.org that can be used to direct the message to the most appropriate list? ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
[OT] Overridding a module's use of a function
I discovered that File::NCopy uses the function glob to expand file
names. My problem is that I need to pass file names that have spaces in
them and glob does not process them. So I did the following override (I
do not need to expand the file names):
package Apache::AAM;
. . .
package File::NCopy;
use subs qw(glob);
sub glob {@_};
package Apache::AAM;
. . .
Is there a better way to do this?
++
| George Sanderson [EMAIL PROTECTED]
| http://www.xorgate.com
++
ANNOUNCE: Apache::RedirectDBI 0.02
I needed a module that would allow users' to login (authenticate) using a single HREF link and then be redirected to their individual web site (URL). I searched CPAN and found Apache::RedirectDBI. It was very close to what I needed, so I modified it so that it would read a URL field from the user's DBI record and do an external REDIRECT to their site location. In the process of upgrading the module, I thought to myself; self, it would be nice if the cpan released module was updated..., so I fired off an email to the author, Michael Smith. Mike replied: This is a bit of a blast from the past! I've not looked at it for quite some time (is it in my name on CPAN?) Anyway, very happy to get your changes in to what's on CPAN ... or better still, hand it over to you. So that is what I did. I contacted CPAN and uploaded RedirectDBI-0.02.tar.gz on 05 Oct. 02 to CPAN. (It will take a while for it to show up on CPAN.) The module is also available for download at: http://www.xorgate.com/Apache/RedirectDBI [I think that the Perl public community is really awesome. . .] ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
ANNOUNCE: Apache::OpenIndex 1.01
An Apache::OpenIndex (OpenIndex-1.01.tar.gz) update was uploaded to CPAN on 27Sep2001 and is currently available for cpan download. This release includes some addational admin autherization checks and the patch provided by Tatsuhiko Miyagawa, which allow the module to be compiled by Perl 5.005_03. Thank you Tatsuhiko! Again, a demo is available at: http://www.xorgate.com/Apache/OpenIndex/demo ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
RE: ANNOUNCE: Apache::OpenIndex
At 07:37 PM 10/2/2001 +0100, you wrote: OpenIndex provides a file manager for an Apache modperl web site using a web browser. Looks good George any plans to add better authentication (eg certificates) and to set it up for secure SSL file transfers? Clinton Gormley Authentication and any other Apache service (SSL) are outside of scope of OpenIndex. The reason why a fake directory (default .XOI) is used to access the site's files through OpenIndex, is to allow location and directory Apache directives to provide the other desired services, while they may not be required for the normal (http) site browser access. Hope this answers your question. ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
ANNOUNCE: Apache::OpenIndex
Apache::OpenIndex (OpenIndex-1.00.tar.gz) was uploaded to CPAN on 14Sep2001 and is currently released. This was my first module. I enjoyed journey. OpenIndex provides a file manager for an Apache modperl web site using a web browser. It is a extensive rewrite of the Apache::AutoIndex module (written by Philippe M. Chiasson), which in turn was a remake of the autoindex Apache module. OpenIndex also provides the same functionality as autoindex, therefore it can be used to both manage and navigate a web site. A demo is available at: http://www.xorgate.com/Apache/OpenIndex/demo ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
Re: Knowledge Base for 2.0
On Sun, 16 Sep 2001, Jim wrote: I'll try to spit out some more thoughts as they come. Please feel free to contribute. Now's the time to get the wish list put together. Some ideas (from the user's perspective): It would be nice if, the user interface provided a configuration and version form that the user could select for responses. For example, an article may cover all Apache versions between certain version numbers, or all versions above some value. It would be nice if, a sort priority were available. In other words, the date of the article or the keywords may be most important. (etc.) It would be nice if, the knowledge base could connect the information dots. In other words, would contain references (links) to related information. Of these, Guide, Debugging, and Archive references would be useful. It would be nice if, the information entered during a query could be used to post to an email list. In other words, if the user does not finding useful information, that it would be easy to post a message, that would get human feedback. ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
RFC: Apache::RedirectDBI mod
I put my modified version 0.02 at http://www.xorgate.com/Apache/RedirectDBI The original functionality was kept intact. This version allows a DBI field to contain the URI to redirect the client too, and also provides for external REDIRECT. Let me know what you think. I am willing to take over the model, if that is acceptable. -Original Message- From: George Sanderson [mailto:[EMAIL PROTECTED]] Sent: Thursday, September 13, 2001 12:12 AM To: Michael Smith; [EMAIL PROTECTED] Subject: Apache::RedirectDBI mod I needed something close to what Apache::RedirectDBI provided. I modified version 0.01 to allow it to get the URI from a DB field and also to do an external REDIRECT. The original functionality of the module should be maintained. Let me know if you are interested in looking at the code. Hi there, This is a bit of a blast from the past! I've not looked at it for quite some time (is it in my name on CPAN?) Anyway, very happy to get your changes in to what's on CPAN ... or better still, hand it over to you. Mike http://www.iii.co.uk Interactive Investor International is a leading UK Internet personal finance service that provides individuals with the capability to identify, compare, monitor and buy online a number of financial products and services. Interactive Investor Trading Limited, a subsidiary of Interactive Investor International plc, is regulated by the SFA. ++ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com ++
Re: ANNOUNCE: Apache::OpenIndex
At 03:18 PM 9/8/2001 +0800, you wrote: Apache::AutoIndex was never released to CPAN. Hehe, funny, it sure look like it's still there ;-) I'm sorry. Of course what I meant to say was: Apache::FileMan was never released to CPAN. Also, for any of you trying to login to the demo late yesterday at: http://www.xorgate.com/Apache/OpenIndex/demo I made a password capital letter typo in the README file (at about the same time I made the above typo) , which by the way, made all the difference in the world when trying to login. It was bad for only about an hour. +==+ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com +==+
ANNOUNCE: Apache::OpenIndex
Back on 13 Dec 00, I posted the creation of module Apache::FileMan.pm. Today, I ANNOUNCE that Apache::FileMan has been renamed to Apache::OpenIndex. The name change was made because, several software applications go by the same name, one of which, provides the same basic functionality. The name OpenIndex was selected because it clearly follows from the name, Apache::AutoIndex, from which it was derived. Apache::AutoIndex was never released to CPAN. I do plan to do a CPAN release Apache::OpenIndex on 14 Sep 01 (next Friday.) DESCRIPTION: OpenIndex provides a file manager for web sites through a web browser. It is a extensive rewrite of the Apache::AutoIndex.pm module (written by Philippe M. Chiasson), which in turn was a remake of the autoindex Apache module. It can provide the same functionality as AutoIndex.pm, therefore it can be used to both to navigate and manage the web site. The source is available at: http://www.xorgate.com/Apache/OpenIndex A full functioning demo is available at: http://www.xorgate.com/Apache/OpenIndex/demo Enjoy. The name change carries with it configuration directive and default value changes. I hope the changes do not cause more problems than they solve. All comments are welcomed. Thank you all for bearing with me during this long development cycle. +==+ | George Sanderson [EMAIL PROTECTED] | http://www.xorgate.com/http://www.xorgate.com/ +==+
Re: RFC: Apache::FileMan 0.07
It turns out that there are several software applications that go by the name FileMan. I have updated Apache::FileMan.pm to pre-release version 0.07. I substantially reorganized to code. FileMan provides a file manager for a web site through a web browser. It is an extensive rewrite of the Apache::AutoIndex.pm module (written by Philippe M. Chiasson), which in turn was a remake of the autoindex Apache C module. FileMan can provide the same functionality as AutoIndex.pm, so it can be used to both navigate and manage a web site. For example; Date: Sun, 02 Sep 2001 18:57:00 -0700 From: Alex Krohn [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: FileMan Hi George, I released an Apache modperl web file manager called FileMan on 13 Dec 00. It's demo is available at: www.xorgate.com/FileMan/demo I thing that we should resolve the name conflict, so that others will not get confused. I had been meaning to email you as I am on the mod_perl list and caught your announcement a while back. We've been using the name FileMan since our first release back in September of 1998. You can see us in third party listings like: http://cgi.resourceindex.com/Programs_and_Scripts/Perl/File_Management/ http://www.hotscripts.com/Detailed/237.html If there's any way you could change your projects name to something else, I'd really appreciate it. If you have any questions, please don't hesitate to ask. Cheers, Alex Here is my short list of URLs that reference other FileMan applications: http://www.gossamer-threads.com/scripts/index.htm http://business.fortunecity.com/hearst/805/fileman.htm http://thegururamana.enmail.com/Fileman.htm http://www.hardhats.org/fileman/u1/index.html http://www.tribbs.com/help/tri00018.htm http://www.suse.de/en/produkte/susesoft/linux/Pakete_prof/paket_xfileman.html The point is, that I think Apache::FileMan should be renamed, to something, which easily distinguishes it from the others, yet at the same time, keeps the name short and to the point, before releasing it to CPAN. Here is a short list of names that might work. NameComment == FileMan.pm Of course, pm stands for Perl Module (not that different) OpenIndex This name maintains Index from AutoIndex. XorfMan Just a thought. FileMod Same initials, Please comment if you have any preferences or other recommendations. Also, please respond and let me know what you think of the current code: good, bad, or ugly. Thank you.
Re: [OT] Reference to a method using a scalar?
Thank you, Andrew.
Andrew, pointed out that Perl provides a way to delay the resolution of a
method.
Date: Sun, 19 Aug 2001 20:04:12 -0700 (PDT)
From: Andrew Ho [EMAIL PROTECTED]
George,
GSIn other words, I would like all my reference to be hard method
GSreferences which are developed from string scalars at run time (delayed
GSresolution and later dereferenced).
Perl allows you to delay resolution of a method until run-time:
$method = 'foo';
$object-$method($a, $b, $c)
Humbly,
Andrew
--
Andrew Ho http://www.tellme.com/ [EMAIL PROTECTED]
NOTE:
Above, I should have said subroutine in place of the word method, since
method implies OO referent calls. Also, I did not understand delayed
resolution. (My mistake.)
I choose to implement the following (snip):
no strict 'refs';
$Commands{$command}{cmd}=\$routine; # $routine contains a string
use strict 'refs';
unless(defined {$Commands{$command}{cmd}}) {
delete $Commands{$command}; # This is bad, so throw it away!
warn $routine routine is not defined;
}
for a couple of reasons.
1) The methods are not OO (yet).
2) The above assignment only happens during initialization.
[I learned a lot about OO Perl.]
Again, thanks, [done]
+===+
| George Sanderson [EMAIL PROTECTED] |
| http://www.xorgate.com/www.xorgate.com |
+===+
RFC: Apache::FileMan 0.07d
I have updated Apache::FileMan.pm to pre-release version 0.07d. I substantially reorganized to code. FileMan provides a file manager for a web sites through a web browser. It is an extensive rewrite of the Apache::AutoIndex.pm module (written by Philippe M. Chiasson), which in turn was a remake of the autoindex Apache C module. FileMan can provide the same functionality as AutoIndex.pm, so it can be used to both to navigate and manage a web site. This update includes HTML frame support, import of arbitrary menu commands, and the HOLD directory was restructured to allow it to manage any directory on the web server. The source is available at: http://www.xorgate.com/FileMan A demo is also available. I need help from others to fill in the FileMan.dic file for Apache::Language with the language phrases. I included all the English entries. I would like to release FileMan to CPAN. Please provide any comments, suggestions, or feedback. Thanks
[OT] Reference to a method using a scaler?
