protocol/5411: probably missing test in If-Modified-Since code
Number: 5411 Category: protocol Synopsis: probably missing test in If-Modified-Since code Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Fri Dec 3 03:40:00 PST 1999 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.9 Environment: Linux RH 6.1 .. doesn't matter Description: Suppose you have apache on IP and only one global web (with It worked..). Someone looked it via SQUID proxy which cached the / dir. Then we created virtual web on the same IP. The same person looked again but it have seen the same old page. When I traced it using sniffit I found that SQUID sends request with If-Modified-Since header with date 11/8 and length=1945. New index.html was older but has DIFFERENT LENGTH. It seems the Apache tests ONLY date and when file-date is older it doesn't check file length too. It's not too serious bug, but probably it can make some headaches .. How-To-Repeat: see above Fix: Yep. Test for file length difference before sending 304 response .. Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, you need] [to include [EMAIL PROTECTED] in the Cc line and make sure the] [subject line starts with the report component and number, with ] [or without any 'Re:' prefixes (such as general/1098: or ] [Re: general/1098:). If the subject doesn't match this ] [pattern, your message will be misfiled and ignored. The ] [apbugs address is not added to the Cc line of messages from ] [the database automatically because of the potential for mail ] [loops. If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request from a ] [developer. Reply only with text; DO NOT SEND ATTACHMENTS! ]
mod_proxy/5315: ProxyRequests Off (no forbidden-message)
Number: 5315 Category: mod_proxy Synopsis: ProxyRequests Off (no forbidden-message) Confidential: no Severity: serious Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Mon Nov 15 06:20:01 PST 1999 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.9 Environment: AIX 4.1.4, gcc2.95 Description: -- Apache, Version 1.2.b8 If Apache is used as a proxy-server: ProxyRequests Off results in forbidden, you do not have permission ... Apache as a web-server is working fine (ProxyPass-Statements) -- Apache, Version 1.3.9 ProxyRequests Off: Instead off the forbidden-message, all requests are sent to my default-site. My question is: How can I configure Apache to get the forbidden-message for the reasen that Apache is used as a proxy-Server! Thank´s a lot! Martin Vodicka How-To-Repeat: Fix: Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, you need] [to include [EMAIL PROTECTED] in the Cc line and make sure the] [subject line starts with the report component and number, with ] [or without any 'Re:' prefixes (such as general/1098: or ] [Re: general/1098:). If the subject doesn't match this ] [pattern, your message will be misfiled and ignored. The ] [apbugs address is not added to the Cc line of messages from ] [the database automatically because of the potential for mail ] [loops. If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request from a ] [developer. Reply only with text; DO NOT SEND ATTACHMENTS! ]
mod_proxy/5138: Can the proxy and http running on the same port ?
Number: 5138 Category: mod_proxy Synopsis: Can the proxy and http running on the same port ? Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: support Submitter-Id: apache Arrival-Date: Tue Oct 12 20:10:30 PDT 1999 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.6 Environment: UNIX NCR 3.02 Description: We configured (not intentionaly) the http server and proxy module over the same port. Is this configuration possible ? Who serve each request first ? How-To-Repeat: Fix: Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, you need] [to include [EMAIL PROTECTED] in the Cc line and make sure the] [subject line starts with the report component and number, with ] [or without any 'Re:' prefixes (such as general/1098: or ] [Re: general/1098:). If the subject doesn't match this ] [pattern, your message will be misfiled and ignored. The ] [apbugs address is not added to the Cc line of messages from ] [the database automatically because of the potential for mail ] [loops. If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request from a ] [developer. Reply only with text; DO NOT SEND ATTACHMENTS! ]
mod_rewrite/5073: mod_rewrite drop the QUERY_STRING using the [P] flag on per-dir context
Number: 5073 Category: mod_rewrite Synopsis: mod_rewrite drop the QUERY_STRING using the [P] flag on per-dir context Confidential: no Severity: serious Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Tue Sep 28 09:20:00 PDT 1999 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.9 Environment: SunOS obelix 5.5.1 Generic_103640-27 sun4m sparc SUNW,SPARCstation-20 egcs-2.9.5 Description: The mod_rewrite in the apache 1.3.9 strip off the QUERY_STRING if the force proxy flag is used in the per-directory context. I think the same bug is reported earlier but the author forgot the per-directory context. If [P] is used in global context all is fine. How-To-Repeat: This configuration doesn't work with 1.3.9. With apache 1.3.6 it worked well! Location /hack RewriteEngine On RewriteRule ^.*/htdocs/hack(.*)$ http://hugo.sat1.de:88/hack$1 [P,L] /Location Fix: I comment the line 1377 in mod_rewrite.c but I didn't check the side effects! After this change the QUERY_STRING is appended 1376: if (r-args != NULL) { 1377:/* Martin Zeh changed 28.09.1999 r-uri == r-unparsed_uri) { */ 1378:/* see proxy_http:proxy_http_canon() */ 1379:r-filename = ap_pstrcat(r-pool, r-filename, 1380:?, r-args, NULL); 1381: } Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, you need] [to include [EMAIL PROTECTED] in the Cc line and make sure the] [subject line starts with the report component and number, with ] [or without any 'Re:' prefixes (such as general/1098: or ] [Re: general/1098:). If the subject doesn't match this ] [pattern, your message will be misfiled and ignored. The ] [apbugs address is not added to the Cc line of messages from ] [the database automatically because of the potential for mail ] [loops. If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request from a ] [developer. Reply only with text; DO NOT SEND ATTACHMENTS! ]
os-windows/4872: Cannot run Apache as a service using a particular configuration
Number: 4872 Category: os-windows Synopsis: Cannot run Apache as a service using a particular configuration Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Tue Aug 17 16:10:01 PDT 1999 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.6 Environment: Windows NT 4.0 SP 3, Binary Distribution Description: Apache does not properly install as a service when using a particular configuration. For example, the command apache -i -f conf\0\httpd.conf. installs Apache as a service, but the service starts with the default configuration conf\http.conf rather than the configuration conf\0\httpd.conf specified during installation. How-To-Repeat: Fix: Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, you need] [to include [EMAIL PROTECTED] in the Cc line and make sure the] [subject line starts with the report component and number, with ] [or without any 'Re:' prefixes (such as general/1098: or ] [Re: general/1098:). If the subject doesn't match this ] [pattern, your message will be misfiled and ignored. The ] [apbugs address is not added to the Cc line of messages from ] [the database automatically because of the potential for mail ] [loops. If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request from a ] [developer. Reply only with text; DO NOT SEND ATTACHMENTS! ]
Re: config/4248: ServerSignature directive only effective in main server config.
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: ServerSignature directive only effective in main server config. State-Changed-From-To: open-closed State-Changed-By: martin State-Changed-When: Wed Apr 28 06:00:16 PDT 1999 State-Changed-Why: Thanks for your patch. I have committed it for apache 1.3.7
Re: mod_proxy/4087: Fix for ftp proxy with spaces in filenames
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: Fix for ftp proxy with spaces in filenames State-Changed-From-To: open-feedback State-Changed-By: martin State-Changed-When: Wed Apr 28 06:12:20 PDT 1999 State-Changed-Why: The patch you mentioned isn't at the location ftp://ftp.characterlink.net/proxy/ any longer. Please submit the patch in a reply or send it to [EMAIL PROTECTED]. Does Squid's license allow inclusion onto apache, or did you rewrite the code yourself based on information in squid's code? Responsible-Changed-From-To: apache-martin Responsible-Changed-By: martin Responsible-Changed-When: Wed Apr 28 06:12:20 PDT 1999 Responsible-Changed-Why: I will check the code; I knew the ftp proxy parser had a very weak error tolerance.
