2 Last modified at [$Date: 2000/02/25 03:44:44 $]
6 Vague plan, assuming more features aren't demanded for the 2.0 release
14 * MPMs should have responsibility for closing sockets and
15 registering the socket cleanup. Stop registering multiple
16 cleanups for each socket. Move socket close code out of
17 http_connection.c and into the MPMs.
20 * Reliable piped logs look broken everywhere. Each MPM includes essentially
21 identical code to ap_register_other_child(), etc. Most of this code can
22 be moved out of the MPMs and into some common file (http_core.c?).
24 * DSO hooks registered during the preflight call to ap_read_config
25 are not 'unregistered' during the preflight clean-up. Thus, when
26 Apache comes up, the hook tables contain two entries for each DSO hook,
27 one of which is bogus, leftover from the preflight.
29 * Pipes to CGI scripts are not being timed out
30 Status: code has been added to APR to support timing out pipes.
31 This needs to be used in Apache now.
33 * Put back resource limit code
36 Status: Paul Reder <rederpj@raleigh.ibm.com> is working on this.
39 Status: Manoj has posted an patch to fix this.
40 <19991103003605.A20612@samosa.mindspring.com>
42 * Win32: Enable the Windows MPM to honor max_requests_per_child
43 Status: Bill will fix this.
45 * Win32: Fix Win9* specific code in the winnt MPM
48 * Win32: Test access logging with multiple threads. Will the
49 native file I/O calls serialize automagically like the
50 CRT calls or do we need to add region locking each time
54 * Win32: Smoke test all ported modules to make sure they work
55 correctly under Windows.
58 * Current 2.0 code is not tested on many Unix platforms. Make 2.0
59 work on most, if not all the systems 1.3 did
60 Status: Autoconf and APR will get us most of the way there.
62 * The module API is a weird combination of the old table and the new
63 hook system. Switch completely to the new hook system
64 Status: Ben Laurie is working on this.
66 * Modules. Which modules do not work yet, can we get a list?
68 * We need a thread-safe resolver, at least on Unix.
69 Status: The best known candidate would be something from BIND
70 (v8 or v9?) The only other option would be to mutex all the
73 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
75 * Win32: Reuse accept socket after transmitfile/close
76 This is not a bug, but would be nice to get this feature in
79 * Win32: Enable the winnt MPM to use the new scoreboard API
81 * Win32: Non-blocking CGI pipes
83 * Win32: Graceful restart of a service is broken (code does shutdown then start)
85 * Dynamically loadable MPMs
86 Enable users of the binary distributions of Apache to be able to
87 dynamically load the MPM.
89 * Platforms that do not support fork (primarily Win32 and AS/400)
90 Consider introducing HAVE_FORK feature macro. Architect start-up code
91 that avoids initializing all the modules in the parent process on
92 platforms that do not support fork.
94 * Clean the code. There are a lot of places we used APR but didn't
95 remove the hacks that were required for the cross-platform code in
96 1.3. We need to make the code look like APR was supposed to be there.
98 * Go throught the Bug DB and research the bugs marked "suspended". People
99 were told these would be considered for inclusion in Apache 2.0, it
100 would be nice to actually do so.
102 * Win32: Migrate the MPM over to use APR thread/process calls. This
103 would eliminate some code in the Win32 branch that essentially
104 duplicates what is in APR.
105 Status: Bill <stoddard@raleigh.ibm.com> is working on this.
107 * Move I/O layering into APR.
109 * There are still a number of places in the code where we are
110 loosing error status (i.e. throwing away the error returned by a
111 system call and replacing it with a generic error code)
113 * Win32: Implement reliable piped logs on Windows
114 Status: Bill <stoddard@raleigh.ibm.com> has prototype code (not
115 reliable) in 1.3. Should be much cleaner with APR in 2.0.
117 * Switch to autoconf, et al. for configuration.
118 Status: Manoj has placed an initial version into the 2.0
119 repository. Todos include:
121 - "make install" should work
122 - a tool to simplfy third-party module building should be
123 written. Something like apxs
124 - Add a lot more checks to satisfy the various platforms, e.g.
127 * Use APR to get rid of more platform dependancies.
128 Status: Ryan Bloom <rbb@raleigh.ibm.com> is working on this.
130 * The connection status table is not very efficient. Also, very few stats
131 are exported to the connection status table (easy to fix), and mod_status
134 Other bugs that need fixing:
136 * MaxRequestsPerChild measures connections, not requests.
137 Until someone has a better way, we'll probably just rename it
138 "MaxConnectionsPerChild".
140 * Regex containers don't work in an intutive way
141 Status: No one has come up with an efficient way to fix this
142 behavior. Dean has suggested getting rid of regex containers
145 * SIGSEGV on Linux (glibc 2.1.2) isn't caught properly by a
146 sigwaiting thread. We need to work around this, perhaps unless
147 there is hope soon for a fixed glibc.
149 Other features that need writing:
151 * Finish infrastructure in core for async MPMs
154 * TODO in source -- just do an egrep on "TODO" and see what's there
156 * Work on other MPMs. Possible MPMs discussed or in progress include:
158 - Dean Gaudet's async MPM
159 Status: dean is unlikely to begin work on this any time soon...
160 has lots of ideas he'll share with folks willing to begin
161 tackling the issues. this is apache 2.1 material.
163 - Zach Brown's <zab@zabbo.net> Linux siginfo MPM
166 Documentation that needs writing:
168 * The concept of MPMs, especially if we ship more than one MPM for a
171 * New directives in the various MPMs
174 Status: Ben Laurie has written some hooks documentation
175 (apache-2.0/htdocs/hooks.html)
177 * Changes since 1.3.9 can be more easily seen in the commitlog file
178 dev.apache.org:/home/cvs/CVSROOT/commitlogs/apache-2.0
179 which includes some of Roy's comments when the changes were
180 committed in rough change-sets by purpose. Note that the commitlog
181 does not show the contents of new files until later.
185 * Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
187 Status: These were written for 1.3, and are awaiting a port to
192 * What do we do about mod_proxy?
194 * Which MPMs will be included with Apache 2.0?
196 * Is conf/highperformance.conf-dist obsolete? It looks obsolete.