I have a Perl scalar ($mystr) that holds a string. I want to use the
scalar to form a hash reference to a method after the Perl code is compiled.
For example, when I know the method name before compile time, I can say:
use strict;
%Hashref=( Entry1={meth1=\method1, do='something'},);
and then later execute method1 by using:
$Hashref{Entry1}{meth1}($a,$b,$c);
How can I assign the hash method reference using the contents of the scalar
$mystr, in such a way, that I can use the same syntax to execute the method
name stored in $mystr.
NOTE: I have already used the following, but would prefer to use method
references for both:
$Hashref{Entry1}{meth1}=$mystr;
...
if(ref($Hashref{Entry1}{meth1}) eq 'CODE') {
$retval =$Hashref{Entry1}{meth1}($a,$b,$c);
} else {
no strict 'refs';
$retval={$Hashref{Entry1}{meth1}}($a,$b,$c);
use strict 'refs';
}
Thanks in advance.
[ANNOUNCE] Apache::FileMan 0.06c
FileMan provides a file manager for a web sites through a web browser. It is a extensive rewrite of the Apache::AutoIndex.pm module (written by Philippe M. Chiasson), which in turn was a remake of the autoindex Apache module. FileMan can provide the same functionality as AutoIndex.pm and can be used to both navigate and manage the web site. This version (0.06c) corrects an Apache::Language problem displaying 1's. Special thanks goes out to "Aaron Johnson" [EMAIL PROTECTED] and his co-worker Ken Clark (who work for ProVillage), for finding the Language problem and making performance suggestions that have been incorporated. The latest release can be found at: httpd://www.xorgate.com/FileMan An on line demo is available at: httpd://www.xorgate.com/FileMan/demo If you have any questions, comments, or success stories please provide it to me, so that a good user base can be established before committing the code to CPAN.
Apache::Language STORE needs a language
Problem solved: I am using the latest version of Apache::Language I noticed that the Apache::Language.pod says that the PlainFile format is: phrase-id:lang-type While the PlainFile.pod says: lang-tag:phrase-id The first one seems to work. It works like Apache::Language::PlainFile describes: (lang-tag:phrase-id). Also I am getting the following error: warn] [Apache::Language (111)] STORE needs a language specification to work for each phrase looked up. After these messages I always get the dreaded: child pid 12280 exit signal Segmentation fault (11) My problems was that I did not have the negotiation_module loaded and Options +MultiViews specified in the httd.conf file.
Apache::Language STORE needs a language
I am using the latest version of Apache::Language I noticed that the Apache::Language.pod says that the PlainFile format is: phrase-id:lang-type While the PlainFile.pod says: lang-tag:phrase-id The first one seems to work. Also I am getting the following error: warn] [Apache::Language (111)] STORE needs a language specification to work for each phrase looked up. After these messages I always get the dreaded: child pid 12280 exit signal Segmentation fault (11) The the page then displays OK. Any solutions?
RFC: Apache::FileMan 0.06
I have updated Apache::FileMan.pm to pre-release version 0.06. FileMan provides a file manager for a web sites through a web browser. It is an extensive rewrite of the Apache::AutoIndex.pm module (written by Philippe M. Chiasson), which in turn was a remake of the autoindex Apache C module. FileMan can provide the same functionality as AutoIndex.pm, so it can be used to both to navigate and manage a web site. The source is available at: http://www.xorgate.com/FileMan A demo is also available. I need help from others to fill in the FileMan.dic file for Apache::Language phrases and nomenclature. I have included all the English entries, but have not a clue what to put in for the other languages.-) Hopefully, I will be able to release this to CPAN in the near future. Please provide any suggestions, feedback, or what ever. Thanks.
Re: Apache FileMan
At 02:25 PM 2/11/2001 +, you wrote: George, Just had a peek at your FileMan demo. Looks pretty neat. I too was about to extend/rewrite Apache::AutoIndex to do some similar things. A few questions Does it use Apache::AutoIndex at all or is it a total rewrite ? There was a lot to do, so I did an extensive rewrite and tried to maintain the existing functionality. What versions are you using ? I'm running perl 5.6.0, mod_perl 1.25 and apache 1.3.17. I just tried using AutoIndex and while it built OK, I got a child seg fault (11) when I tried to access a directory. Whats your setup ? I am similar but at Apache 1.3.14 and mod_perl 1.24_01. Try setting the following directive and let me know if you have better success. FileManOptions language 0 I will be updating Apache and mod_perl soon, but I also had seg faults when using language. However, I have been updating the code to include more Apache::Language support, so I will be looking at this more. How extensible is FileMan ? (I'm about to get the source). The thing I wanted to do over AutoIndex was have control over the Description field for each file. I want this to be a link to an RDF search engine I'm thinking about, so each file would have a description like ( e.g. dir = /bar/baz ) I have not over ridden any of the functions within the module, but with the latest code, I made the arguments to the internal menu commands have all the same format. My goal is to make it so the module OO in the future, but I doubt that it is there today. The description field works fine today, however. Just try the classic autoindex directives, such as AddDescription "HTML text" .html .htm Also a couple other points/suggestions. - I was thinking about having the directory path across the top (where you have "Directory index of x/y/z " ) into links also, like Directory index of a href='/x/' x /a / a href='/y/' y /a / z which would allow rapid movement around directory tree instead of popping up via parent. I am a little slow with suggestions, (it just a personal thing). Do you mean to have a go to root link? (Kind of like a go to top for a long page.) Or are you suggesting dynamic book marks or some thing alone those lines? - I'd like the option to apply a CSS link to the directory page What is a CSS link? - Is it because its a demo or a fundamental aspect of the module that when you go to the directory itself as a URL you don't get the FileMan menu structure / form, i.e there is a distinction between /demo/foo/ and /demo/.XFM/foo#root. Could the FileMan page be made the Directory Index or am I missing something ? The module uses the fake .XFM directory so that you can provide authentication when accessing the root. In other words, lets say you wanted to set FileMan up to access http://www.website.com/users/john You would have to access http://www.website.com/users/john/.XFM in order to activate the file manager functionality. This way if you require authentication, you will need to apply it to http://www.website.com/users/john/.XFM, that way hopefully only john will have access to his files. Anyway I might install your latest version and give it a go. I will have an update available in about two days.
Re: httpd keeps crashing overnight
Hi George, Thanks for the reply... My problem is that the mod_perl httpd is sometimes crashing overnight. In the last three days, it has mysteriously crashed twice. When I restart it with "apachectl_modperl start" (apachectl_modperl is just apachectl but with the config file path set differently), it comes up with no problem, but I suppose it might crash again in the future. How long had it been running OK before you started having problems? Did something change just before the problem started occurring? Previously, I only had one mod_perl httpd running on the system. I split it into a non-mod_perl httpd and a mod_perl httpd because the system was running out of memory. This change happened 4 days ago. Before that, I did not have this crashing problem. What ports are you using for your httpd that does and does not have the problem? Both httpds listen on port 80. The mod_perl enabled one listens on 216.74.79.145:80 and 216.74.79.194:80, while the non-mod_perl enabled one listens on port 80 of all other IP addresses on the machine. Is there any indications in the access_log at about the time of the crash? 203.177.3.11 - - [19/Jan/2001:05:05:21 -0500] "GET /anime/seraphimcall/ HTTP/1.1" 200 3256 "http://www.animelyrics.com/anime/_S" "Mozilla/4.0 (compatible; MSIE 5.0; Windows 98; DigExt)" 207.35.188.14 - - [19/Jan/2001:08:40:20 -0500] "GET / HTTP/1.0" 200 6205 "-" "Mozilla/4.7 [en]C-CCK-MCD (WinNT; U)" Those are the two log entries for animelyrics.com before and after the crash; I don't see anything unusual. I also looked at slayers.aaanime.net: 24.67.224.12 - - [19/Jan/2001:05:03:24 -0500] "GET /~linazel/fanfics/fanfic.asp?fanfic=nobilitypart=10 HTTP/1.0" 200 15547 "http://slayers.aaanime.net/~linazel/fanfics/fanfic.asp?fanfic=nobilitypart =9" "Mozilla/4.0 (compatible; MSIE 5.0; Windows 98; DigExt; AtHome0107; sureseeker.com; Hotbar 2.0)" 172.133.91.48 - - [19/Jan/2001:08:35:32 -0500] "GET /~linazel/banners/ybanner.gif HTTP/1.1" 200 20544 "-" "Opera/5.02 (Windows 98; U) [en]" I don't see anything out of the ordinary here, either. Perhaps you could run a cron job to scan the processes in order narrow down the exact time of the problem. What would I be looking for? Is there any indication of a burst load (or a similar pattern) just before crash? Is there a back end data base involved? It took about 4 hours before the httpd process was restarted. It would be nice to know how long after the last request the httpd root process crashed. If a cron job ran once a minute to scan for the httpd root process and report when it disappears, it might be a clue as to the nature of the problem. In the report you might want to include information about the last 10 minutes (last 10 scans from temp files 1 through 10) of all the httpd process running via the `ps -gaux|grep httpd`. It would be interesting to know how many httpd process were running and also what `vmstat` had to say before and after the crash. Often when trying to solve an intermittent problem, it is good, if you can duplicate the problem a will. The information obtained about the problem should help you to achieve this. For example, the access_log indicates that the last browser access before the problem were both: "Mozilla/4.0 (compatible; MSIE 5.0; Windows 98; DigExt;". However, it is difficult to tell if this is just a coincidence or not until a pattern can be established.
Re: [Fwd: AuthenDBI idea]
At 08:37 PM 1/15/2001 -0800, you wrote:
On Mon, 15 Jan 2001, Edmund Mergl wrote:
any comments ?
[count number of times a user has logged in and such things]
Hope I am not out of place here, and also that the ideas are generic enough
to be applied to a wide number of authentication requirements.
Here are two ideas.
I.=
The first idea for authentication:
Provide a directive to perform a comparison on any or all fields of the
current user's record. If the comparison is true, provide a URL to
REDIRECT the original request.
The supporting directives could be something like:
Auth_DBI_comp {regexp}
Auth_DBI_url"http://www.redirect.com/ok/"
Where regexp is a comparison string and url is where to REDIRECT the user
if the comparison is true.
The original request URL should be passed as a url ? argument, so that a
REDIRECT cgi target script could determine the original requested url. The
target script could update any fields as required.
The regexp needs to be able to easily access any arbitrary field values for
the current user's record. Perhaps simply by pre appending a '$' to the
field name. For example:
Auth_DBI_comp {$username='xyz' $usecount4}
This would REDIRECT every login with field "usercount" less than 4 for
field "username" equal 'xyz'.
A pass and fail condition would also be needed, perhaps just designated as
PASS and FAIL.
Being able to specify multiple conditions per authorization attempt would
be useful.
II.
A second idea (for authorization) is to provide a generic way to set an
Apache environment variable with the contents of any field for the current
user. For example:
Auth_DBI_env field1,field2
This would set two environment variables called: "FIELD1" and "FIELD2" with
their field data content of the current user's record.
I suppose if the data base had multiple records for a user, then the
environment variables would contain a list of values.