Re: mod_proxy/3859: cache problems
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: cache problems State-Changed-From-To: open-feedback State-Changed-By: martin State-Changed-When: Wed Apr 28 06:25:00 PDT 1999 State-Changed-Why: It seems that two different errors are mixed in this PR. a) [error] (17)File exists: proxy: error linking cache file This error can occur under normal situations when two requests to the same URL happen in parallel (two proxy children). A better concept of shared memory interprocess request table between proxy children might resolve this (not anytime soon...). b) [error] (16)Device busy: Hmmm. What can apache do about that? It seems to be an error in your OS. All the proxy does is call unlink(name). If Solaris can't handle that, then :-( Also, I don't see how apache can be involved in the inability to handle large disks, other than the fact that one GC must have terminated before the next one starts. Did you observe any core dumps? Beware, there might be no log message about it (GC is detached!). Try to trigger a GC by touching 0101010170 cache/.time and watch the background process. Recently in 1.3.7-dev, we had a core dump candidate. Try to report any findings. Thank you.
Re: mod_proxy/3776: Proxy Authentication appears to be broken
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: Proxy Authentication appears to be broken State-Changed-From-To: open-analyzed State-Changed-By: martin State-Changed-When: Wed Apr 28 08:16:48 PDT 1999 State-Changed-Why: The line which ignores Proxy-Auth at this place is intentional. We don't want to pass the auth information to any server further down the chain (the assumption here is that it was our own proxy authentication). The error is the fact that we don't check if *WE* requested and used the proxy authentication, or if it was another server further down. We should only strip out the proxy auth info if we requested and used it ourselves. This test is indeed currently missing, and therefore we always strip the Proxy-Auth header to be on the safe side. It's somewhere on my TODO list...
Re: mod_proxy/2577: Return error 403 instead of 500 due to ProxyBlock directive
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: Return error 403 instead of 500 due to ProxyBlock directive State-Changed-From-To: open-closed State-Changed-By: martin State-Changed-When: Wed Apr 28 08:30:29 PDT 1999 State-Changed-Why: Sorry for the late reply. This has been solved in a more flexible way in recent apache releases already. (e.g., 1.3.6). Anyway, thanks for your bug report!
Re: mod_proxy/3859: cache problems
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: cache problems Comment-Added-By: martin Comment-Added-When: Wed Apr 28 08:38:56 PDT 1999 Comment-Added: Oops! I should have read PR#3233 first. Yes, large disks will make problems when you configure a CacheSize with more than 2 gig. I'll fix it RSN.
mod_rewrite/4276: Rewriting with {prg:/my/rewriteprog} fails if URI contains %0A
Number: 4276 Category: mod_rewrite Synopsis: Rewriting with {prg:/my/rewriteprog} fails if URI contains %0A Confidential: no Severity: serious Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Tue Apr 20 06:10:01 PDT 1999 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.6 Environment: Linux infong1 2.0.36 #1 Wed Mar 24 21:40:07 MET 1999 i686 unknown gcc version egcs-2.91.60 19981201 (egcs-1.1.1 release) Description: passing a URL to the rewriteprog wich contains %0A in fact passes 2 lines to the rewriteprog. Because of the recommeded structure of the rewriteprog (one line in, one line out) the apache gets back two lines wich is very confusing for it. How-To-Repeat: Install a {prg:/my/rewriteprog} and pass a url containing %0A to it. Fix: A workaround is to prepend a magic string to the url. If the rewriteprg cannot find the magic string at the begin of a inputline it should produce no output Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
documentation/4101: http.conf.default protects .htaccess - better protect .htaccess*
Number: 4101 Category: documentation Synopsis: http.conf.default protects .htaccess - better protect .htaccess* Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: change-request Submitter-Id: apache Arrival-Date: Mon Mar 22 11:20:01 PST 1999 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.4 Environment: not relevant here Description: http.conf.default contains Files .htaccess Order allow,deny Deny from all /Files It would be better to use Files .htaccess* because editors like emacs leave backup files with suffixes. How-To-Repeat: emacs .htaccess Fix: change Files .htaccess to Files .htaccess* (Except that would be costly in terms of CPU time, I don't that's really an issue here...) Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
mod_status/3936: 'extendedstatus on' doesn't work
Number: 3936 Category: mod_status Synopsis: 'extendedstatus on' doesn't work Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Mon Feb 22 03:30:01 PST 1999 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.4 Environment: win nt 4.0 Description: the 'extendedstatus on' parameter doesn't work with the default httpd.conf because it is placed before the 'loadmodule' line. How-To-Repeat: Fix: move the 'extendedstatus on', so that the module is loaded Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
Re: mod_proxy/3455: 500 Proxy Error raised too often
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: 500 Proxy Error raised too often State-Changed-From-To: open-closed State-Changed-By: martin State-Changed-When: Tue Feb 16 06:20:16 PST 1999 State-Changed-Why: Based in part on your patch, two changes in http_protocol.c and proxy_util.c address this problem. First, the proxy doesn't always return HTTP_INTERNAL_SERVER_ERROR any longer, and second, the proxy sets an expose flag so that the error printing routine displays the error-notes text. Thanks for your patch/suggestion. Release-Changed-From-To: 1.3.3-1.3.4 Release-Changed-By: martin Release-Changed-When: Tue Feb 16 06:20:16 PST 1999
Re: general/3573: [PATCH] a wrote a module to process macro which needs a fix in main/http_config.c
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: [PATCH] a wrote a module to process macro which needs a fix in main/http_config.c State-Changed-From-To: open-closed State-Changed-By: martin State-Changed-When: Mon Dec 28 04:32:30 PST 1998 State-Changed-Why: Your proposal was taken into http_config.c as of rev. 1.138; I see no ill side effects, and I think modules can take advantage of the added flexibility of modifying the file name.
