- 12 Jul, 1998 6 commits
-
-
Jim Jagielski authored
PR: Obtained from: Submitted by: Reviewed by: Oops... gotta recall that older (and esp. SysV-based) tr's require the '[]' method. This works fine with BSD tr's since it maps the [ to [ and the ] to ]. Cover all bases. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81713 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
we forgot them for the next release ;-) git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81712 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
change its priority, it doesn't hurt is when we are additionally be even more conservative: Let mod_mmap_static be more reserved unless at least one "mmapfile" directive is present in the configuration. Because this experimental module has to do some black magic to operate inside the current API and thus forces side-effects for other modules under some circumstances. PS: For Apache 2.0 we really need dynamically configurable priorities _per API hook_ for each module, yes.... git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81711 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
to avoid problems under brain-dead platforms like Digital Unix (OSF1). PR: 2596 PS: The error on OSF1 also could mean that OSF1 really wants "egrep -e <pattern>" instead of "egrep <pattern>". But when this is the problem other egreps in src/Configure also would fail. Because they don't we can assume that the problem for OSF1 is really the missing quotes because of the ^ char. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81710 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81709 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
At least under the Win32 environment it is not available. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81708 13f79535-47bb-0310-9956-ffa450edef68
-
- 11 Jul, 1998 4 commits
-
-
dgaudet authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81707 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81706 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
thus the child processes had no write access to the files. Now a chown() is done to the uid of the childs if applicable. Submitted by: Lars Eilebrecht Reviewed and fixed by: Ralf S. Engelschall PR: 2341 PS: Lars, I've changed s->server_uid to ap_user_id because s->server_uid can be different inside virtual hosts for the suEXEC mechanism. But we need the uid of the process, so ap_user_id is correct IMHO. And I've searched for the PR in the bugdb for you and noted it above. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81705 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
================================================== PR#2093, PR#2361, PR#2377, PR#2434, PR#2524, PR#2525, PR#2533, PR#2569 Background ---------- Currently we have some places (especially in conf.h and os/unix/os.h) where we make decisions about the existence of particular header files based on the underlying platform. In general this is sufficient, but expercience showed (especially the Linux/dlfcn.h PRs) that making general implications from platforms to file existance is very problematic because of too much variants and version of particular platforms. Problem ------- The problem is that we decide particular header files exists even when they don't exist. This caused compile time failures and a lot of PRs. Solution -------- The short-hand solution in the past was to use ``-D<PLATFORM>=<VERSION>'' defines instead of ``-D<PLATFORM>'' and use ``#if defined(<PLATFORM>) && <PLATFORM> >= <VERSION>'' instead of ``#ifdef <PLATFORM>''. This is ok for platforms where we really have access to and where we can be sure that a particular platform version has a file or hasn't it. This is not the case for esoteric Unix derivates and not the case of the huge amount of Linux variants. So a long-term solution is needed. As often GNU Autoconf is the father of the idea: We actually check for the existance of particular header files and define HAVE_XXXX_H if ``#include <xxxx.h>'' works (which means xxxx.h exists). The patch --------- The patch actually does the following: 1. It src/Configure another check-cycle is done where we test for various header files which later conditionally can be included via #ifdef HAVE_XXX_H #include <xxx.h> #endif This is done by calling TestCompile with a new command "header" which checks for existing header files. As a result the file include/conf_auto.h is generated containing these HAVE_XXX_H defines. Currently the following header files are tested (can be extended in the future): dlfcn.h dl.h bstring.h crypt.h unistd.h sys/resource.h sys/select.h sys/processor.h 2. In include/conf.h all HAVE_XXX_H defines/undefines were replaced by a single ``#include "conf_auto.h"'' statement at the top. And in all related files the inclusion of such headers are now based on HAVE_XXX_H defines. The generated conf_auto.h file looks like this: | /* Automatically generated file - DO NOT EDIT */ | #ifndef APACHE_CONF_AUTO_H | #define APACHE_CONF_AUTO_H 1 | | /* <dlfcn.h> */ | #ifndef HAVE_DLFCN_H | #define HAVE_DLFCN_H 1 | #endif | | /* <dl.h> */ | #ifdef HAVE_DL_H | #undef HAVE_DL_H | #endif | : | : | #endif /* APACHE_CONF_AUTO_H */ Submitted by: Ralf S. Engelschall Reviewed by: Dean Gaudet, Jim Jagielski PR: 2093, 2361, 2377, 2434, 2524, 2525, 2533, 2569 git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81704 13f79535-47bb-0310-9956-ffa450edef68
-
- 10 Jul, 1998 8 commits
-
-
rasmus authored
modules to load their own modules dynamically. This improves mod_perl and mod_php3 when these modules are loaded dynamically into Apache. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81703 13f79535-47bb-0310-9956-ffa450edef68
-
Ken Coar authored
Add a missing 'Override' clause. PR: 2582 Submitted by: Andrew Gillespie <apache@FutureQuest.net> Reviewed by: Ken Coar git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81702 13f79535-47bb-0310-9956-ffa450edef68
-
Ken Coar authored