Re: FileMan - Not enough arguments for mkdir
At 08:59 AM 1/17/2001 +0200, you wrote: Hi, This is me again. Thanks for quick response. Another two questions: in your demo http://www.xorgate.com/FileMan/demo/.XFM/ just tried to upload file "1", it reported me "ERROR: MkFile: Parent access denied" but I suspect it catchup to do open() before... so you can see now "1" with 0k size. a hole? - someone maliciose can create as many files as he wants this way... Fixed. The parent directory check was being checked after the open. The source is available from: http://www.xorgate.com/FileMan as file FileMan-0.05a.tar.gz Thanks and another one. I tried to install it(0.04 version), and after a hour of struggle it got it compiled and make test passed, but I can't get it running in apache(1.3.14+1.24_01). Sorry about the struggle part. This is my first module. Since you are just installing this as a stand alone module, the other Perl module dependances are not being automatically install for you like hopefully CPAN will be doing for you once the code get released. Others may have some advise to me about how CPAN determines the Perl module dependencies and automatically install them. You need the following other Perl modules installed (CPAN works great for this.) use Archive::Zip use File::Copy use File::Path use File::NCopy use Data::Dumper use Storable use HTML::HeadParser
Re: FileMan - Not enough arguments for mkdir
At 03:12 PM 1/16/2001 +0200, you wrote: Some linuxes required to have correct permission set in mkdir (it fails on mkdir $dir; needs to be mkdir $dir,0775;), so make test fails... (FileMan.pm line 771 and several times below..) I fixed all the mkdirs with "mkdir $dir,0755". I think 0755 is all that is required. Let me know if this doesn't fix your problem. Thank you for your input.
Need: Autherization one time password.
I want an authorization module to generate a one time password for a particular user name. For example, lets say I target some user names to have a one time password service by using a directive like: Auth_onetimeruser1, user2, ... If the password was blank (empty), the module would automagicly create and store a new password per some format string (perhaps): Auth_password_format"hello%16s" After a user logs in some specified number of times (default 1), the password would be blanked out. Of course, someone would have to be able to view the account information, but this should be different module. A module that would allow an "administrator" to view and change the user account data base.
Re: Need: Autherization one time password.
At 11:08 PM 1/14/2001 +, you wrote: I seem to recall Randal did a column in Web Techniques on just such a thing. Check http://www.stonehenge.com/merlyn/http://www.stonehenge.com/merlyn/ I want an authorization module to generate a one time password for a particular user name. I think I found the article: http://www.webtechniques.com/archives/2000/05/perlhttp://www.webtechnique s.com/archives/2000/05/perl/ Self-Registering Password Protection, Part 1 Randal's requirements were some what similar. I wanted to randomly generate the password for a particular user after they log in a limited number of times and then send email to someone. After reading the article, I thinking the best way to implement my requirements would be to create (or modify) an authentication module that would keep track of how many times a user logs in and also be able to trigger (REDIRECT to) a cgi script when certain conditions are meet. My cgi script (custom requirements) could then regenerate the new password and do any the notifications required. The authentication trigger (REDIRECT) would be the trick, in that it should be generic enough to allow for a whole range of other situations.
Re: [ANNOUNCE] Apache-AuthzCache 0.03
When a request that requires authorization is received, Apache::AuthzCache looks up the REMOTE_USER in a shared-memory cache (using IPC::Cache) and compares the list of groups in the cache against the groups enumerated within the "require" configuration directive. If a match is found, the handler returns OK and clears the downstream Authz handlers from the stack. Otherwise, it returns DECLINED and allows the next PerlAuthzHandler in the chain to be called. After the primary authorization handler completes with an OK, Apache::AuthzCache::manage_cache adds the new group (listed in REMOTE_GROUP) to the cache. I would like the module to be able to cache selected environment variables along with the user and group information. A directive could be used to select a list of environment variables to be cache. For example; AuthzCacheOption CacheEnv REMOTE_EMAIL SPECIAL_VAL FIELD23 Then when the following authorize request are processed, these environment variables would be set. The down steam handlers can now access the variables without having to go back to the original data source (which may change). I have done some work with AuthzDBI.pm to provide this functionality. I hope to work with Edmund Mergl to get it added to Apache::AuthzDBI.pm.
Re: Whither gozer?
Philippe M. Chiasson, the mail address given in Apache::AutoIndex is bouncing (today, at least): [EMAIL PROTECTED] If you are reading this and have a working address, please let me know. I have been tring to get in touch with Philippe for about 6 months with no success.
Re: Debugging information sent to browser
Why do you think it would be hard to return AUTH_REQUIRED if the
DEBUG=1 param is in the URL? Granted, the browser issues involved won't
make it the best solution, but no worse than passwords are already. Someone
would have to be using a machine where a valid uid/passwd are in browser
cache,
or someone would have to know a valid uid/password.
I like your idea. Yes this should be no problem. How I think I will do it
is to
require a login before presenting the DEBUG submit button. That way it won't
confuse or frustrate the unwary.
My main worry is how I'm going to produce the best debugging information
without requiring scripts to be rewritten. As it is, CF displays debugging
information based on IP. I can mimic the code from cgiwrap to find what I
would like to send, and I can probably find a way to send the best debugging
and error reporting to the debugging uid's browser.
Try something like:
if($debug) {
use Data::Dumper;
print "PRE";
print "\%list\n";
print Dumper \%$list;
print "/PREHR";
}
Searching the mod_perl archive works
I had a problem and first tried: http://perl.apache.org/guide/perl.html It said to search the mod_perl archive, but it did not provide it's URL (or I did not find it). So then I did a search engine search for "mod_perl archive" and found: http://www.bitmechanic.com/mail-archives/modperl/ Once there and after searching for a while (using the "Advanced search":), I found information that help me solve my problem. It would have been nice if I did not have to resort to the search engine to locate the archive. Also, I looked for a mod_perl function (method) perldoc type of reference doc, but did not find any easy (URL) references. What I had hoped to locate was a page where you enter the mod_perl function (method) name and it provides the perldoc equivalent reference infomation for the function.
OT: HTML/XHTML syntax checking tools?
What HTML/XHTML syntax tools are recommended? I was hoping that the tool would run on the browser (client) side. I prefer a PC version, but hey, a good ap is best.-) For example, I am working on a mod_perl module, and I made a HTML typo (TEST instead or TEXT), (belzlare slip). Perhaps I or someone else has made other mistakes. Since the browser ignored the TEST tag, and the default was reasonable, it looked OK. Now I think I should be checking with something other than my eye.
RFC: Apache::FileMan pre-release 0.04
FileMan provides a web site file manager to web browsers. It is an extensive rewrite of the Apache::AutoIndex.pm module (written by Philippe M. Chiasson), which in turn was a remake of the autoindex Apache module. FileMan can provide the same functionality as AutoIndex.pm which means that it can be used to both navigate and manage a web site. The current pre-release version is 0.04 and is available for download at: httpd://www.xorgate.com/FileMan A restricted demo is available by accessing: httpd://www.xorgate.com/FileMan/demo/.XFM user name=fileman password=13dec00 In version 0.04, I added the "Edit", "MkFile", and "Revoke" (only for admin users) commands. This completes the development of the module. The command list includes: "Browse", "Upload", "Unzip", "Delete", "Mkdir", "MkFile", "Copy", "Move", "Edit", "Rename", "Help", "SetGID", and "Revoke". If you notice any problems or have any comments, please, let me know. All comments are welcomed.
Re: RFC: Apache::FileMan pre-release 0.04
At 09:19 PM 1/7/2001 -0600, you wrote: I just tried the demo (coursoity got the better part of me ;) and the username: fileman password 13dec00 did not work Sorry, it appears, I left the demo access wide open and the directory got deleted. Try it again. It should work. (Its now read only though.)
mod_perl large scalar cache
I have a mod_perl hash scalar that can store up to 4 MB of string data.
I was wondering what happens to the memory usage after a "delete
$myhash{large}"?
I was thinking (I know this can be dangerous:-), it would be nice if I
could tell mod_perl to cache scalars that become larger than some specified
value, so that they do not chew up too large chunks of memory.
Re: HTTP authentication based sessioning with logout ability
At 06:28 PM 12/19/00 +, you wrote: Hi there. I've been trying to write a system to perform authentication using the www-authenticate (http authenatication) method. However, I need a client visiting the page and having been authenticated to be able to logout and have their browser forget the information (or at least be able to force the browser to re-authenticate within the same browser session). There are two phases of Apache user authorization, namely, authentication and authorization. I think that you need to set a flag when the user logs-out to force authorization to reject requests and force a new login (which would then clear the flag on success). The flag could also be tied to a timer such that if no request was received after a defined time span, the flag could be set for that user. This approach would not require a cookie. I have never done this, but it should work.
Re: RFC: Apache::FileMan
At 01:40 PM 12/18/00 -0800, you wrote: you should take a look at the interface of the file management of some of the free webspace providers. I have looked at some of these. They do not look and feel like a file manager. They tend to be fragmented and wordy. i haven't really looked at the Apache::FileMan module, but the biggest thing i would suggest is using something like the Template Toolkit to move the html pages into templates. I am not there yet. FileMan does not require any client software and it is independent of the file content. It just doesn't care what is in a files or directory. (personally, i've moved on to using dav (http://www.webdav.org/) for most of my web-based file management needs.) I know very little about WebDAV. I looked at the site and found the DAV concepts interesting. Perhaps you could point me to a demo or more information about how the requirements are implemented. WebDAV appears to integrate the CASE information within the XML documents themselves. I noticed that WebDAV provides file sharing. I had not considered this, but it would be easy to do in FileMan.
Re: RFC: Apache::FileMan
At 04:55 PM 12/18/00 -0800, you wrote: On Dec 18, George Sanderson wrote: At 01:40 PM 12/18/00 -0800, you wrote: a matter of taste, of course. it also depends on your target audience. The user interface is the most sacred information ground. It's application depends on the requirements (marketing). A benefit gained from a consistent and portable interface is that the user can progress from system A to system B without having to navigate (interface) differently. The seamless interface that DAV provides is it's biggest advantage, because to the user it has the same interface to different (servers) services. Which is why I thought, that since FileMan looks and feels like autoindex it would be a natural extension for Apache users. biggest thing i would suggest is using something like the Template Toolkit to move the html pages into templates. http://www.template-toolkit.org/tpc4/paper.html#cgi (well, the whole document, really) is a good place to read about what i'm talking about. After reading the first part of the document, it made a lot of sense. (Thanks) well, webdav is an extension to http that allows you to "put", "delete", "move", "copy", etc You have given me a lot to think about. I need to read RFC2518 a couple or three times. http://andrew2.andrew.cmu.edu/rfc/rfc2518.html
Re: Email (mod_perl) Apache module?
At 09:32 AM 12/15/00 -0300, you wrote: Jeremy Howard wrote: IMHO, the best open source WebMail servers are PHP based I have this dangling idea of building a TWIG lookalike (in Perl), with a 'plug-in'/'module' structure, so I may write the email client, and others fill with their desired modules. Anyway, it's a seriuos undertaking, but it's in my plans to rip as much code and design choices from stable OS webmails as possible. I test drove the WebMail server, and, ... it was nice! What I had in mind was a simple stand-alone (with a robust email server API: POP IMAP, NEWS) Apache mod_perl email module, that would allow authorization, read, compose, send, forward, and delete. I don't think it would need to read attachments, but being able to save and attach files would be useful. A simple directory interface would also be useful but not hard requirement. An eligant simple interface would cover about 90% of most needs. Having it simple and easy to install is key.