Re: mod_proxy/3530: Apache proxy confused by superfluous ':' in host part of URL
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: Apache proxy confused by superfluous ':' in host part of URL State-Changed-From-To: open-closed State-Changed-By: martin State-Changed-When: Mon Dec 28 05:00:12 PST 1998 State-Changed-Why: You are right. Thanks for the patch. It was added as Rev. 1.73 for apache-1.3.4 Category-Changed-From-To: general-mod_proxy Category-Changed-By: martin Category-Changed-When: Mon Dec 28 05:00:12 PST 1998
Re: os-windows/3416: UserDir for directory on a different Server not working
The following reply was made to PR os-windows/3416; it has been noted by GNATS. From: [EMAIL PROTECTED] (Martin Leyrer) To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: os-windows/3416: UserDir for directory on a different Server not working Date: Thu, 19 Nov 1998 23:34:51 +0100 Thank you for your fast response! I am looking forward to the next release. Cheers Martin 'm3' Leyrer =8-]== mailto:[EMAIL PROTECTED] - The world is filled with willing people; some willing to work, the rest willing to let them. --- Robert Frost Synopsis: UserDir for directory on a different Server not working State-Changed-From-To: open-closed State-Changed-By: [EMAIL PROTECTED] State-Changed-When: Wed Nov 18 09:54:55 PST 1998 State-Changed-Why: Yep, this was a problem in the os_canonical_filename routine in 1.3.2. The routine would add the incorrect slash. This routine has been fixed in CVS, and should be available in the next release of apache for Windows. Thanks for using Apache!
mod_rewrite/3320: RewriteEngine On directive in Directory
Number: 3320 Category: mod_rewrite Synopsis: RewriteEngine On directive in Directory Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Fri Oct 30 17:40:01 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.3. Environment: RedHat Linux 5.1 w/ 2.0.35 Kernel on a dual PII gcc 2.7.2.3 Description: I've upgraded my apache from 1.2.6 to 1.3.3 and my RewriteMap mapname prg:/file... stoped to work. Apache stopped to care about it. The rewrite module was configured like this: Virtualhost ... RewriteLog RewriteLogLevel 1 RewriteMap mapname prg:/file... /Virtualhost Directory ... RewriteEngine on RewriteRule ... /Directory I've discovered that Apache is not even initializing my program, until I've moved the RewriteEngine On directive to the VirtualHost definition context. I truly don't know if this is a bug or the rewrite module is intended to behave like that. So far I have not found a word about this behaviour. With the previous version of Apache I've had (1.2.6) it was working fine. How-To-Repeat: Fix: Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
general/3308: binary distribution available - how to submit to binary archive?
Number: 3308 Category: general Synopsis: binary distribution available - how to submit to binary archive? Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: support Submitter-Id: apache Arrival-Date: Wed Oct 28 12:50:00 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.3 Environment: uname -a: HP-UX urmel 7.05 B 9000/350 urmel compiler: gcc version 2.7.2.3 Description: NO ERROR or anything. I am building the Apache binary for almost any release and can make this available if interest is sufficient. Supported platforms are HP9000 Series 300 and 400 (all M68K based machines from HP). Build environment is as described above but the binaries should run fine on HP-UX 7.x, 8.x and 9.x for M68K. If you are interested, let me know how to submit the binary packages. (BTW: I have read the chapter about binary distributions on this Web- site, but it is not clear to me, how the files are actually transferred over to you) How-To-Repeat: N/A Fix: N/A Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
general/3309: binary distribution available - how to submit to binary archive?
Number: 3309 Category: general Synopsis: binary distribution available - how to submit to binary archive? Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: support Submitter-Id: apache Arrival-Date: Wed Oct 28 12:50:01 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.3 Environment: uname -a: HP-UX urmel 7.05 B 9000/350 urmel compiler: gcc version 2.7.2.3 Description: NO ERROR or anything. I am building the Apache binary for almost any release and can make this available if interest is sufficient. Supported platforms are HP9000 Series 300 and 400 (all M68K based machines from HP). Build environment is as described above but the binaries should run fine on HP-UX 7.x, 8.x and 9.x for M68K. If you are interested, let me know how to submit the binary packages. (BTW: I have read the chapter about binary distributions on this Web- site, but it is not clear to me, how the files are actually transferred over to you) How-To-Repeat: N/A Fix: N/A Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
Re: general/3309: binary distribution available - how to submit to binary archive?
The following reply was made to PR general/3309; it has been noted by GNATS. From: Martin Bayreuther [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: Subject: Re: general/3309: binary distribution available - how to submit to binary archive? Date: Wed, 28 Oct 1998 21:52:38 +0100 (MET) From: Martin Bayreuther Subject: Autoreply Priority: Normal - Autoreply - Sie haben den E-Mail Briefkasten von Martin Bayreuther Hewlett Packard GmbH ich bin bis auf weiteres nicht im Buero. Bei dringenden Faellen wenden Sie sich bitte an Sibylle Schock oder Bert Anderson in Andover. Gelegentlich werde ich von zu Hause aus meine E-Mail anschauen. Vielen Dank fuer Ihre Nachricht Martin Bayreuther . You have contacted the e-mail box of Martin Bayreuther Hewlett Packard GmbH I will be out of the office for some time. In urgent cases please contact Sibylle Schock or Bert Anderson in Andover. Occasionally I will check my e-mail account from home. Thank you for your message Martin Bayreuther
Re: general/3308: binary distribution available - how to submit to binary archive?
The following reply was made to PR general/3308; it has been noted by GNATS. From: Martin Bayreuther [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: Subject: Re: general/3308: binary distribution available - how to submit to binary archive? Date: Wed, 28 Oct 1998 21:52:37 +0100 (MET) From: Martin Bayreuther Subject: Autoreply Priority: Normal - Autoreply - Sie haben den E-Mail Briefkasten von Martin Bayreuther Hewlett Packard GmbH ich bin bis auf weiteres nicht im Buero. Bei dringenden Faellen wenden Sie sich bitte an Sibylle Schock oder Bert Anderson in Andover. Gelegentlich werde ich von zu Hause aus meine E-Mail anschauen. Vielen Dank fuer Ihre Nachricht Martin Bayreuther . You have contacted the e-mail box of Martin Bayreuther Hewlett Packard GmbH I will be out of the office for some time. In urgent cases please contact Sibylle Schock or Bert Anderson in Andover. Occasionally I will check my e-mail account from home. Thank you for your message Martin Bayreuther
Re: general/3190: httpd processes hang and never free up their slot, until MaxClients reached
The following reply was made to PR general/3190; it has been noted by GNATS. From: Kevin Martin [EMAIL PROTECTED] To: [EMAIL PROTECTED] (Marc Slemko) Cc: [EMAIL PROTECTED], [EMAIL PROTECTED] Subject: Re: general/3190: httpd processes hang and never free up their slot, until MaxClients reached Date: Wed, 14 Oct 1998 21:27:27 -0400 (EDT) No problem, I understand, thank you for clarifying. If this problem recurs in any fashion on 1.3.3, I'll re-open this. Thanks, Kevin On Tue, 13 Oct 1998 [EMAIL PROTECTED] wrote: Please try to duplicate this behavior with version 1.3.3. We can't support old versions of Apache. I had thought that 1.2.6 was still supported to some degree and that there There may be a 1.2.7 release if necessary, but there are so many bug fixes that have gone in between 1.2.x and 1.3.x that we don't have the time to spend trying to debug problems that may well be fixed in the current code.