Make mod_setenvif work like mod_rewrite - namely, let it use "^$" to match missing fields. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81700 13f79535-47bb-0310-9956-ffa450edef68
-
Marc Slemko authored
out how is good; telling people the most probable cause (ie. lack of #! in a script) is even better. PR: Obtained from: Submitted by: Reviewed by: git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81699 13f79535-47bb-0310-9956-ffa450edef68
-
Marc Slemko authored
to get access to a directory because they aren't in the list of those allowed access. PR: Obtained from: Submitted by: Reviewed by: git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81698 13f79535-47bb-0310-9956-ffa450edef68
-
Marc Slemko authored
it. PR: Obtained from: Submitted by: Reviewed by: git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81697 13f79535-47bb-0310-9956-ffa450edef68
-
Ken Coar authored
Yes, I know this is style-guide/indent stuff, but I'm tracking down a possible bug and want to have a clean basis for any changes. I.e., I'm not just being capricious.. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81695 13f79535-47bb-0310-9956-ffa450edef68
-
Ken Coar authored
Add a note to the FAQ about increasing file descriptor limits [sic] on Win95. Does this have any effect on WinNT? Submitted by: "The Smiths" <smiths@erols.com> Reviewed by: Ken Coar git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81694 13f79535-47bb-0310-9956-ffa450edef68
-
- 09 Jul, 1998 10 commits
-
-
brian authored
We are not public domain. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81689 13f79535-47bb-0310-9956-ffa450edef68
-
brian authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81687 13f79535-47bb-0310-9956-ffa450edef68
-
Martin Kraemer authored
the configured percentage of the file has already been transfered. It works for http transfers only. The new httpd.conf directive is: CacheForceCompletion <percentage> PR: 2277 Submitted by: Glen Parker <glenebob@nwlink.com> Reviewed by: Martin Kraemer, Brian Behlendorf git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81686 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
ap_http_method() based variant (which is still ok for Apache-SSL because it patches ap_http_method() itself). This way there is now really _no_ sort of crypto hook in the official Apache distribution ;-) git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81684 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
when he renamed the define. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81683 13f79535-47bb-0310-9956-ffa450edef68
-
Ken Coar authored
Add the "<!DOCTYPE HTML" magic cookie for mod_mime_magic. PR: 2574 Submitted by: Anna Shergold <anna@inext.co.uk> Reviewed by: Ken Coar git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81682 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
apply to the complete expression or GCC still complains with: proxy_cache.c:334: warning: long int format, different type arg (arg 6) proxy_cache.c:365: warning: long int format, different type arg (arg 6) git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81681 13f79535-47bb-0310-9956-ffa450edef68
-
dgaudet authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81680 13f79535-47bb-0310-9956-ffa450edef68
-
dgaudet authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81679 13f79535-47bb-0310-9956-ffa450edef68
-
dgaudet authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81678 13f79535-47bb-0310-9956-ffa450edef68
-
- 08 Jul, 1998 9 commits
-
-
dgaudet authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81677 13f79535-47bb-0310-9956-ffa450edef68
-
brian authored
maybe /include/hsregex.h would be better, I'm not sure. This fixes it for now. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81676 13f79535-47bb-0310-9956-ffa450edef68
-
Jim Jagielski authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81675 13f79535-47bb-0310-9956-ffa450edef68
-
Jim Jagielski authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81674 13f79535-47bb-0310-9956-ffa450edef68
-
dgaudet authored
suggestion because Iain Brown <iain@iain.com> pointed out that my logic regarding the casting was wrong. This compiles and runs/works on both Linux and Solaris 2.6. I'm totally willing to back it out if there are complaints, I just figure this is the best way to get widespread multiplatform testing. To test you need to request a URL with character 246 (�), and you should create a .htaccess in a directory with an AddDescription containing character 246 as well. (Actually, trying 240 through 255 would be good.) PR: 800, 2282, 2553 (and others) git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81673 13f79535-47bb-0310-9956-ffa450edef68
-
dgaudet authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81672 13f79535-47bb-0310-9956-ffa450edef68
-
Ralf S. Engelschall authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81671 13f79535-47bb-0310-9956-ffa450edef68
-
Jim Jagielski authored
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81670 13f79535-47bb-0310-9956-ffa450edef68
-
Jim Jagielski authored
PR:2398 Submitted by: Kristian Kraemmer Nielsen <jkkn@ctav.com> Reviewed by: Jim Jagielski Change 'nobody' to 'nogroup for DEFAULT_GROUP under AIX git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81669 13f79535-47bb-0310-9956-ffa450edef68
-
- 07 Jul, 1998 3 commits
-
-
Ken Coar authored
Put back the bits that aren't controversial. :-) git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81668 13f79535-47bb-0310-9956-ffa450edef68
-
dgaudet authored
progress into CHANGES. PR: 800, 2252, 2553 git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81666 13f79535-47bb-0310-9956-ffa450edef68
-
dgaudet authored
error ever occurs then it's a bug in Apache, and we can put debugging code in then. I can't see how Apache would ever make this mistake by inspecting the code in util_script.c. Is there a PR or something that reported this as a bug?? Reverting to 1.39. git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/1.3.x@81665 13f79535-47bb-0310-9956-ffa450edef68
-