RFC: Apache::FileMan
I have been getting some good feedback on the Apache::FileMan module. FileMan provides a file manager for a web sites through a web browser. It is a extensive rewrite of the Apache::AutoIndex.pm module (written by Philippe M. Chiasson), which in turn was a remake of the autoindex Apache module. FileMan can provide the same functionality as AutoIndex.pm and can be used to both navigate and manage the web site. I added a destination text field, which helps make the commands more intuitive. I also added a "View" command that allows you to view files (MiME) within FileMan. Now the command list includes: "Browse", "Upload", "Unzip", "Move", "Copy", "Rename", "Delete", "Mkdir", "View", "Help", and "SetGID". A demo can be viewed at www.xorgate.com/FileMan/example/.XFM user name=fileman password=13dec00 A new pre-release (version 0.02) tar ball file can be down loaded from www.xorgate.com. On Wed, 14 Dec 2000 Stas wrote: I already tell you that FileMan is not a very good choice as it reminds me of man page utility... Apache::FM Or Apache::FileMgr sounds like a better choice. But let the list do its job. Well I like the FileMan name, and I did a cpan i /fileman/ and i /FileMan/ with no response. I plan to add an "Edit" command which I will provide an on-line text file (HTML) editor from within FileMan. A Michigan friend suggested learning aids, "FileMan for Dummies" interface. All comments welcomed.
RFC: Email (mod_perl) Apache module?
I think it would be cool to have an email Apache module. I was thinking that if the URL was something like: http://www.site.com/user/mail it would activate the module. The module would allow the users to read and send email. Kind of like how I did Apache::FileMan (an Apache web site file manager). This would allow sites to get email interfaces up and running very quickly and provide a common user interface. Now that would be advocacy;-)
RFC: Apache::FileMan
I have created a new module that I called, Apache::FileMan.pm. FileMan provides a file manager for a web sites through a web browser. It is a extensive rewrite of the Apache::AutoIndex.pm module (written by Philippe M. Chiasson), which in turn was a remake of the autoindex Apache module. FileMan can provide the same functionality as AutoIndex.pm and can be used to both to navigate and manage the web site. Hopefully this module will open a new realm for web service providers (WSP). This is my first Apache module. I am not quiet sure what others will expect me to do, nor how I should provide the code to the mod_perl user community for inspection. The module has some interesting aspects, such as how to provide file permissions (security) in a shared environment.
Re: RFC: Apache::FileMan
Stas Bekman [EMAIL PROTECTED] At 11:11 PM 12/13/00 +0100, you wrote: On Wed, 13 Dec 2000, George Sanderson wrote: At 10:29 PM 12/13/00 +0100, you wrote: Do you want me to email you a *.tar.gz file? No. Please post the URL. http://www.xorgate.com/FileMan/FileMan-0.01.tar.gzhttp://www.xorgate.com/ FileMan/FileMan-0.01.tar.gz What do you mean by having aspects? Is it a question, or you does the module provides all the answers? I set the GID on the files and directories for the users that are authenticated. Also the GID of the files and directories are check before any significant commands are executed. Hmm, are you running the server as root? How can you set different GID than. No, Apache runs with a common "webuser" ID, just like normal Apache children. All I do is create a new UNIX /etc/group and make webuser a member of that group. When Apache creates files via FileMan, their UID is "webuser" (Apache's PUID). When a user is authorized, FileMan picks up a GID from an Apache environment variable for that user. Since Apache owns the file and is a member of the group, it can change the GID of the files. I'm investigating multiple membership and how that might be done. May be you should be more explicit in your question? What are you trying to accomplish and you don't get it? This is not really a question. I am just mumbling. Currently I have only two levels, of access control in FileMan, namely, "admin", and "user". I think there is a need for an intermediate level, that is, being able to run all the commands, but not able to change GIDs of the files.
Re: RFC: Apache::FileMan
At 02:32 PM 12/13/00 -0800 Carlos Ramirez [EMAIL PROTECTED] wrote: Looks very nice and useful! Does it use the default apache icons when the files are listed? when do you plan to make it available to the public on CPAN? docs? Yes, it does use the apache icons, just like Apache::AutoIndex.pm. My intention is to make it available on CPAN. I have the docs, test, and etc., however I have never released anything to CPAN before, so I am sure I am missing something. I put a copy of the current package at: http://www.xorgate.com/FileMan/FileMan-0.01.tar.gz
Re: Changing a file's UID from within an Apache module?
Prior to creating any file, the user would have been authenticated and an authorize module would have supplied the UID to be used for the request in an environment variable. The location of the user's files has to read by Apache so such that it can serve the files (HTML). The rational behind setting the UID of the files (and directories) is to allow each user to be able to create files that are protected from other users who may also be providing files. Each users files are confined to certain sub-directory stubs. Perhaps, I should use the GID instead of the PID to differentiate between users and leave the PID alone. Then perhaps setting the directory permission to 0770 and 0460 for files on creation, along with checking the GID of the files and dirs before any modifcation, could be used to achieve the same effect. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Changing a file's UID from within an Apache module?
I am creating a mod_perl Apache module. The last functionality that I need to implement is the ability to set the UID of the files and directories which are created and modified by the Apache module to something other than Apache's child UID. For example, if when I built Apache 1.3.14 on Linux 2.2.17, I used: APACI_ARGS=--server-uid=300,--server-gid=300 Later, my module creates a file "myfile.html". I want to change the file's UID from 300 to 3000, from within the Apache module. I was going to set the file permissions to 750 and keep the Apache GID. The directory permissions could be set to 770 in order to let the Apache module create, modify, and delete files within the directories. In this way each file would provide exclusive (modify and create) access rights in a shared user environment. Any ideas about the best way to change the permissions and UID? - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Apache::AuthenDBI request request
I need to access DBI table fields in the AuthenDBI/AuthzDBI data base. I think a good way to do this, would be to provide a new directive which would specify the table field values to be placed into the Apache notes. For example: Auth_DBI_note_field UID would add a key/value pair (field_name/value) into the Apache notes for each request accepted (not rejected) by Apache::AuthzDBI. If more that one field name directive was provided, each would then in turn have a notes key/value entry. In this way, a module down stream could retrieve the field/value pair from Apache notes to process their requirements. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re:coredump on Carps in startup.pl
Hey thanks a lot. I'm just going blind or something, because when I see a URL and click... If it quacks, it's a duck :-). But seriously, I have had the problem that this patch is suppose to fix. I do appreciate your efforts. I'll let you know how it turns out. Thanks. At 04:34 AM 11/21/00 +, you wrote: OK, now I'm using a mailer that doesn't insert newlines, so you should find this hyperlink clickable. I've cc'd myself so I'll check that it works. Here it is: http://forum.swarthmore.edu/epigone/modperl/sningvezerd/26091135.MAA145 [EMAIL PROTECTED] -- Jeremy Howard [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [OT] $r-header_out(Location=$dst) browser response
Follow up:
Well I screwed up, but at least now I know it!
The 303 status code did give me the result I wanted, but after I removed
the bug which caused the problem, I no longer need to use it. (I had a
global variable which was previously set and latter caused my REDIRECT to
cycle right back to the previous REDIRECT location. Kind of like a,
"REDIRECT pong bug". :-).
Sorry to have bothered you about this one, but I did learn something about
the http status codes and such.
| Is there a way to tell the browser(s) not to make this same substitution
in
| the near future?
| Or, is there a better way to do a one time only redirect from within a
| dynamic page?
I don't expect IE to do anything by the rules, but maybe you can experiment
with sending an explicit Status: 303, as described on
http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html#sec10.3.4
Yes, the 303 status code is exactly what I needed. I also tried setting
the "Expire" time to a negative and a small positive time in conjunction
with the "normal" REDIRECT (302) status return code, but IE 5 still did not
hack the response favorably.
I hope that SEE_OTHER {302} can be added to mod_perl's:
FakeRequest.pm
for future reference.
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: [OT] $r-header_out(Location=$dst) browser response
At 11:02 PM 11/19/00 +0100, you wrote:
| Is there a way to tell the browser(s) not to make this same substitution
in
| the near future?
| Or, is there a better way to do a one time only redirect from within a
| dynamic page?
I don't expect IE to do anything by the rules, but maybe you can experiment
with sending an explicit Status: 303, as described on
http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html#sec10.3.4
Yes, the 303 status code is exactly what I needed. I also tried setting
the "Expire" time to a negative and a small positive time in conjunction
with the "normal" REDIRECT (302) status return code, but IE 5 still did not
hack the response favorably.
I hope that SEE_OTHER {302} can be added to mod_perl's:
FakeRequest.pm
for future reference.
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
[OT] $r-header_out(Location=$dst) browser response
I know this is not directly a mod_perl issue, but I did use mod_perl to
present the issue. . .
So here it is:
When I issue: $r-header_out(Location="$dst"),
The browser (IE 5 in my case so far) goes to the fully specified $dst URL.
My problems is that, if the browser attempts to go back to the originating
(dynamic) URL, it again goes to the previous $dst location. (It has either
cached or remembered the new Location.)
I have also set:
$r-no_cache(1);
before:
$r-send_http_header("text/html");
Is there a way to tell the browser(s) not to make this same substitution in
the near future?
Or, is there a better way to do a one time only redirect from within a
dynamic page?
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Apache::AuthenDBI NOTICE message
So far, using Apache::AuthenDBI, I get an error_log message:
NOTICE: Unrecognized variable client_encoding
Every time the DBI ping is called.
I have not been able to tack down the origin of this message.
Any ideas or suggestions?
never seen before. What database do you use ?
I am using PostgreSQL 7.0.2
More details. . .
I am also using the latest Apache::AuthenDBI and Apache::AuthzDBI.
In startup.pl I have:
Apache::DBI-connect_on_init("dbi:Pg:dbname=atable", "webuser", "webuser",
{
PrintError = 1, # warn() on errors
RaiseError = 0, # don't die on error
AutoCommit = 1 # commit executes immediately
}
);
#
I did try seting PrintError=0 with no change on the NOTICE message.
Note: I am not using any of the following (in startup.pl):
Apache::DBI-setPingTimeOut("dbi:driver:database", $timeout);
Apache::AuthDBI-setCacheTime(0);
Apache::AuthDBI-setCleanupTime(-1);
Apache::AuthDBI-initIPC(5);
When I start httpd I get the following error log output.
==
[Wed Nov 15 17:47:48 2000] [info] created shared memory segment #12928
[Wed Nov 15 17:47:48 2000] [notice] Apache/1.3.14 (Unix) mod_perl/1.24_01
mod_layout/2.8 configured -- resuming normal operations
[Wed Nov 15 17:47:48 2000] [notice] suEXEC mechanism enabled (wrapper:
/usr/local/apache/bin/suexec)
[Wed Nov 15 17:47:48 2000] [info] Server built: Nov 5 2000 10:26:07
12052 Apache::DBI PerlChildInitHandler
12051 Apache::DBI PerlChildInitHandler
12053 Apache::DBI PerlChildInitHandler
NOTICE: Unrecognized variable client_encoding
Database handle destroyed without explicit disconnect at
/usr/local/lib/perl5/5.6.0/site_perl/Apache/DBI.pm line 140.
require 0 called at
/usr/local/lib/perl5/5.6.0/site_perl/Apache/DBI.pm line 140
Apache::DBI::childinit('Apache=SCALAR(0x874565c)') called at
/dev/null line 0
require 0 called at /dev/null line 0
12054 Apache::DBI PerlChildInitHandler
NOTICE: Unrecognized variable client_encoding
Database handle destroyed without explicit disconnect at
/usr/local/lib/perl5/5.6.0/site_perl/Apache/DBI.pm line 140.
require 0 called at
/usr/local/lib/perl5/5.6.0/site_perl/Apache/DBI.pm line 140
Apache::DBI::childinit('Apache=SCALAR(0x874565c)') called at
/dev/null line 0
require 0 called at /dev/null line 0
12055 Apache::DBI PerlChildInitHandler
12056 Apache::DBI PerlChildInitHandler
=
The NOTICE and etc messages are repeated several (12) times for the 6
Apache children and the 1 Apache root processes running (i.e. 2 NOTICE
messages per child).