RE: protocol/2614: Upgrafe from Apache 1.0 to 1.3 on Solaris 2.4 Breaks Netscape Proxy 1.0
Currently upgrading to Solaris 2.6. Will try again once the upgrade is completed. Martin -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Friday, October 09, 1998 9:54 PM To: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject:Re: protocol/2614: Upgrafe from Apache 1.0 to 1.3 on Solaris 2.4 Breaks Netscape Proxy 1.0 [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: Upgrafe from Apache 1.0 to 1.3 on Solaris 2.4 Breaks Netscape Proxy 1.0 Comment-Added-By: lars Comment-Added-When: Fri Oct 9 18:54:15 PDT 1998 Comment-Added: [This is a standard response.] This Apache problem report has not been updated recently. Please reply to this message if you have any additional information about this issue, or if you have answers to any questions that have been posed to you. If there are no outstanding questions, please consider this a request to try to reproduce the problem with the latest software release, if one has been made since last contact. If we don't hear from you, this report will be closed. If you have information to add, BE SURE to reply to this message and include the [EMAIL PROTECTED] address so it will be attached to the problem report!
mod_so/3189: httpd: module Segmentation Fault at startup using DSO
Number: 3189 Category: mod_so Synopsis: httpd: module Segmentation Fault at startup using DSO Confidential: no Severity: serious Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Mon Oct 12 04:10:00 PDT 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.3 Environment: OS: Solaris 2.6 Compiler: SUN Proworks cc (/usr/ucb/cc) Description: Loading a module at runtime with DSO used to work fine with Apache 1.3.0 on the exact same system. I tried compiling mod_status as shared. This is how I configured Apache 1.3.3: CC=cc ./configure --prefix=/opt/local/apache/1.3.3 --enable-module=so --enable-shared=log_referer --enable-rule=SHARED_CORE And this is what it says when I run it (with a LoadModule directive): httpd: module Segmentation Fault No coredump (as far as I can see) or nothing, just that. How-To-Repeat: Fix: Compare 1.3.0 and 1.3.3, and see what has changed. Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
general/3190: httpd processes hang and never free up their slot, until MaxClients reached
Number: 3190 Category: general Synopsis: httpd processes hang and never free up their slot, until MaxClients reached Confidential: no Severity: serious Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Mon Oct 12 08:00:01 PDT 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.2.6 Environment: FreeBSD noldo.pair.com 2.1.7.1-RELEASE FreeBSD 2.1.7.1-RELEASE #1: Wed Nov 12 09:53:51 EST 1997 [EMAIL PROTECTED]:/usr/src/sys/compile/PAIRd i386 gcc version 2.6.3 (also observed under FreeBSD 2.2.7-STABLE) Description: We have only observed this recently on two or three of our 150+ Web servers. What is happening is that on an otherwise normal and relatively quiet server, the httpd processes will stick around in the R state, with no activity, until eventually all slots are full. The most obvious datum is that on the mod_status report, the field Acc is always of the form X/Y/Y, where the last two numbers are the same. This is *not* true on any of our servers except for a couple. We have checked for SYN floods, unusual network activity, tried disabling busier sites on the server one at a time, cleaned up all .htaccess files, etc. We are not using serialized accept. We have tried with and without KeepAlive, at different levels. Our Timeout is set to 2400. I have verified that the configuration is identical to all other servers, except for the virtual hosts listed. ktrace on the parent shows only a series of sleeps and signals. ktrace on any child shows no system calls whatsoever. Attached below is the leftmost columns of our mod_status output after the server runs for a few minutes. HTMLHEAD TITLEApache Status/TITLE /HEADBODY H1Apache Server Status for *.pair.com/H1 Current Time: Mon Oct 12 10:36:08 1998 br Restart Time: Mon Oct 12 10:26:32 1998 br Server uptime: 9 minutes 36 secondsbr Total accesses: 3132 - Total Traffic: 12.0 MBbr CPU Usage: u3.79688 s5.71875 cu14.6406 cs7.21094 - 5.45% CPU loadbr 5.44 requests/sec - 21.4 kB/second - 4032 B/requestbr 172 requests currently being processed, 17 idle servers PRE RKRRR_RRR_RRR___RRR__RRR__RRR_WW__W_W... /PRE Scoreboard Key: br Bcode_/code/B Waiting for Connection, BcodeS/code/B Starting up, BcodeR/code/B Reading Request,BR BcodeW/code/B Sending Reply, BcodeK/code/B Keepalive (read), BcodeD/code/B DNS Lookup,BR BcodeL/code/B Logging, BcodeG/code/B Gracefully finishing, Bcode./code/B Open slot with no current processP p table border=0trthSrvthPIDthAccthMthCPU thSSthConnthChildthSlotthHostthVHostthRequest/tr trtdb0/btd10628td0/20/20tdbR/b td0.42td458td0.0td0.14td0.14 td199.227.99.68/tr trtdb1/btd10627td0/23/23tdbR/b td0.12td480td0.0td0.16td0.16 td206.181.40.100/tr trtdb2/btd10626td0/0/0tdbR/b td0.00td572td0.0td0.00td0.00 td/tr trtdb3/btd10625td0/19/19tdbR/b td0.19td499td0.0td0.12td0.12 td198.8.253.11/tr trtdb4/btd10624td0/5/5tdbR/b td0.02td556td0.0td0.02td0.02 td193.14.10.129/tr trtdb5/btd10623td0/5/5tdbR/b td0.06td554td0.0td0.00td0.00 td209.143.18.64/tr trtdb6/btd10622td0/0/0tdbR/b td0.00td572td0.0td0.00td0.00 td/tr trtdb7/btd10621td0/39/39tdbR/b td0.20td422td0.0td0.15td0.15 td12.75.167.161/tr trtdb8/btd10620td0/21/21tdbR/b td0.05td518td0.0td0.07td0.07 td147.155.3.52/tr trtdb9/btd10619td0/4/4tdbR/b td0.14td516td0.0td0.04td0.04 td209.43.151.208/tr trtdb10/btd10618td0/15/15tdbR/b td0.34td498td0.0td0.06td0.06 td207.226.104.4/tr trtdb11/btd10617td0/1/1tdbR/b td0.00td570td0.0td0.00td0.00 td38.231.95.149/tr trtdb12/btd10616td0/42/42tdbR/b td0.35td402td0.0td0.13td0.13 td152.163.217.99/tr trtdb13/btd10615td0/10/10tdbR/b td0.02td518td0.0td0.07td0.07 td207.203.199.130/tr trtdb14/btd10614td0/8/8tdbR/b td0.08td537td0.0td0.05td0.05 td167.240.254.1/tr trtdb15/btd10613td0/1/1tdbR/b td0.00td568td0.0td0.00td0.00 td199.227.99.68/tr trtdb16/btd10612td0/6/6tdbR/b td0.02td553td0.0td0.03td0.03 td143.183.152.10/tr trtdb17/btd10611td0/11/11tdbR/b td0.06td537td0.0td0.04td0.04 td167.240.254.1/tr trtdb18/btd10610td0/0/0tdbR/b td0.00td571td0.0td0.00td0.00 td/tr trtdb19/btd10609td0/30/30tdbR/b td0.07td461td0.0td0.05td0.05 td150.176.210.3/tr trtdb20/btd10608td0/26/26tdbR/b td0.27td480td0.0td0.14td0.14 td206.181.40.100/tr trtdb21/btd10607td0/5/5tdbR/b td0.19td556td0.0td0.05td0.05 td130.126.241.232/tr trtdb22/btd10606td0/20/20tdbR/b td0.08td479td0.0td0.11td0.11 td209.48.216.1/tr trtdb23/btd10605td0/11/11tdbR/b td0.05td536td0.0td0.01td0.01 td166.84.249.234/tr trtdb24/btd10637td0/16/16tdbR/b td0.17td517td0.0td0.04td0.04 td167.240.254.1/tr trtdb25/btd10641td0/10/10tdbR/b td0.04td537td0.0td0.04td0.04 td12.79.33.118/tr trtdb26/btd10643td0/6/6tdbR/b td0.02td553td0.0td0.02td0.02
mod_perl/3135: httpd eats all available cpu cycles and does not die or respond
