2 Last modified at [$Date: 1999/11/30 00:10:23 $]
6 2.0: In pre-alpha development
10 * Vague goal of an alpha or beta release in 1999. Commit-then-review
15 * Complete HAVE_SENDFILE work. Need to add iovec head and tail pointers to
16 the iol sendfile API and make it work under Unix.
17 Status: Bill <stoddard@raleigh.ibm.com> is working on this
19 * Pipes to CGI scripts are not being timed out
20 Status: code has been added to APR to support timing out pipes.
21 This needs to be used in Apache now.
23 * Put back resource limit code
26 Status: Paul Reder <rederpj@raleigh.ibm.com> is working on this.
29 Status: Paul Reder <rederpj@raleigh.ibm.com> is working on this.
32 Status: Manoj has posted an patch to fix this.
33 <19991103003605.A20612@samosa.mindspring.com>
35 * Windows NT port isn't done
36 Status: Bill <stoddard@raleigh.ibm.com> is working on MPM and APR.
38 1. Add back ability to run Apache as a service
39 2. Fix Win9* specific code in the winnt MPM
40 3. Get the MPM working in multi process mode (one parent & one child)
41 4. Test access logging with multiple threads. Will the native file I/O
42 calls serialize automagically like the CRT calls or do we need to
43 add region locking each time we write to the access/error logs?
45 * Current 2.0 code is not tested on many Unix platforms. Make 2.0
46 work on most, if not all the systems 1.3 did
47 Status: Autoconf and APR will get us most of the way there.
49 * The module API is a weird combination of the old table and the new
50 hook system. Switch completely to the new hook system
51 Status: Ben Laurie is working on this.
53 * Modules. Which modules do not work yet, can we get a list?
55 RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP:
57 * Clean the code. There are a lot of places we used APR but didn't
58 remove the hacks that were required for the cross-platform code in
59 1.3. We need to make the code look like APR was supposed to be there.
61 * Go throught the Bug DB and research the bugs marked "suspended". People
62 were told these would be considered for inclusion in Apache 2.0, it
63 would be nice to actually do so.
65 * Win32: Migrate the MPM over to use APR thread/process calls. This would eliminate
66 some code in the Win32 branch that essentially duplicates what is in APR.
67 Status: Bill <stoddard@raleigh.ibm.com> is working on this.
69 * Move I/O layering into APR.
71 * There are still a number of places in the code where we are loosing error
72 status (i.e. throwing away the error returned by a system call and replacing
73 it with a generic error code)
75 * APRize the stat function. CRT stat() is about 30% slower on Windows than
76 the equivalent native Windows call. The APR call should return the Unix
77 style stat structure, just to keep it familier.
79 * Implement reliable piped logs on Windows
80 Status: Bill <stoddard@raleigh.ibm.com> has prototype code (not reliable) in 1.3.
81 Should be much cleaner with APR in 2.0.
83 * Switch to autoconf, et al. for configuration.
84 Status: Manoj has placed an initial version into the 2.0
85 repository. Todos include:
87 - Support selection of modules (including MPMs; only Dexter
88 works right now). Probably involved creating a directory for
90 - "make install" and "make depend" should work
91 - a tool to simplfy third-party module building should be
92 written. Something like apxs
93 - Add a lot more checks to satisfy the various platforms, e.g.
96 * Use APR to get rid of more platform dependancies.
97 Status: Ryan Bloom <rbb@raleigh.ibm.com> is working on this.
99 * The connection status table is not very efficient. Also, very few stats
100 are exported to the connection status table (easy to fix), and mod_status
103 Other bugs that need fixing:
105 * MaxRequestsPerChild measures connections, not requests.
106 Until someone has a better way, we'll probably just rename it
107 "MaxConnectionsPerChild".
109 * Regex containers don't work in an intutive way
110 Status: No one has come up with an efficient way to fix this
111 behavior. Dean has suggested getting rid of regex containers
114 * SIGSEGV on Linux seems to only kill a thread, not a whole process;
115 we need to work around this, probably by bouncing the signal to
116 the sigwait thread. But this will hurt debugability.
118 Other features that need writing:
120 * Finish infrastructure in core for async MPMs
123 * TODO in source -- just do an egrep on "TODO" and see what's there
125 * Work on other MPMs. Possible MPMs discussed or in progress include:
127 - Dean Gaudet's async MPM
130 - Zach Brown's <zab@zabbo.net> Linux siginfo MPM
133 Documentation that needs writing:
135 * The concept of MPMs, especially if we ship more than one MPM for a
138 * New directives in the various MPMs
141 Status: Ben Laurie has written some hooks documentation
142 (apache-2.0/htdocs/hooks.html)
144 * Changes since 1.3.9 can be more easily seen in the commitlog file
145 dev.apache.org:/home/cvs/CVSROOT/commitlogs/apache-2.0
146 which includes some of Roy's comments when the changes were
147 committed in rough change-sets by purpose. Note that the commitlog
148 does not show the contents of new files until later.
152 * Mike Abbott's <mja@trudge.engr.sgi.com> patches to improve
154 Status: These were written for 1.3, and are awaiting a port to
159 * What do we do about mod_proxy?
161 * Which MPMs will be included with Apache 2.0?
163 * Is conf/highperformance.conf-dist obsolete? It looks obsolete.