The Apache::DBI.pm is version 0.87 and near line 140 I see:
sub childinit {
my $prefix = "$$ Apache::DBI";
print STDERR "$prefix PerlChildInitHandler \n" if $Apache::DBI::DEBUG 1;
#if (defined @ChildConnect) {
if (@ChildConnect) {
for my $aref (@ChildConnect) {
shift @$aref;
DBI-connect(@$aref);
$LastPingTime{@$aref[0]} = time;
}
}
1;
}
Note that I changed the "if(defined @ChildConnect" line because, "defined
on aggregates (hashes and arrays) is deprecated." Line 140 is:
$LastPingTime{@$aref[0]} = time;
RE: database access
At 08:04 AM 11/13/00 -0500, you wrote: I haven't followed this thread too closely, so forgive me if I'm not quite with it... you can already do this with Apache::DBI... change DBI-connect() to the (undocumented) 6 parameter call: Don't get me wrong here, "but", it would be nice if the undocumented somehow made it to the documented status. A related difficulty, is reading and understanding the documentation with my set of problems (problems are the reason why I read most documentation). I understand it is a difficult task, but someone has to do it (we will all benifit). Thanks.
Apache::AuthenDBI NOTICE message
It turns out that I am learning about DBI and Apache::DBI. So far, using Apache::AuthenDBI, I get an error_log message: NOTICE: Unrecognized variable client_encoding Every time the DBI ping is called. I have not been able to tack down the origin of this message. Any ideas or suggestions?
Apache::AutoIndex author
I want to enhance Apache::AutoIndex. I have tried to contact Philippe M. Chiasson, but have not had any success. Can someone please tell me how to contact Philippe or whom ever is maintaining the module. Thanks,
Apache_1.3.14 mod_layout 1.24_01 core dump
=
I am resending this email because I initially sent it out late Friday
evening and I only received further questions from one person. I was
hoping that someone could shed some light on why I am having problems
loading DSO modules when mod_perl is statically linked and loaded.
=
First, off, let me clarify, I am a Newbie in almost every way. . .
I have the following configuration Linux 2.2.17, Apache 1.3.14,
mod_perl-1.24_01 (static), Perl-5.6.0.
If I load Apache::Icon and mod_layout (static) I do not have a problem.
Also, I can not load mod_perl as a DSO and load Apache::Icon without core
dump. When I load mod_perl (static), Apache::Icon, and mod_layout 2.8 as
a DSO, Apache will also core dump and the following gdb bt output is produced:
=
#0 0x8089c9d in ap_push_array (arr=0x) at alloc.c:1029
#1 0x2ac788f4 in push_item () from
/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/auto/Apache/Icon/Icon.so
#2 0x2ac78a48 in add_alt () from
/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/auto/Apache/Icon/Icon.so
#3 0x2ac78ad8 in add_icon () from
/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/auto/Apache/Icon/Icon.so
#4 0x2ac78bdf in add_raw () from
/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/auto/Apache/Icon/Icon.so
#5 0x2ac78c55 in add_icon_raw () from
/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/auto/Apache/Icon/Icon.so
#6 0x808ec79 in invoke_cmd (cmd=0x2ac7a670, parms=0x7bd4,
mconfig=0x81c115c,args=0x7fffdb5e "(CMP,/icons/compressed.gif) x-compress
x-gzip") at http_config.c:808
#7 0x808f703 in ap_handle_command (parms=0x7bd4, config=0x81c0534,
l=0x7fffdb4c "AddIconByEncoding (CMP,/icons/compressed.gif) x-compress
x-gzip") at http_config.c:1028
#8 0x808f7a7 in ap_srm_command_loop (parms=0x7bd4, config=0x81c0534)
at http_config.c:1042
#9 0x808fe68 in ap_process_resource_config (s=0x81bf95c, fname=0x81c0f4c
"/usr/local/apache/conf/httpd.test", p=0x81bf934,ptemp=0x84121dc) at
http_config.c:1298
#10 0x8090762 in ap_read_config (p=0x81bf934, ptemp=0x84121dc,
confname=0x81af760 "conf/httpd.test") at http_config.c:1580
#11 0x809a6d1 in standalone_main (argc=4, argv=0x7ce4) at http_main.c:4640
#12 0x809af43 in main (argc=4, argv=0x7ce4) at http_main.c:5004
=
The tail of the strace was:
=
stat("/usr/local/apache/lib/perl/Apache/Icon.pmc", 0x7fffd7ac) = -1 ENOENT
(No such file or directory)
open("/usr/local/apache/lib/perl/Apache/Icon.pm", O_RDONLY) = -1 ENOENT (No
such file or directory)
stat("/usr/local/lib/perl5/5.6.0/i586-linux/Apache/Icon.pmc", 0x7fffd7ac) =
-1 ENOENT (No such file or directory)
open("/usr/local/lib/perl5/5.6.0/i586-linux/Apache/Icon.pm", O_RDONLY) = -1
ENOENT (No such file or directory)
stat("/usr/local/lib/perl5/5.6.0/Apache/Icon.pmc", 0x7fffd7ac) = -1 ENOENT
(No such file or directory)
open("/usr/local/lib/perl5/5.6.0/Apache/Icon.pm", O_RDONLY) = -1 ENOENT (No
such file or directory)
stat("/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/Apache/Icon.pmc",
0x7fffd7ac) = -1 ENOENT (No such file or directory)
open("/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/Apache/Icon.pm",
O_RDONLY) = 5
fstat(5, {st_mode=S_IFREG|0444, st_size=2333, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0x2aac6000
read(5, "package Apache::Icon;\n\nuse stric"..., 4096) = 2333
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
close(5)= 0
munmap(0x2aac6000, 4096)= 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++
=
I have a similar problem when I replace mod_layout with mod_auth_mda (DSO),
however I do not have a problem if I do not load mod_perl with the other
modules or if I do not load the other DSO modules (mod_layout or
mod_auth_mda).
No error_log output is generated.
Looks like some kind of munmap() problem.
I appreciate any help I can get. Thanks in advance.
Apache_1.3.14 mod_layout 1.24_01 core dump
First, off, let me clairfy, I am a Newbie in almost every way. . .
I have the following configuration Apache 1.3.14, mod_perl-1.24_01
(static), Perl-5.6.0, Linux 2.2.14.
If I load load Apache::Icon, and mod_layout 2.8 (DSO), Apache will core
dump and the following gdb bt output is produced:
#0 0x8089c9d in ap_push_array (arr=0x) at alloc.c:1029
#1 0x2ac788f4 in push_item () from
/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/auto/Apache/Icon/Icon.so
#2 0x2ac78a48 in add_alt () from
/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/auto/Apache/Icon/Icon.so
#3 0x2ac78ad8 in add_icon () from
/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/auto/Apache/Icon/Icon.so
#4 0x2ac78bdf in add_raw () from
/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/auto/Apache/Icon/Icon.so
#5 0x2ac78c55 in add_icon_raw () from
/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/auto/Apache/Icon/Icon.so
#6 0x808ec79 in invoke_cmd (cmd=0x2ac7a670, parms=0x7bd4,
mconfig=0x81c115c,args=0x7fffdb5e "(CMP,/icons/compressed.gif) x-compress
x-gzip") at http_config.c:808
#7 0x808f703 in ap_handle_command (parms=0x7bd4, config=0x81c0534,
l=0x7fffdb4c "AddIconByEncoding (CMP,/icons/compressed.gif) x-compress
x-gzip") at http_config.c:1028
#8 0x808f7a7 in ap_srm_command_loop (parms=0x7bd4, config=0x81c0534)
at http_config.c:1042
#9 0x808fe68 in ap_process_resource_config (s=0x81bf95c, fname=0x81c0f4c
"/usr/local/apache/conf/httpd.test", p=0x81bf934,ptemp=0x84121dc) at
http_config.c:1298
#10 0x8090762 in ap_read_config (p=0x81bf934, ptemp=0x84121dc,
confname=0x81af760 "conf/httpd.test") at http_config.c:1580
#11 0x809a6d1 in standalone_main (argc=4, argv=0x7ce4) at http_main.c:4640
#12 0x809af43 in main (argc=4, argv=0x7ce4) at http_main.c:5004
The tail of the strace was:
stat("/usr/local/apache/lib/perl/Apache/Icon.pmc", 0x7fffd7ac) = -1 ENOENT
(No such file or directory)
open("/usr/local/apache/lib/perl/Apache/Icon.pm", O_RDONLY) = -1 ENOENT (No
such file or directory)
stat("/usr/local/lib/perl5/5.6.0/i586-linux/Apache/Icon.pmc", 0x7fffd7ac) =
-1 ENOENT (No such file or directory)
open("/usr/local/lib/perl5/5.6.0/i586-linux/Apache/Icon.pm", O_RDONLY) = -1
ENOENT (No such file or directory)
stat("/usr/local/lib/perl5/5.6.0/Apache/Icon.pmc", 0x7fffd7ac) = -1 ENOENT
(No such file or directory)
open("/usr/local/lib/perl5/5.6.0/Apache/Icon.pm", O_RDONLY) = -1 ENOENT (No
such file or directory)
stat("/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/Apache/Icon.pmc",
0x7fffd7ac) = -1 ENOENT (No such file or directory)
open("/usr/local/lib/perl5/5.6.0/site_perl/i586-linux/Apache/Icon.pm",
O_RDONLY) = 5
fstat(5, {st_mode=S_IFREG|0444, st_size=2333, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0x2aac6000
read(5, "package Apache::Icon;\n\nuse stric"..., 4096) = 2333
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
close(5)= 0
munmap(0x2aac6000, 4096)= 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++
I have a similar problem when I replace mod_layout with mod_auth_mda (DSO),
however I do not have a problem if i do not load mod_perl with the other
modules or if I do not load the other DSO modules (mod_layout or
mod_auth_mda).
No error_log output is generated.
Looks like some kind of munmap() problem to me.
If will appreciate any help I can get. Thanks in advance.
Apache::Icon SIGHUP does not read directives
I have very strange problem. With a minimum configuration, (httpd -l) gives:
Compiled-in modules:
http_core.c
mod_env.c
mod_log_config.c
mod_mime.c
mod_negotiation.c
mod_dir.c
mod_alias.c
mod_access.c
mod_perl.c
suexec: enabled; valid wrapper /usr/local/apache/bin/suexec
and with the following build:
Apache/1.3.13-dev (Unix) mod_perl/1.24_01-dev configured
running on Linux 2.2.14 with Perl 5.6.0
The only mod_perl module loaded is Apache:Icon. (httpd.conf is attached to
the end of this message)
When I first launch httpd, Apache:Icon reads and processes all of the
autoindex directives.
However, if I SIGHUP the running httpd root process with,
'kill -HUP httpd-root-pid'
I get the following in the error_log file:
Syntax error on line 11 of /usr/local/apache/conf/httpd.dev:
Invalid command 'AddIconByType', perhaps mis-spelled or defined by a module
not included in the server configuration
Note that this is a valid directive and that it works fine on the initial
launch.
Can someone please point me in the right direction to solve this problem?
The following is my complete httpd.conf file.