Number: 3135 Category: mod_perl Synopsis: httpd eats all available cpu cycles and does not die or respond Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Sun Oct 4 17:30:00 PDT 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.1 Environment: Redhat 5.0 Linux kernel: 2.0.35, Intel pent 150mhz Description: I hope I am sending this to the right people... I have installed a copy of apache 1.3.1 and mod_perl 1.15 from redhat in src.rpm form. they both compile fine, but when mod_perl is installed and run. It starts the httpd processes, they immediately expand to take up all available process time, but do not respond to http requests of any kind. If I comment out the LoadModule and AddModule directives in the httpd.conf then the server returns to normal, but no mod_perl. How-To-Repeat: Fix: I have read ( dejanews) that DSO is broken on ELF file systems (like mine). Is this true? Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
mod_autoindex/3130: Directories have size shown as 0k instead of - in Fancy Indexing
Number: 3130 Category: mod_autoindex Synopsis: Directories have size shown as 0k instead of - in Fancy Indexing Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: mistaken Submitter-Id: apache Arrival-Date: Sat Oct 3 12:00:00 PDT 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.2 Environment: - Description: Fancy Indexing shows size of directories as 0k instead of -. In Apache 1.2.6 it was OK. Probably only a mistake (if not intended) in Apache 1.3.x (even 1.3.3-dev). How-To-Repeat: See e.g. http://www.apache.org/dist/ Fix: The following patch for Apache 1.3.1 should solve it: @@ -762,7 +762,7 @@ if (!(p-alt = find_alt(d, rr, 1))) { p-alt = DIR; } - p-size = 0; + p-size = -1; p-name = ap_pstrcat(r-pool, name, /, NULL); } else { Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
Re: mod_negotiation/1987: Apache dumps core when handling unacceptable variants
The following reply was made to PR mod_negotiation/1987; it has been noted by GNATS. From: Martin Plechsmid [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: mod_negotiation/1987: Apache dumps core when handling unacceptable variants Date: Sat, 3 Oct 1998 22:22:42 +0200 (MET DST) The bug is still present in Apache 1.3.1, and nothing has changed about it in 1.3.2. So it's there also. How-to-repeat: Uncoment #define TCN_02 at the beggining of the file mod_negotiation.c and recompile server. Put files xxx.html.en and xxx.html.fr into a directory and send the following headers to the server: GET //xxx.html HTTP/1.0 Negotiate: 1.0 Accept-Language: it The server will crash. Fix: The following patch for Apache 1.3.1 solves the problem: --- mod_negotiation.c.orig Sat Oct 3 22:05:04 1998 +++ mod_negotiation.c Sat Oct 3 22:06:27 1998 @@ -1566,6 +1566,7 @@ variant-definite); #endif +if (q == 0) return 0; if (q bestq) { *p_bestq = q; return 1;
config/3112: official patch not solving htaccess/php3 problem
Number: 3112 Category: config Synopsis: official patch not solving htaccess/php3 problem Confidential: no Severity: serious Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Wed Sep 30 12:30:00 PDT 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.2, 1.3_19980930131214 Environment: Linux 2.0.35 GCC 2.7.2.1 libc 5.4.33 Description: Have compiled apache-1.3.2 and latest CVS snapshot + php3.0.4, not solving the htaccess error. It worked with the first patch from Roy Fielding: diff -u -r1.128 http_config.c --- http_config.c 1998/09/21 21:09:47 1.128 +++ http_config.c 1998/09/25 07:41:49 @@ -1229,7 +1229,10 @@ filename = ap_make_full_path(r-pool, d, w); f = ap_pcfg_openfile(r-pool, filename); } -if (f) { +if (!access_name[0]) { + dc = NULL; +} +else if (f) { but it looked like some other problems occured with cgi execution.. The second official patch gives no results. How-To-Repeat: Fix: Jordan Mendelson wrote: In case anyone cares.. simply adding _REENTRANT=1 to src/Configure like: CFLAGS=$CFLAGS -DLINUX=2 -D_REENTRANT=1 ... posted on comp.infosystems.www.servers.unix Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
general/3017: browser hangs then gives secure channle error
Number: 3017 Category: general Synopsis: browser hangs then gives secure channle error Confidential: no Severity: critical Priority: medium Responsible:apache State: open Class: support Submitter-Id: apache Arrival-Date: Thu Sep 17 02:50:00 PDT 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.1 Environment: BSD/OS lon1.dpe.net 3.0 BSDI BSD/OS 3.0 Kernel #0: Thu Jan 30 13:02:43 MST 1997 [EMAIL PROTECTED]:/rel/proto/3.0RC1-i386/usr/src/sys/compile/GENERIC i 386 Description: The following is logged to our error_log We have built the server with mod_ssl wanted to rule out general server config problem first [error] Invalid method in request % How-To-Repeat: point browser at https://www1.dpe.net Fix: Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
Re: config/2958: UserDir disabled doesn't work
The following reply was made to PR config/2958; it has been noted by GNATS. From: Martin Husemann [EMAIL PROTECTED] To: [EMAIL PROTECTED] (Marc Slemko) Cc: [EMAIL PROTECTED] Subject: Re: config/2958: UserDir disabled doesn't work Date: Sun, 6 Sep 1998 07:34:12 +0200 (MEST) Are you sure it isn't being cached by your browser or by a proxy? Are you sure you don't have any other userdir settings or rewrite settings that could be impacting this? Ooops... Must have been cached by my browser. I can't reproduce it myself now. Please close the PR, very sorry for the noise Martin
config/2958: UserDir disabled doesn't work
Number: 2958 Category: config Synopsis: UserDir disabled doesn't work Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Sat Sep 5 14:20:00 PDT 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.1 Environment: NetBSD rumolt.teuto.de 1.3G NetBSD 1.3G (RUMOLT) #0: Fri Aug 21 10:55:15 CEST 1998 Compiler: egcs-2.90.27 980315 (egcs-1.0.2 release) Description: My server configuration includes (in srm.conf) the following lines: ---8--- # UserDir: The name of the directory which is appended onto a user's home # directory if a ~user request is recieved. UserDir disabled root UserDir public_html ---8--- as suggested by the install documents. Now to try this I made a directory ~root/public_html and an index.html file there. The effect is: request to (servername)/~root fail, but (servername)/~root/ works und gives the right index.html file. You can call this pilot error, that's why I set the severity to non-critical. How-To-Repeat: See above. Fix: Either fix it (i.e. let the server return access denied for that URI) or document it. Sorry, no patches available. Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
Re: mod_proxy/2914: Segmentation fault on invalid server response
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: Segmentation fault on invalid server response State-Changed-From-To: open-closed State-Changed-By: martin State-Changed-When: Mon Aug 31 07:45:12 PDT 1998 State-Changed-Why: I wish I had read your report earlier ;-) I fixed the problem independently (sorry, without mentioning you) in Rev. 1.59 of proxy_http.c on 1998/08/27. You are right, an improvement would be to accept, log, and ignore microsoft specific bugs in the proxy.
