1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
|
Apache 2.0 STATUS:
Last modified at [$Date: 2000/05/05 20:19:25 $]
Release:
2.0a4 : vague plans to release mid-May
2.0a3 : released April 28, 2000
2.0a2 : released March 31, 2000
2.0a1 : released March 10, 2000
RELEASE SHOWSTOPPERS:
* Win32: Get mod_auth_digest working under win32
- APR_HAS_RANDOM should be defined on windows and there is a
lib/apr/misc/win32/rand.c which is basically a copy of what
mod_auth_digest used to use.
* Re-work configuration for top level Apache. Work should start with
trying to clean the autoconf stuff. If and only if this proves
impossible to do (very unlikely), autoconf should be removed and we
will roll our own config implementation.
NOTE: When this is implemented, it would be really nice if the
default cgi module on Unix with a threaded MPM is the cgid
module. This should be override-able of course.
Status: Jim Jagielski is looking into this.
* Fix SAFEREAD.
see <Pine.LNX.4.21.0004101511380.7822-100000@twinlark.arctic.org>
* MPMs should have responsibility for closing sockets and
registering the socket cleanup. Stop registering multiple
cleanups for each socket. Move socket close code out of
http_connection.c and into the MPMs.
Status:
* Put back resource limit code
* suEXEC doesn't work
Status: Manoj has posted an patch to fix this.
<19991103003605.A20612@samosa.mindspring.com>
* Win32: Enable the Windows MPM to honor max_requests_per_child
Status: Bill will fix this.
* Win32: Get Apache working on Windows 95/98. The following work
(at least) needs to be done:
- winnt MPM: Fix 95/98 code paths in the winnt MPM. There is some NT
specific code that is still not in NT only code paths
- APR sendfile uses TransmitFile which is not available on
95/98.
- ap_stat uses GetFileAttributeEx which is not available on
95/98
Status:
* Win32: Test access logging with multiple threads. Will the
native file I/O calls serialize automagically like the
CRT calls or do we need to add region locking each time
we access the logs?
Status:
* The handler hook needs to use the new registration system
* We need a thread-safe resolver, at least on Unix.
Status: The best known candidate would be something from
BIND v9.
* Switch from HSRegex to PCRE. New code should be src/lib/pcre/
(ftp://cus.cam.ac.uk/pub/software/programs/pcre/)
Status: Greg +1
RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
* the top-level configure.in appears to use $USE_MAINTAINER_MODE = yes
to set -Wall/etc. flags, and the APR configure.in appears to use
--with-debug. both seem like more of a hardwired pain compaired
to the pre-2.0 method of setting EXTRA_CFLAGS.
* Win32: Reuse accept socket after transmitfile/close
This is not a bug, but would be nice to get this feature in
before ship.
* Win32: Enable the winnt MPM to use the new scoreboard API
* Win32: Graceful restart of a service is broken (code does shutdown then start)
* Dynamically loadable MPMs
Enable users of the binary distributions of Apache to be able to
dynamically load the MPM.
Dean says this probably isn't a good idea. For example consider
-lpthread, you couldn't have the same core run with pthreads,
prefork, and gnu pth without a lot of headaches is my guess.
* Platforms that do not support fork (primarily Win32 and AS/400)
Consider introducing HAVE_FORK feature macro. Architect start-up code
that avoids initializing all the modules in the parent process on
platforms that do not support fork.
* Clean the code. There are a lot of places we used APR but didn't
remove the hacks that were required for the cross-platform code in
1.3. We need to make the code look like APR was supposed to be there.
* Go throught the 1.3 Bug DB and research the bugs marked "suspended".
People were told these would be considered for inclusion in Apache 2.0,
it would be nice to actually do so.
* Win32: Migrate the MPM over to use APR thread/process calls. This
would eliminate some code in the Win32 branch that essentially
duplicates what is in APR.
Bill says we need a new procattr, APR_CREATE_SUSPENDED (or
something similar) to direct ap_create_process to create the
process suspended. We also need a call to wake up the suspended
process This may not be able to be implemented everywhere though.
Status:
* Move I/O layering into APR.
* There are still a number of places in the code where we are
loosing error status (i.e. throwing away the error returned by a
system call and replacing it with a generic error code)
* Win32: Implement reliable piped logs on Windows
Status:
* Switch to autoconf, et al. for configuration.
Status: Manoj has placed an initial version into the 2.0
repository. Todos include:
* Use APR to get rid of more platform dependancies.
Status: Ryan Bloom <rbb@covalenet.net> is working on this.
* The connection status table is not very efficient. Also, very few stats
are exported to the connection status table (easy to fix), and mod_status
is ugly.
* Optimise the name-based <VirtualHost> implementation so that it uses
a hash table instead of a linear linked list scan.
* Mass vhosting version of suEXEC.
* Provide a sane API for handling the request's environment variables.
* configuration option to use *DBM
Status: Greg +1 (volunteers)
* add SDBM into src/lib/sdbm/ as a default/fallback DBM implementation.
SDBM is used by Perl, mod_dav, mod_sssl, others for basic DBM support.
Status: Greg +1 (volunteers)
Other bugs that need fixing:
* MaxRequestsPerChild measures connections, not requests.
Until someone has a better way, we'll probably just rename it
"MaxConnectionsPerChild".
* Regex containers don't work in an intutive way
Status: No one has come up with an efficient way to fix this
behavior. Dean has suggested getting rid of regex containers
completely.
* SIGSEGV on Linux (glibc 2.1.2) isn't caught properly by a
sigwaiting thread. We need to work around this, perhaps unless
there is hope soon for a fixed glibc.
Other features that need writing:
* Finish infrastructure in core for async MPMs
Status: post 2.0
* TODO in source -- just do an egrep on "TODO" and see what's there
Documentation that needs writing:
* Mod_status docs are needed.
* The concept of MPMs, especially if we ship more than one MPM for a
given platform
* New directives in the various MPMs and appropriate links from
obsolete directives in core.html to the MPM documentation.
* Revise manual/stopping.html and the last part of
manual/misc/perf-tuning.html to take account of the MPMs.
* API documentation
Status: Ben Laurie has written some hooks documentation
(apache-2.0/htdocs/hooks.html)
* Changes since 1.3.9 can be more easily seen in the commitlog file
dev.apache.org:/home/cvs/CVSROOT/commitlogs/apache-2.0
which includes some of Roy's comments when the changes were
committed in rough change-sets by purpose. Note that the commitlog
does not show the contents of new files until later.
Available Patches:
* Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
performance
Status: These were written for 1.3, and are awaiting a port to
2.0
Open issues:
* What do we do about mod_proxy?
* Which MPMs will be included with Apache 2.0?
* Is conf/highperformance.conf-dist obsolete? It looks obsolete.
|