#===
ClearModuleList
AddModule mod_log_config.c
AddModule mod_dir.c
AddModule mod_perl.c
#PerlFreshRestart On
DocumentRoot /home/httpd/test/htdocs
DirectoryIndex index.html index.htm index.shtml index.cgi Default.htm
default.htm
PerlModule Apache::Icon
#
#AddIconByEncoding (CMP,/icons/compressed.gif) x-compress x-gzip
AddIconByType (TXT,/icons/text.gif) text/*
AddIcon /icons/binary.gif .bin .exe
AddIcon /icons/tar.gif .tar
AddIcon /icons/compressed.gif .Z .z .tgz .gz .zip
AddIcon /icons/a.gif .ps .ai .eps
AddIcon /icons/text.gif .txt
DefaultIcon /icons/unknown.gif
#
ServerType standalone
Port 81
HostnameLookups off
User webuser
Group webgroup
ServerAdmin [EMAIL PROTECTED]
ServerRoot /usr/local/apache
ErrorLog logs/error_log.dev
LogLevel warn
LogFormat "%h %l %u %t \"%r\" %s %b \"%{Referer}i\" \"%{User-Agent}i\""
combined
LogFormat "%h %l %u %t \"%r\" %s %b" common
LogFormat "%{Referer}i - %U" referer
LogFormat "%{User-agent}i" agent
CustomLog logs/access_log.dev combined
PidFile logs/httpd.pid.dev
ScoreBoardFile logs/httpd.scoreboard.dev
LockFile logs/httpd.lock.dev
ServerName test.xorgate.com
UseCanonicalName on
Timeout 300
KeepAlive On
MaxKeepAliveRequests 100
KeepAliveTimeout 15
MinSpareServers 4
MaxSpareServers 10
StartServers 6
MaxClients 50
MaxRequestsPerChild 200
Missing src/include/alloc.h ?
I down loaded the CVS file apache-1.3_2825041201.tar.gz and attempted a make, that included: --add-module=/usr/src/mod_layout-2.3/mod_layout.c It expects src/include/alloc.h I noticed that this header file was included in the apache_1.3.12 distribution. Is the file missing?
Missing src/include/alloc.h ?
Please ignore my previous message, I let this message go out to the wrong mailing list. I down loaded the CVS file apache-1.3_2825041201.tar.gz and attempted a make, that included: --add-module=/usr/src/mod_layout-2.3/mod_layout.c It expects src/include/alloc.h I noticed that this header file was included in the apache_1.3.12 distribution. Is the file missing?
Re: Apache::Icon fix
Update:
Apache/1.3.12 (Unix) mod_perl/1.24_01-dev/Perl-5.6.0/linux-2.2.14.
(mod_perl linked statically.) with Icon.pm.
I hope I am not bugging anyone with my problem and I appreciate any help.
I have made some progress isolating the problem.
With a minimal config. (with only Apache::Icon loaded), it's directives are
not recognized after:
"kill -HUP httpd-pid"
The error_log.test output was:
Syntax error on line 17 of /usr/local/apache/conf/httpd.try:
Invalid command 'AddIconByEncoding', perhaps mis-spelled or defined by a
module not included in the server configuration
Hence the httpd (apache) processes terminate.
If I comment out all of the Icon directives, the SIGHUP works.
Of course, when I just launch httpd for the first time, every thing is fine.
Here is my minimal httpd.conf file:
LoadModule config_log_modulelibexec/mod_log_config.so
LoadModule access_modulelibexec/mod_access.so
LoadModule dir_module libexec/mod_dir.so
ClearModuleList
AddModule mod_log_config.c
AddModule mod_dir.c
AddModule mod_access.c
AddModule mod_so.c
AddModule mod_perl.c
DocumentRoot /home/httpd/test/htdocs
DirectoryIndex index.html index.htm index.shtml index.cgi Default.htm
default.htm
PerlModule Apache::Icon
#
AddIconByEncoding (CMP,/icons/compressed.gif) x-compress x-gzip
AddIconByType (TXT,/icons/text.gif) text/*
AddIcon /icons/binary.gif .bin .exe
AddIcon /icons/tar.gif .tar
AddIcon /icons/compressed.gif .Z .z .tgz .gz .zip
AddIcon /icons/a.gif .ps .ai .eps
AddIcon /icons/text.gif .txt
DefaultIcon /icons/unknown.gif
#
ServerType standalone
Port 81
HostnameLookups off
User webuser
Group webgroup
ServerAdmin [EMAIL PROTECTED]
ServerRoot /usr/local/apache
ErrorLog logs/error_log.test
LogLevel warn
LogFormat "%h %l %u %t \"%r\" %s %b \"%{Referer}i\" \"%{User-Agent}i\""
combined
LogFormat "%h %l %u %t \"%r\" %s %b" common
LogFormat "%{Referer}i - %U" referer
LogFormat "%{User-agent}i" agent
CustomLog logs/access_log.test combined
PidFile logs/httpd.pid.test
ScoreBoardFile logs/httpd.scoreboard.test
LockFile logs/httpd.lock.test
ServerName test.xorgate.com
UseCanonicalName on
Timeout 300
KeepAlive On
MaxKeepAliveRequests 100
KeepAliveTimeout 15
MinSpareServers 4
MaxSpareServers 10
StartServers 6
MaxClients 50
MaxRequestsPerChild 200
Re: Apache::Perfmon 0.011
At 11:26 AM 8/22/00 +0200, you wrote: On Tue, 22 Aug 2000, Lupe Christoph wrote: On Monday, 2000-08-21 at 21:06:54 -0400, Greg Stark wrote: Lupe Christoph [EMAIL PROTECTED] writes: Hmm. Apache::Benchmark sounds more like a benchmark driver to me. Apache::Instrumentation or so? Apache::Probe? Profile or even just Prof. That would be misleading. A profiler is something much more sophisticated and provides you a lot of information about the code that gets executed. More over the Apache::DProf and Apache::SmallProf are already in place. I thought about it a little more. What is does is find the (cpu|real)time taken to process a request. So what about Apache::RequestTimer or maybe Apache::Perf::RequestTimer, creating a new namespace for all performance-related modules? Both sound good time, especially the second :) How about a take off from vmstat (a great UNIX tool) and call it Apache::HTStat
Re: Apache::Icon fix
I have to admit that a lot of this stuff is a mistry. Recap: My configuration is: apache_1.3.12/mod_perl-1.24/Perl-5.6.0/linux-2.2.14. (I have mod_perl linked statically.) with Apache::AutoIndex.pm loaded. Doug, could you please explain what problem(s) your patch solved? I applied the patch below to Apache-Icon-0.02\Icon.xs, but all the Apache processes still die when I do "kill -HUP ###" on the Apache root PID. I noticed that when the order of the httpd.conf following two directives is (as you stated): PerlModule Apache::Icon PerlModule Apache::AutoIndex "apachectl start" produces the same error_log entry as my "kill -HUP ###", namely: Syntax error on line 156 of /usr/local/apache/conf/httpd.conf: Invalid command 'IndexOptions', perhaps mis-spelled or defined by a module not included in the server configuration However, with the order of the two directives reversed (as follows): PerlModule Apache::AutoIndex PerlModule Apache::Icon The launch is just fine. [The SIGHUP still fail however:[ Recap: My configuration is: apache_1.3.12/mod_perl-1.24/Perl-5.6.0/linux-2.2.14. (I have mod_perl linked statically.) I noticed that you are using development code with DSO. Perhaps I am missing some other patch that would alow it to work. Is the stability of DSO mod_perl improving? I remember reading that when httpd is started, it runs the init twice in order to verity that it can survive a SIGHUP. If this is the case, why would the SIGHUP fail? At 10:04 AM 8/21/00 -0700, Doug MacEachern [EMAIL PROTECTED] wrote: with the patch below, Apache::Icon/Apache::AutoIndex works fine for me. running apache_1.3.13-dev, mod_perl-1.24_01-dev, perl-current (5.7.0-dev) mod_perl built with EVERYTHING=1, USE_DSO=1 copy-n-pasted config from Apache::AutoIndex/README: PerlModule Apache::Icon PerlModule Apache::AutoIndex PerlTransHandler Apache::AutoIndex::transhandler PerlHandler Apache::AutoIndex kill -HUP works just fine too. --- Icon.xs 1999/01/18 19:21:46 1.1.1.1 +++ Icon.xs 2000/08/21 16:50:04 @@ -18,6 +18,12 @@ PROTOTYPES: DISABLE +void +END() + +CODE: +ap_remove_module(icon_module); + Apache::Icon new(class, r=default_r) char *class
Re: Producing an error page
At first, when I read the first message, I said to myself; (self) Why would anyone want an error page with all the Perl HTML STDOUT/STDOUT stuff mixed up together? However, after reading and thinking some more, . . . (just from a conceptional point of view and not from an implementation view point), , , what a powerful concept! If you (a developer) could regain the Perl command line style of output from the web server (itself) without having to piece together the STDOUT/STDERR during a debugging effort. Now, lets say that you had an Apache module that output a web page that looked just like the stuff you could see from the browsers "View/(Page)Source" with the STDERR output included. All one would have to do, is configure the DEBUG handler in the httpd.conf, which would in turn produce what appeared as HTML code with STDOUT included all within a HTML page. The STDERR output would be distinguished, with lets say: STDERR Opps got an error here or something. \STDERR I'm sure you would also want a submit (DEBUG) button which could generate the above, on-demand for form pages and such. (Kind of a DEBUG on demand option.) Is this what is being discussed here? (This is probably, not a mod_perl topic, but then again, it could be:-). At 12:04 PM 8/21/00 -0700, you wrote: The stuff that the server sends back comes from STDOUT if CGIs or within mod-perl, either $req-print or a regular print since it is tied. The error messages go to STDERR which apache redirects internalls so that the messages go to the error log. I don't know that it's possible and I'm certain it's not a good idea to try and circumvent that and send STDERR to the browser. And besides that, you'd still have to send content-type headers. I suppose you could write one perl program that executes your perl program of interest from within an eval block and capture the output that way. But isn't simply looking at the error log easier? -- Rob --On Monday, August 21, 2000 8:56 AM -0500 Jay Strauss [EMAIL PROTECTED] wrote: Hi, I'm asking this again, due to lack of response (but I can't believe no one out there knows how to do this). How do I produce an error page (in HTML), when I call the script from a browser, that looks just like the error screen I get when I run a script at the command line? That is, if I run the following script from the command line: --- #!/usr/bin/perl -w use strict; use diagnostics; ($first, $second) = @ARGV; exit; --- I'll get a whole bunch of messages telling me I "use strict" and I have variables that I didn't define with "my". But, if I call it from my browser, I just get back a "Internal Server Error" page. Instead I want all the diagnostics messages. Thanks Jay Jay Strauss [EMAIL PROTECTED] (h) 773.935.5326 (c) 312.617.0264
Re: Milage may vary comments
2) 'Do not': "PerlFreshRestart On" in httpd.conf.
Curiously enough, "PerlFreshRestart On" has always worked for me, even
on the production servers.
It's pretty damn useful too, giving servers a graceful restart everytime
you upload a module... it'd be a pain to have to fully restart the
server every time, that would mean like 10 seconds of downtime for the
whole site.
I should be more observant, because I think this problem ("PerlFreshRestart
On") may be related to my latest email posting "SIGHUP failure". Perhaps
if I solve it, the solution to this problem will follow. Hopefully, someone
will respond. It just seems like it's one think after another.
Perhaps as my skills improve, it will become easier to get more difficult!