config/2744: cannot get make to complete
Number: 2744 Category: config Synopsis: cannot get make to complete Confidential: no Severity: critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Thu Jul 30 08:10:00 PDT 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.1 Environment: SunOS rmdsu5 5.5.1 Generic sun4m sparc SUNW,SPARCstation-20 gcc -v = gcc version cygnus-2.3.3 Description: attempting to make httpd === os/unix gcc -c -I../../os/unix -I../../include -DSOLARIS2=251 os.c gcc -c -I../../os/unix -I../../include -DSOLARIS2=251 os-inline.c rm -f libos.a ar cr libos.a os.o os-inline.o ranlib libos.a === os/unix === ap gcc -c -I../os/unix -I../include -DSOLARIS2=251 ap_execve.c gcc -c -I../os/unix -I../include -DSOLARIS2=251 ap_cpystrn.c gcc -c -I../os/unix -I../include -DSOLARIS2=251 ap_signal.c ap_signal.c: In function `ap_signal': ap_signal.c:68: storage size of `act' isn't known ap_signal.c:68: storage size of `oact' isn't known *** Error code 1 make: Fatal error: Command failed for target `ap_signal.o' Current working directory /s19/3/BinBuild/apache_1.3.1/src/ap *** Error code 1 make: Fatal error: Command failed for target `subdirs' error: httpd binary not made ERROR: abnormal exit, binary release not properly built How-To-Repeat: run make Fix: No idea. Any hints as to what is wrong with my system setup would be greatly appreciated. Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ]
Re: documentation/2359: corrections needed in htdocs/manual/suexec.html
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] Synopsis: corrections needed in htdocs/manual/suexec.html State-Changed-From-To: open-closed State-Changed-By: martin State-Changed-When: Thu Jul 30 09:44:32 PDT 1998 State-Changed-Why: I modified suexec.html for the 1.3.2-dev version of apache, based on your suggestions. Thanks for helping Apache!
RE: protocol/2614: Upgrafe from Apache 1.0 to 1.3 on Solaris 2.4 Breaks Netscape Proxy 1.0
The following reply was made to PR protocol/2614; it has been noted by GNATS. From: [EMAIL PROTECTED] (Giguere, Martin) To: 'Marc Slemko' [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: RE: protocol/2614: Upgrafe from Apache 1.0 to 1.3 on Solaris 2.4 Breaks Netscape Proxy 1.0 Date: Mon, 20 Jul 1998 10:21:55 -0400 I will have the box upgraded to the Latest Solaris and let you know. In the mean time I am using Apache 1.2.6. Martin Giguere EDS Canada [EMAIL PROTECTED] voice (905)644-3640 fax (905)644-7509 -Original Message- From: Marc Slemko [mailto:[EMAIL PROTECTED] Sent: July 15, 1998 12:11 AM To: Giguere, Martin Cc: [EMAIL PROTECTED] Subject: RE: protocol/2614: Upgrafe from Apache 1.0 to 1.3 on Solaris 2.4 Breaks Netscape Proxy 1.0 On Wed, 15 Jul 1998, Giguere, Martin wrote: The basic problem is that I am unable to retrieve certain files from Apache 1.3.0 running on Solaris 2.4 but I am able to do it with Apache 1.3.0 running on Redhat Linux 5.1 i386. I try with various browser yet it still won't display the image, but if I try it from home, everything loads fine. I unfortunately do not control the Sparc environment nor am I able to get a newer version of the proxy server installed. Do you have the latest Solaris patch cluster installed? If not, install it. This has fixed similar problems for other people.
protocol/2614: Upgrafe from Apache 1.0 to 1.3 on Solaris 2.4 Breaks Netscape Proxy 1.0
Number: 2614 Category: protocol Synopsis: Upgrafe from Apache 1.0 to 1.3 on Solaris 2.4 Breaks Netscape Proxy 1.0 Confidential: no Severity: critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Tue Jul 14 12:10:01 PDT 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3 Environment: The Web Server is running on SunOS 5.4 Generic_101945-10 sun4m sparc The Proxy Server is Netscape Proxy Server 1.0 Description: I am running Apache 1.0 on port 80 and Apache 1.3 on port 8080. When I try to connect to version 1.3 from behind a Netscape Proxy Server version 1.0 I get the following error. I am still running both version (1.0 in the production environment, 1.3 as a test for now) Both Apache versions point to the same document root. Note that this error only shows with gif files of a certain size. I am anxious to move my productive environment to 1.3, but this problem is holding me back. good (Apache 1.0 access_log) canlink.eds.com - - [14/Jul/1998:14:44:41 -0400] GET /images/title.gif HTTP/1.0 304 - http://www.eds.ca/images/title.gif bad (Apache 1.3 access_log) 205.239.196.2 - - [14/Jul/1998:14:36:58 -0400] GET /images/title.gif HTTP/1.0 200 9971 http://www.eds.ca:8080/images/title.gif If you have any questions please feel free to contact me. How-To-Repeat: good (Apache 1.0 access_log) canlink.eds.com - - [14/Jul/1998:14:44:41 -0400] GET /images/title.gif HTTP/1.0 304 - url http://www.eds.ca/images/title.gif bad (Apache 1.3 access_log) 205.239.196.2 - - [14/Jul/1998:14:36:58 -0400] GET /images/title.gif HTTP/1.0 200 9971 url http://www.eds.ca:8080/images/title.gif Fix: version 1.0 works fine. Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ]
os-windows/2503: Install
Number: 2503 Category: os-windows Synopsis: Install Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: support Submitter-Id: apache Arrival-Date: Thu Jun 25 10:10:00 PDT 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3 Environment: WinNT4.0 Description: Where and what is the filename to get the v1.3 of this server. I can't locate it. How-To-Repeat: Fix: Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ]
Re: apache-api/2337: Modules are initialized twice at startup.
Ben Laurie writes: The justification I was given, a long time ago, was that modules are reinitialised on a server restart (i.e. SIGHUP or SIGUSR1) so they may as well go wrong at startup if they can't deal with that. I believe that that is just a rationalisation, though. The real reason is that it is simpler to let it do it that way. Ok, but why then does the SSL patch subvert that behaviour? One problem, for example, is that the FastCGI module expects to be called twice at startup, but after the SSL patch, this is not true anymore so it fails to initialize itself. It took me some time to figure that out. In general, module initialization can be costly, so Apache shouldn't initialize all its modules twice at startup, there's just no reason for it, as far as I can see.
Re: apache-api/2337: Modules are initialized twice at startup.