SIGHUP failure
The problem is when I "apachectl restart" (SIGHUP), all Apache processes die (no core dump). However, if I "apachectl stop", (SIGTERM) followed by "apachectl start" everything is OK (no error_log error messages and every thing works fine). My configuration is: apache_1.3.12/mod_perl-1.24/Perl-5.6.0/linux-2.2.14. (I have mod_perl linked statically.:-) In the httpd.conf file, I have: # PerlModule Apache::AutoIndex PerlModule Apache::Icon IndexOptions FancyIndexing # In error_log, after a SIGHUP, the following is recorded: Syntax error on line 156 of /usr/local/apache/conf/httpd.conf: Invalid command 'IndexOptions', perhaps mis-spelled or defined by a module not included in the server configuration Any help or guidance will be appreciated
Milage may vary comments
I the past 6 weeks I have been working on various problems with Apache/mod_perl/AutoIndex.pm. There needs to be a "Your Mileage May Vary" (YMMV) reference page that is easily accessible. For Example,after scratching at the mouse pad, I found the following URLs with practical information: http://www.perlreference.com/mod_perl/guide/warnings.html#Evil_things_might _happen_when_us http://www.perlreference.com/mod_perl/guide/control.html There are no doubt many more URLs, but, my point is "perspective"; Practical information about, "what works and what doesn't" (WWWDuh:-). I want to point out a couple of observations during my trouble shooting: !) 'Do not' load mod_perl as a DSO. 2) 'Do not': "PerlFreshRestart On" in httpd.conf. 3) Always put any httpd.conf module directives after the module is loaded. (Be careful about the load sequence of Perl modules from within your "startup.pl" file.) 4) Trouble shooting is a bag of worms, which includes error_log, gdb, strace, PERL_TRACE (on UNIX of course). Setting up a testing environment on a different port (I used port 81), and then "divide and concur" (core and all:-). Which points out, that a lot of details, change for the various platforms. . . . Does anyone know what happened to or how to contact GOZER (Philippe M. Chiasson)? Or perhaps who maintains AutoIndex.pm? I have some questions and thoughts to pursue. . .
Re: Milage may vary comments
On Fri, 18 Aug 2000, Roger Espel Llima wrote: !) 'Do not' load mod_perl as a DSO. This should be a golden rule if you plan on HUPing your server with any regularity. 2) 'Do not': "PerlFreshRestart On" in httpd.conf. Curiously enough, "PerlFreshRestart On" has always worked for me, even on the production servers. It's pretty damn useful too, giving servers a graceful restart everytime you upload a module... it'd be a pain to have to fully restart the server every time, that would mean like 10 seconds of downtime for the whole site. It's even more useful if you want to cache a large quantity of data in ram with the root process and then periodically refresh that data. I agree! What was happening to me, was that with "PerlFreshRestart On", I would get a segmentation fault every time used AutoIndex.pm. I removed the directive and the problem went away. I have seen other references to memory leaks and such. If anyone has suggestions as to why the "PerlFreshRestart On" would aggervate segment faults, I would like to here them. -- Douglas Leonard [EMAIL PROTECTED]
Re: Apache 1.3.12/mod_perl 1.24/Perl 5.6.0 crash
I was able to get the Apache::AutoIndex module to load and run by
statically building mod_perl into Apache. Thanks.
Somebody should fix the flaky DSO. I did notice that the problems are
timing related. I would could get different results when I did a fast
"apachectl stop", "apachectl start" (the RedHat /etc/rc.d/init.d/httpd does
this) as opposed to a slow sequence.
I have Apache 1.3.12 using mod_perl 1.24 as a DSO, built with Perl 5.6.0
using
Apache::AutoIndex 0.08 which is running on Linux 2.2.14.
Everything works fine, until, I `apachectl stop`, then add the following
to the httpd.conf:
PerlModule Apache::AutoIndex
When I do `bin/httpd -X` Linux does a core dump.
In the process, I noticed the awkwardness of the httpd.conf
"LoadModule","AddModule" directive pair.
Apache's httpd knows which modules are statically linked (evidenced by
httpd -l). Modules should not be loaded ("LoadModule") if they are linked
statically, and conversely, DSO modules can not be added ("AddModule") if
they are not loaded (LoadModule).
Perhaps a better directive would just be "EnableModule", so that a DSO
module would be both loaded and enabled and a statically linked module
would not activated unless it was enabled ("EnableModule"). (I know this
sounds simple, and probably why it can't be implemented. :)
AutoIndex.pm reading autoindex directives
Well I finally got AutoIndex to load and run. But, . . . It does not read the autoindex directives from the httpd.conf file. I believe these are suppose to be processed by Apache::Icon. My httpd.conf is: # PerlModule Apache::AutoIndex PerlModule Apache::Icon SetHandler perl-script PerlHandler Apache::AutoIndex PerlTransHandler Apache::AutoIndex-transhandler # Note that in order to get the last line to work I had modify the transhandler method in order to make it a object-oriented (sub transhandler ($$)) method. Can someone please point me in the right direction so that AutoIndex.pm/Icon.pm read and process the httpd.conf autoindex directives? Thank you.
Re: Apache 1.3.12/mod_perl 1.24/Perl 5.6.0 crash
Let me try this again.
Recap:
I have Apache 1.3.12 using mod_perl 1.24 as a DSO, built with Perl 5.6.0
using Apache::AutoIndex 0.08 which is running on Linux 2.2.14.
Everything works fine, until, I `apachectl stop`, then add the following to
the httpd.conf:
PerlModule Apache::AutoIndex
When I do `bin/httpd -X` Linux does a core dump.
Has anyone tried this configuration with AutoIndex.pm with either
successful or with problems?
The following is the gdb bt listing:
=
(gdb) bt
Program received signal SIGSEGV, Segmentation fault.
0x8054b09 in ap_remove_module ()
(gdb) bt
#0 0x8054b09 in ap_remove_module ()
#1 0x8054bdf in ap_remove_loaded_module ()
#2 0x804f2cc in unload_module ()
#3 0x805126e in run_cleanups ()
#4 0x804f910 in ap_clear_pool ()
#5 0x8060683 in standalone_main ()
#6 0x8060f23 in main ()
#7 0x2ab304a5 in __libc_start_main () from /lib/libc.so.6
#8 0x815b in ?? ()
Cannot access memory at address 0xe853.
The following is tail end of a verbose strace ouput during a core dump.
=
open("/usr/local/apache/logs/error_log.test", O_WRONLY|O_APPEND|O_CREAT,
0666) = 4
fcntl(4, F_DUPFD, 15) = 15
close(4)= 0
fcntl(15, F_GETFL) = 0x401 (flags O_WRONLY|O_APPEND)
fstat(15, {st_mode=S_IFREG|0644, st_size=13212, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,0) =
0x2aac
_llseek(15, 0, [0], SEEK_CUR) = 0
dup2(15, 2) = 2
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
open("/usr/local/apache/conf/apache-mime.types", O_RDONLY) = 4
fstat(4, {st_mode=S_IFREG|0644, st_size=7374, ...}) = 0
fstat(4, {st_mode=S_IFREG|0644, st_size=7374, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,0) =
0x2aac8000
read(4, "# This is the default mime.types"..., 4096) = 4096
read(4, "m.EXT\napplication/vnd.osa.netdep"..., 4096) = 3278
brk(0x848) = 0x848
read(4, "", 4096) = 0
close(4)= 0
munmap(0x2aac8000, 4096)= 0
open("/usr/local/apache/logs/access_log.test", O_WRONLY|O_APPEND|O_CREAT,
0644)= 4
fcntl(4, F_DUPFD, 15) = 16
close(4)= 0
getpid()= 1404
time(NULL) = 965956118
close(16) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
munmap(0x2ad9d000, 14288) = 0
munmap(0x2ae0a000, 16308) = 0
munmap(0x2ae0e000, 13712) = 0
munmap(0x2ae12000, 22668) = 0
munmap(0x2ada1000, 9064)= 0
munmap(0x2ae18000, 46408) = 0
munmap(0x2ae24000, 29000) = 0
munmap(0x2ae2c000, 9736)= 0
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++
=
The following is the tail end of a mod_perl PERL_TRACE output:
=
5: _compile();
entering CGI::_compile
665:my($func) = $AUTOLOAD;
666:my($pack,$func_name);
667:{
668:local($1,$2); # this fixes an obscure variable
suicide problem.
668:local($1,$2); # this fixes an obscure variable
suicide problem.
669:$func=~/(.+)::([^:]+)$/;
670:($pack,$func_name) = ($1,$2);
671:$pack=~s/::SUPER$//;# fix another obscure problem
673:unless defined(${"$pack\:\:AUTOLOADED_ROUTINES"});
672:$pack = ${"$pack\:\:AutoloadClass"} ||
$CGI::DefaultClass
675:my($sub) = \%{"$pack\:\:SUBS"};
675:my($sub) = \%{"$pack\:\:SUBS"};
676:unless (%$sub) {
682: my($code) = $sub-{$func_name};
684: $code = "sub $AUTOLOAD { }" if (!$code and
$func_name eq 'DESTROY');
685: if (!$code) {
695: die "Undefined subroutine $AUTOLOAD\n" unless $code;
696: eval "package $pack; $code";
697: if ($@) {
702:CORE::delete($sub-{$func_name}); #free storage
703:return "$pack\:\:$func_name";
exited CGI::_compile
2: foreach (@_) {
exited CGI::AUTOLOAD for CGI::_compile_all
exited CGI::_setup_symbols
exited CGI::compile
46: 1;
3:
entering Config::DESTROY
exited Config::DESTROY
entering IO::Handle::DESTROY
exited IO::Handle::DESTROY
Carp::(/usr/local/lib/perl5/5.6.0/Carp.pm:97):
97: { local $@; require Carp::Heavy; } # XXX fix require to not
clear $@?
Carp::(/usr/local/lib/perl5/5.6.0/Carp.pm:97):
97: { local $@; require Carp::Heavy; } # XXX fix require to not
clear
Re: Apache 1.3.12/mod_perl 1.24/Perl 5.6.0 crash
At 03:59 PM 8/11/00 -0700, you wrote: Just curious, but how did you build 5.6.0? Running 5.6.0 and mdo_perl dso, I got the same thing. Have you tried compiling it statically? On 11 Aug 2000, at 16:52, George Sanderson wrote: Let me try this again. Recap: I have Apache 1.3.12 using mod_perl 1.24 as a DSO, built with Perl 5.6.0 using Apache::AutoIndex 0.08 which is running on Linux 2.2.14. Everything works fine, until, I `apachectl stop`, then add the following to the httpd.conf: PerlModule Apache::AutoIndex When I do `bin/httpd -X` Linux does a core dump. I really like the modular nature of DSO. One benifit from the DSO is that, I can just rebuild mod_perl independently of httpd. I normally build a production and a debug version of mod_perl then start up httpd on a differnt port to toubleshoot problems. I have tried the static build in the past but not resently. Do you know if there is more success one way or the other (static vr. DSO)? I've attached the Perl 5.6.0 config.sh file. For mod_perl I used a makepl_args.mod_perl file in the src root as follows: EVERYTHING=1 #PERL_DEBUG=1 #PERL_TRACE=1 USE_APXS=1 WITH_APXS=/usr/bin/apxs #DO_HTTPD=1 #NO_HTTPD=1 #PREP_HTTPD=1 APACHE_PREFIX=/usr/local/apache APACHE_SRC=/usr/src/apache_1.3.12/src #!/bin/sh # # This file was produced by running the Configure script. It holds all the # definitions figured out by Configure. Should you modify one of these values, # do not forget to propagate your changes by running "Configure -der". You may # instead choose to run each of the .SH files by yourself, or "Configure -S". # # Package name : perl5 # Source directory : . # Configuration time: Fri Aug 11 12:29:20 CDT 2000 # Configured by : george # Target system : linux estore.xorgate.com 2.2.14 #3 sat jan 15 08:41:41 cst 2000 i586 unknown Author='' Date='$Date' Header='' Id='$Id' Locker='' Log='$Log' Mcc='Mcc' RCSfile='$RCSfile' Revision='$Revision' Source='' State='' _a='.a' _exe='' _o='.o' afs='false' alignbytes='4' ansi2knr='' aphostname='/bin/hostname' api_revision='5' api_subversion='0' api_version='5' api_versionstring='5.005' ar='ar' archlib='/usr/local/lib/perl5/5.6.0/i586-linux' archlibexp='/usr/local/lib/perl5/5.6.0/i586-linux' archname64='' archname='i586-linux' archobjs='' awk='awk' baserev='5.0' bash='' bin='/usr/bin' bincompat5005='undef' binexp='/usr/bin' bison='' byacc='byacc' byteorder='1234' c='' castflags='0' cat='cat' cc='gcc' cccdlflags='-fpic' ccdlflags='-rdynamic -Wl,-rpath,/usr/local/lib/perl5/5.6.0/i586-linux/CORE' ccflags='-fno-strict-aliasing -I/usr/local/include' ccsymbols='__GNUC_MINOR__=91 __i386=1 __i386__=1 __i686=1 __i686__=1 __linux=1 __linux__=1 __pentiumpro=1 __pentiumpro__=1 __unix=1 __unix__=1 cpu=i386 machine=i386 pentiumpro=1 system=posix system=unix' cf_by='george' cf_email='[EMAIL PROTECTED]' cf_time='Fri Aug 11 12:29:20 CDT 2000' charsize='1' chgrp='' chmod='' chown='' clocktype='clock_t' comm='comm' compress='' contains='grep' cp='cp' cpio='' cpp='cpp' cpp_stuff='42' cppccsymbols='__ELF__=1 __GNUC__=2 i386=1 i686=1 linux=1 unix=1' cppflags='-fno-strict-aliasing -I/usr/local/include' cpplast='-' cppminus='-' cpprun='gcc -E' cppstdin='gcc -E' cppsymbols='__GNUC_MINOR__=91 _POSIX_C_SOURCE=199506 _POSIX_SOURCE=1 __STDC__=1 __i386=1 __i386__=1 __i686=1 __i686__=1 __linux=1 __linux__=1 __unix=1 __unix__=1' crosscompile='undef' cryptlib='' csh='csh' d_Gconvert='gcvt((x),(n),(b))' d_PRIEldbl='define' d_PRIFldbl='define' d_PRIGldbl='define' d_PRIX64='define' d_PRId64='define' d_PRIeldbl='define' d_PRIfldbl='define' d_PRIgldbl='define' d_PRIi64='define' d_PRIo64='define' d_PRIu64='define' d_PRIx64='define' d_access='define' d_accessx='undef' d_alarm='define' d_archlib='define' d_atolf='undef' d_atoll='define' d_attribut='define' d_bcmp='define' d_bcopy='define' d_bincompat5005='undef' d_bsd='undef' d_bsdgetpgrp='undef' d_bsdsetpgrp='undef' d_bzero='define' d_casti32='undef' d_castneg='define' d_charvspr='undef' d_chown='define' d_chroot='define' d_chsize='undef' d_closedir='define' d_const='define' d_crypt='define' d_csh='define' d_cuserid='define' d_dbl_dig='define' d_difftime='define' d_dirnamlen='undef' d_dlerror='define' d_dlopen='define' d_dlsymun='undef' d_dosuid='undef' d_drand48proto='define' d_dup2='define' d_eaccess='undef' d_endgrent='define' d_endhent='define' d_endnent='define' d_endpent='define' d_endpwent='define' d_endsent='define' d_endspent='define' d_eofnblk='define' d_eunice='undef' d_fchmod='define' d_fchown='define' d_fcntl='define' d_fd_macros='define' d_fd_set='define' d_fds_bits='undef' d_fgetpos='define' d_flexfnam='define' d_flock='define' d_fork='define' d_fpathconf='define' d_fpos64_t='undef' d_fs_data_s='undef' d_fseeko='define' d_fsetpos='define' d_fstatfs='define' d_fstatvfs='define' d_ftello='define' d_ftime='undef' d_getcwd='define' d_getfsstat='undef' d_getgrent='define' d_getgrps='define' d_gethbyaddr='define' d_gethbyname='define' d_gethent
Re: httpd.conf directive PerlHandler Apache::Hello-handler errors
Ah I C, said the Perl man! Thank you for the clarification. At 09:55 AM 8/10/00 -0500, you wrote: [EMAIL PROTECTED] (G.W. Haywood) wrote: Hi there, On Wed, 9 Aug 2000, George Sanderson wrote: PerlModule Apache::Hello Location /hello/world SetHandler perl-script PerlHandler Apache::Hello-handler /Location # Results in the following error_log output: [Sun Aug 6 21:48:02 2000] [error] Undefined subroutine Apache::Hello-handler::handler called at PerlHandler subroutine `Apache::Hello-handler' line 1. Page 70, Eagle Book. You mean PerlHandler Apache::Hello::handler To clarify - some handlers can be called using object-oriented techniques, and some can't. The switch for this behavior is that the handler is prototyped with ($$). Apache::Hello isn't, so you can't call its handler as a method, only as a function. Eagle book, p. 190.
Re: Apache 1.3.12/mod_perl 1.24/Perl 5.6.0 crash
At 10:49 PM 8/10/00 +0200, you wrote:
On Thu, Aug 10, 2000 at 07:39:24AM -0500, George Sanderson wrote:
At 12:22 PM 8/10/00 +0200, you wrote:
with almost the same conf as your, I had a core dump even with httpd
-t. I did a strace, and I noticed this is due to an inexistant module :
Tie::*hash or something similar. A quick search on CPAN give me the
package to install. After that, no more segfault ...
I did a strace and the last entries were:
munmap(0x2ae2c000, 9736)= 0
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++
Could this be a lack of memory?
dunno...
if you haven't answer on the list, perhaps a more verbose strace trace
could help...
Good luck,
Thanks (I'd rather be lucky than good!-)
Recap:
I have Apache 1.3.12 using mod_perl 1.24 as a DSO, built with Perl 5.6.0
using Apache::AutoIndex 0.08 which is running on Linux 2.2.14.
Everything works fine, until, I `apachectl stop`, then add the following to
the httpd.conf:
PerlModule Apache::AutoIndex
When I do `bin/httpd -X` Linux does a core dump.
I added more memory the computer but the program still crashes at the same
point.
Here's more of a more verbose strace ouput during the core dump.
Any clues?
open("/usr/local/apache/logs/error_log.test", O_WRONLY|O_APPEND|O_CREAT,
0666) = 4
fcntl(4, F_DUPFD, 15) = 15
close(4)= 0
fcntl(15, F_GETFL) = 0x401 (flags O_WRONLY|O_APPEND)
fstat(15, {st_mode=S_IFREG|0644, st_size=13212, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0x2aac
_llseek(15, 0, [0], SEEK_CUR) = 0
dup2(15, 2) = 2
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
open("/usr/local/apache/conf/apache-mime.types", O_RDONLY) = 4
fstat(4, {st_mode=S_IFREG|0644, st_size=7374, ...}) = 0
fstat(4, {st_mode=S_IFREG|0644, st_size=7374, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0x2aac8000
read(4, "# This is the default mime.types"..., 4096) = 4096
read(4, "m.EXT\napplication/vnd.osa.netdep"..., 4096) = 3278
brk(0x848) = 0x848
read(4, "", 4096) = 0
close(4)= 0
munmap(0x2aac8000, 4096)= 0
open("/usr/local/apache/logs/access_log.test", O_WRONLY|O_APPEND|O_CREAT,
0644) = 4
fcntl(4, F_DUPFD, 15) = 16
close(4)= 0
getpid()= 1404
time(NULL) = 965956118
close(16) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
munmap(0x2ad9d000, 14288) = 0
munmap(0x2ae0a000, 16308) = 0
munmap(0x2ae0e000, 13712) = 0
munmap(0x2ae12000, 22668) = 0
munmap(0x2ada1000, 9064)= 0
munmap(0x2ae18000, 46408) = 0
munmap(0x2ae24000, 29000) = 0
munmap(0x2ae2c000, 9736)= 0
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++
[
I also did a PERL_TRACE of mod_perl. The end of which was:
5: _compile();
entering CGI::_compile
665:my($func) = $AUTOLOAD;
666:my($pack,$func_name);
667:{
668:local($1,$2); # this fixes an obscure variable
suicide problem.
668:local($1,$2); # this fixes an obscure variable
suicide problem.
669:$func=~/(.+)::([^:]+)$/;
670:($pack,$func_name) = ($1,$2);
671:$pack=~s/::SUPER$//;# fix another obscure problem
673:unless defined(${"$pack\:\:AUTOLOADED_ROUTINES"});
672:$pack = ${"$pack\:\:AutoloadClass"} ||
$CGI::DefaultClass
675:my($sub) = \%{"$pack\:\:SUBS"};
675:my($sub) = \%{"$pack\:\:SUBS"};
676:unless (%$sub) {
682: my($code) = $sub-{$func_name};
684: $code = "sub $AUTOLOAD { }" if (!$code and
$func_name eq 'DESTROY');
685: if (!$code) {
695: die "Undefined subroutine $AUTOLOAD\n" unless $code;
696: eval "package $pack; $code";
697: if ($@) {
702:CORE::delete($sub-{$func_name}); #free storage
703:return "$pack\:\:$func_name";
exited CGI::_compile
2: foreach (@_) {
exited CGI::AUTOLOAD for CGI::_compile_all
exited CGI::_setup_symbols
exited CGI::compile
46: 1;
3:
entering Config::DESTROY
exited Config::DESTROY
entering IO::Handle::DESTROY
Apache 1.3.12/mod_perl 1.24/Perl 5.6.0 crash
I have Apache 1.3.12 using mod_perl 1.24 as a DSO, built with Perl 5.6.0 using Apache::AutoIndex 0.08 which is running on Linux 2.2.14. Everything works fine, until, I `apachectl stop`, then add the following to the httpd.conf: PerlModule Apache::AutoIndex When I do `bin/httpd -X` Linux does a core dump. Since I compiled in DEBUG the 'gdb bin/httpd core` reports a segmentation fault with: Program received signal SIGSEGV, Segmentation fault. 0x8054b09 in ap_remove_module () (gdb) bt Program received signal SIGSEGV, Segmentation fault. 0x8054b09 in ap_remove_module () (gdb) bt #0 0x8054b09 in ap_remove_module () #1 0x8054bdf in ap_remove_loaded_module () #2 0x804f2cc in unload_module () #3 0x805126e in run_cleanups () #4 0x804f910 in ap_clear_pool () #5 0x8060683 in standalone_main () #6 0x8060f23 in main () #7 0x2ab304a5 in __libc_start_main () from /lib/libc.so.6 #8 0x815b in ?? () Cannot access memory at address 0xe853. == I don't understand why Apache/mod_perl has to remove a module when it should be installing it. Maybe its just a confusing naming convention or something? Note: that I believe I was careful to clean out old versions before the makes. I did have to 'not use' "large file support" for the Perl 5.6.0 build in order to get mod_perl to launch. Please help! ... and thankyou for any input. (You all do great work!)