[EMAIL PROTECTED] writes: Synopsis: Modules are initialized twice at startup. State-Changed-From-To: open-closed State-Changed-By: marc State-Changed-When: Mon Jun 1 15:55:49 PDT 1998 State-Changed-Why: Modules are supposed to be initialized twice, and they have been for a long time and will be for the forseeable future in the 1.x code. I suspect your confusion comes from some third-party patch you applied to your 1.2.x code that makes initialization only happen once. Right. It was added by the SSL patch. I'm surprised. Note that doing so can break some modules that rely on it happening twice. Unfortunately, I couldn't find documentation on that behaviour, it doesn't make sense to me to call the initializers twice, what's the reason? Thanks for the clarifications, Martin
Re: mod_proxy/2066: JDK1.2beta[23] have a bug in the URLConnection - here is simple server side workaround
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] Synopsis: JDK1.2beta[23] have a bug in the URLConnection - here is simple server side workaround State-Changed-From-To: open-analyzed State-Changed-By: martin State-Changed-When: Thu May 14 14:25:08 PDT 1998 State-Changed-Why: Excellent analysis. Responsible-Changed-From-To: apache-martin Responsible-Changed-By: martin Responsible-Changed-When: Thu May 14 14:25:08 PDT 1998 Responsible-Changed-Why: I'll have another look at the cache source, but your fix looks very good. Category-Changed-From-To: other-mod_proxy Category-Changed-By: martin Category-Changed-When: Thu May 14 14:25:08 PDT 1998
Re: other/2066: JDK1.2beta[23] have a bug in the URLConnection - here is simple server side workaround
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] Synopsis: JDK1.2beta[23] have a bug in the URLConnection - here is simple server side workaround State-Changed-From-To: analyzed-open State-Changed-By: martin State-Changed-When: Thu May 14 14:29:00 PDT 1998 State-Changed-Why: I'm terribly sorry. I edited the wrong PR#. Responsible-Changed-From-To: martin-apache Responsible-Changed-By: martin Responsible-Changed-When: Thu May 14 14:29:00 PDT 1998 Responsible-Changed-Why: Undo wrong PR-Edit Category-Changed-From-To: mod_proxy-other Category-Changed-By: martin Category-Changed-When: Thu May 14 14:29:00 PDT 1998
Re: mod_proxy/2094: Documents get not cached, when updated from origin host *and* Content-Length has changed
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] Synopsis: Documents get not cached, when updated from origin host *and* Content-Length has changed State-Changed-From-To: open-analyzed State-Changed-By: martin State-Changed-When: Thu May 14 14:30:48 PDT 1998 State-Changed-Why: Excellent analysis. Responsible-Changed-From-To: apache-martin Responsible-Changed-By: martin Responsible-Changed-When: Thu May 14 14:30:48 PDT 1998 Responsible-Changed-Why: I'm going to have another look at proxy_cache.c, but it looks like your patch fixes the bug. Thank you!
os-windows/1851: fopen and httpd error on startup
Number: 1851 Category: os-windows Synopsis: fopen and httpd error on startup Confidential: no Severity: serious Priority: medium Responsible:apache State: open Class: support Submitter-Id: apache Arrival-Date: Fri Feb 20 11:00:02 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.5 Environment: Windows NT SP3, MSVC 5.0 Description: When I attempt to start apache (apache /s or apache -s), I recieve the message fopen: No such file or directory httpd: could not open document config file at /apache/conf/httpd.conf (please notice the direction of the slashes). Apache then quits. How-To-Repeat: Fix: none, sorr Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ]
Re: mod_proxy/1831: ProxyDomain directive logs incorrect error messages
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] Synopsis: ProxyDomain directive logs incorrect error messages State-Changed-From-To: open-closed State-Changed-By: martin State-Changed-When: Thu Feb 19 11:14:45 PST 1998 State-Changed-Why: The report was not based on 1.3b5, but on part of an older apache version which had slipped in my source base. Class-Changed-From-To: sw-bug-mistaken Class-Changed-By: martin Class-Changed-When: Thu Feb 19 11:14:45 PST 1998 Responsible-Changed-From-To: apache-martin Responsible-Changed-By: martin Responsible-Changed-When: Thu Feb 19 11:14:45 PST 1998 Responsible-Changed-Why: I wrote the PR (sorry), I better close it now..
mod_proxy/1831: ProxyDomain directive logs incorrect error messages
Number: 1831 Category: mod_proxy Synopsis: ProxyDomain directive logs incorrect error messages Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Wed Feb 18 08:50:00 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3b5 Environment: any Description: The ProxyDomain directive uses aplog_error() to log possible redirections. Unfortunately, an URI can (and often will) contain '%'-encoded metachars like '~'-'%7E' which are then mistaken as printf() percent tokens. Oh, and the table_get(Referer) semantics apparently changed: it now returns - instead of NULL which then leads to a log message ... from - (the intention was to suppress the from part if the referer was undefined). How-To-Repeat: Fix: For the first part, the fix is easy: add an extra %s, in front of the pstrdup() call in the aplog_error() call. The second problem is easy to fix as well: check for (ref!= NULL !strcmp(ref, -)) Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ]
mod_proxy/1832: ProxyDomain directive logs incorrect error messages
Number: 1832 Category: mod_proxy Synopsis: ProxyDomain directive logs incorrect error messages Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Wed Feb 18 08:50:01 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3b5 Environment: any Description: The ProxyDomain directive uses aplog_error() to log possible redirections. Unfortunately, an URI can (and often will) contain '%'-encoded metachars like '~'-'%7E' which are then mistaken as printf() percent tokens. Oh, and the table_get(Referer) semantics apparently changed: it now returns - instead of NULL which then leads to a log message ... from - (the intention was to suppress the from part if the referer was undefined). How-To-Repeat: Fix: For the first part, the fix is easy: add an extra %s, in front of the pstrdup() call in the aplog_error() call. The second problem is easy to fix as well: check for (ref!= NULL !strcmp(ref, -)) Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ]
Re: os-other/1588: GuessOS does not know ReliantUNIX
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] Synopsis: GuessOS does not know ReliantUNIX State-Changed-From-To: feedback-closed State-Changed-By: martin State-Changed-When: Fri Feb 13 08:48:02 PST 1998 State-Changed-Why: In the forthcoming 1.2.6 release (as well as in the current 1.3 beta), GuessOS was corrected appropriately to detect ReliantUNIX and SINIX as equivalent. Additionally, the _XPG_IV define is set to work around an interface problem of the gettimeofday() call. So, 1.2.6 should run as-is on both SINIX and ReliantUNIX. Responsible-Changed-From-To: apache-martin Responsible-Changed-By: martin Responsible-Changed-When: Fri Feb 13 08:48:02 PST 1998 Responsible-Changed-Why: I work for Siemens Nixdorf, so you best complain to me if it does NOT work ;-) %0
general/1365: Large download from IE broken up and logged many times
Number: 1365 Category: general Synopsis: Large download from IE broken up and logged many times Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Wed Nov 5 12:20:00 PST 1997 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.2.4 Environment: SunOS 5.4 Generic_101945-34 sun4m sparc Description: When a Internet Explorer client iniates a large download, seen at 400K, the file is transmited in stages. This causes the sames file and transaction to be logged seperate for each stage. This in turn causes most log analysis to overstate the hit count for this document. Seen with IE3 IE4, However not seen with Ntscpe 4. Didn't see this prior version. How-To-Repeat: Fix: Audit-Trail: Unformatted:
protocol/1194: error 500 generated with SCRIPT
Number: 1194 Category: protocol Synopsis: error 500 generated with SCRIPT Confidential: no Severity: serious Priority: medium Responsible:apache (Apache HTTP Project) State: open Class: support Submitter-Id: apache Arrival-Date: Fri Oct 3 07:50:00 1997 Originator: [EMAIL PROTECTED] Organization: apache Release:1.2.4 Environment: Sun Sparcstaion 2, uname -a gives SunOS powers 4.1.3 3 sun4c compilor is gcc (I think) Description: MY news feed is dead so I've recieved no help from there. when I try to access a HTML file conatining :- SCRIPT LANGUAGE=JavaScript some code here.. /SCRIPT the browser gives an error 500 and this is the message I get in my error_log file test.html failed for dalwhinnie, reason: Premature end of script headers How-To-Repeat: you can get the error 500 message by accessing http://powers.dcs.st-andrews.ac.uk:82/warp/finesse-bin/test.html Fix: Audit-Trail: Unformatted:
Re: protocol/1194: error 500 generated with SCRIPT
The following reply was made to PR protocol/1194; it has been noted by GNATS. From: Martin N.T. Bramley [EMAIL PROTECTED] To: Marc Slemko [EMAIL PROTECTED] Subject: Re: protocol/1194: error 500 generated with SCRIPT Date: Fri, 3 Oct 1997 16:36:12 +0100 (BST) You really need to go through your configuration again. For whatever reason, your Apache is setup so that it thinks test.html is a CGI. Otherwise it would not try to run it. The error log message clearly indicates it is trying to execute it as a CGI. If you remove the contents from it and just put a single line of text in test.html, does it still have the same problem? If I have plain HTML without using the SCRIPT /SCRIPT it works absolutly fine it appears only to this one directive which doesn't work The error_log seam to suggest the SCRIPT LANGUAGE=JavaScript is more incomplete reather than unacceptable. Martin
config/851: Default switches for SNI machines should be improved
Number: 851 Category: config Synopsis: Default switches for SNI machines should be improved Confidential: no Severity: non-critical Priority: medium Responsible:apache (Apache HTTP Project) State: open Class: change-request Submitter-Id: apache Arrival-Date: Thu Jul 10 08:20:01 1997 Originator: [EMAIL PROTECTED] Organization: apache Release:1.2.1 Environment: SINIX 5.4 (SVR4) on different platforms SINIX-D 5.41 i386 SINIX-N 5.43 MIPS R4000 / ReliantUNIX /opt/bin/cc or /usr/ccs/bin/cc or gcc Description: Apache compiles (almost) painlessly under SVR4, except for the gettimeofday() call, which was broken by the SYSV people: mod_usertrack.c 150: [error]: CFE1140 too many arguments in function call gettimeofday(tv, tz); ^ The system has the following definition for gettimeofday(), and apparently does NOT define _XPG_IV by default: #if defined(_XPG_IV) int gettimeofday(struct timeval *, void *); #else int gettimeofday(struct timeval *); #endif /* defined(_XPG_IV) */ How-To-Repeat: Fix: The following patch fixes this problem for SINIX (this is the current SVR4 version of SIEMENS NIXDORF (=SNI)). It also defines the shared memory flags which work correctly in SINIX versions. --- Configure.orig Thu Jul 3 07:59:43 1997 +++ Configure Thu Jul 10 16:53:16 1997 @@ -423,7 +423,7 @@ ;; *-sni-sysv4*) OS='SVR4' - CFLAGS=$CFLAGS -DSVR4 + CFLAGS=$CFLAGS -DSVR4 -D_XPG_IV -DHAVE_MMAP -DHAVE_SHMGET DEF_WANTHSREGEX=yes LIBS=$LIBS -lsocket -lnsl -lc ;; %0 Audit-Trail: Unformatted:
other/776: httpd_monitor won't compile
Sorry about this ... during an upgrade, I mixed up the old and new code for the support directory. So, this bug report is not applicable and can be deleted. Martin On Tue, 24 Jun 1997 [EMAIL PROTECTED] wrote: I'm not sure what you are talking about. In the distributed httpd_monitor.c from 1.2.0, httpd.h is included before scoreboard.h. Please be sure you have made no local changes and try redownloading the file.
other/776: httpd_monitor won't compile
Number: 776 Category: other Synopsis: httpd_monitor won't compile Confidential: no Severity: non-critical Priority: medium Responsible:apache (Apache HTTP Project) State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Mon Jun 23 18:50:01 1997 Originator: [EMAIL PROTECTED] Organization: apache Release:1.2.0 Environment: OSF1 V4.0 464 alpha, using the Digital's bundled cc (DEC C V5.2-030 on Digital UNIX V4.0 (Rev. 464)) Description: [EMAIL PROTECTED]:~web/support 114-ROOT# make CC=cc cc -I../src -g httpd_monitor.c -o httpd_monitor cc: Error: ../src/scoreboard.h, line 102: In the declaration of servers, HARD_SERVER_LIMIT is not declared. short_score servers[HARD_SERVER_LIMIT]; ^ *** Exit 1 Stop. How-To-Repeat: Fix: Solution is simply to transpose the following two lines in httpd_monitor.c: #include scoreboard.h #include httpd.h so that httpd.h is read in before scoreboard.h is. %0 Audit-Trail: Unformatted:
protocol/610: Multiple virtual hosts over single connection don't work
Number: 610 Category: protocol Synopsis: Multiple virtual hosts over single connection don't work Confidential: no Severity: serious Priority: medium Responsible:apache (Apache HTTP Project) State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Fri May 23 01:50:00 1997 Originator: [EMAIL PROTECTED] Organization: apache Release:1.2b10 Environment: Linux 2.0.30, GCC 2.7.2.2 Description: If using multiple HTTP/1.1 GET requests on single connection to retrieve data from different virtual hosts, the virtual host list is scanned starting with the most recently used VH, not with the first one which causes that most VH's are ignored. How-To-Repeat: GET / HTTP/1.1, Host: first_virtual_host; GET / HTTP/1.1, Host: primary_server_name Fix: Set current_conn-server back to the original value after each call to read_request from the http_main main loop Audit-Trail: Unformatted:
mod_proxy/271: Access control for proxy does not work.
Number: 271 Category: mod_proxy Synopsis: Access control for proxy does not work. Confidential: no Severity: serious Priority: medium Responsible:apache (Apache HTTP Project) State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Thu Mar 27 08:40:01 1997 Originator: [EMAIL PROTECTED] Organization: apache Release:1.2b8-dev Environment: SVR4-intel Description: I'm using apache with the mod_proxy module and the following access control(s): IfModule mod_proxy.c Directory proxy:* order deny,allow deny from all allow from 127.0.0.1 139.25.113.10 192.168.123.1 #allow from 139.25.112.104 /Directory /IfModule Then I try to access http://www.geocities.com/ from the host 139.25.112.104 and get (correctly): [Thu Mar 27 17:06:54 1997] access to proxy:http://www.geocities.com/ failed for pgtd0119, reason: Client denied by server configuration pgtd0119 unknown - [27/Mar/1997:17:16:42 +0100] GET http://www.geocities.com/ HTTP/1.0 403 1089 But when I send a second request http://www.geocities.com/foo.bar then the server passes the request to www.geocities.com, i.e., it performs the proxy service that should be disallowed: pgtd0119 unknown - [27/Mar/1997:17:16:53 +0100] GET http://www.geocities.com/foo.bar HTTP/1.0 404 1064 BTW: It would be nice if proxy (or any) access could be limited on host+path level, not just host level. How-To-Repeat: See above. Fix: Audit-Trail: Unformatted: Kraemer [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] X-send-pr-version: 3.2