3 *) Fix a bug in fcgi_config_set_authoritative_slot(). Tetsuya Furukawa
4 [tetsuya@secom-sis.co.jp]
6 *) Eliminate the use of locks to assist in the clean shutdown of
7 applications. Instead, it is assumed that applications handle
8 termination signals properly (this is now embedded in the C
11 *) Fix Win32 process termination. Proper operation requires the use of
12 an updated application lib (termination is now signalled with an
13 Event and handled by specialized thread).
15 *) Docs cleanup. Andrew Benham [adsb@bigfoot.com]
17 *) Added code so if the last instance of a dynamic application died without
18 provocation, then don't restart it if singleThreshold > 0 (i.e. if the
19 configuration allows the last instance to be killed, then allow it to die).
20 Andrew Benham [adsb@bigfoot.com]
22 *) Fix the loadFactor calculation used to determine when dyanmic
23 applications could be killed off due to low demand [adsb@bigfoot.com].
25 *) Fix a deadlock condition that could occur with Win32 named
28 *) Fix a potential deadlock condition when FastCGI application
29 sent responses while still reading the client request (POST data).
33 *) Allow absolute pathnames in the -socket argument. Suggested by
34 Christian Jaeger [christian.jaeger@sl.ethz.ch].
36 *) Don't invoke suexec when the user/group for the fastcgi application
37 is the same as the apache main server. This is consistent with
38 apache's suexec handling. Suggested by Nikolaus Rath [Nikolaus@rath.org].
40 *) Reset the apache drop dead timer upon successful read or writes
41 to/from the client. This eliminates timeouts that were occuring
42 during the large file transfers to/from slow clients.
44 *) Support generic wrappers such as cgiwrap by eliminating dependencies
45 on Apache's SUEXEC, renaming the FastCgiSuexec directive
46 FastCgiWrapper and eliminating any checks regarding the target
47 application (this is the repsonibility of the wrapper).
49 *) Fix a nasty bug that occurred when a client aborted a POST request
50 before the connection to a dynamic FastCGI application was opened.
51 The application's lock file descriptor wasn't setup but was being
52 closed which resulted in FD0 being closed. Normally this is open to
53 /dev/null and should pose no problems except that because the FD
54 was available it was being returned by Apache's accept(). This
55 caused it to be registered for pool cleanup. mod_cgi though moves
56 the CGI stdin pipe to FD0 and thus it was getting waxed during pool
57 cleanup. Problem identified by checksum@163.net.
59 Changes with mod_fastcgi 2.2.8
61 *) Eliminate the concept of disabled applications. If a failure
62 occurs trying to setup an application (e.g. bind() error) its
63 tried repeatedly every init-start-delay seconds.
65 *) Tweak to Makefile.tmpl to support DSOs. Dave Hill [ddhill@zk3.dec.com]
67 Changes with mod_fastcgi 2.2.6
69 *) Shutdown the PM when Apache appears to have disappeared.
71 *) seteuid() tweak for HP-UX 11. Milton L. Hankins [mlh@swl.msd.ray.com]
73 *) (Win32) More dynamic fixes.
75 *) (Win32) Eliminate the per application named pipe mutex. Its now
76 per process to keep the lib happy (the use _FCGI_MUTEX_ should
77 removed from the lib). This allows pipe-based applications to
78 now handle multiple simoultanous requests (one per process).
80 *) Increase the number of open FDs we look to close when spawning apps.
82 *) Prevent an assert from popping unnecessarily.
84 *) (Win32) Add support for interpretter scripts.
86 *) (Win32) Fix named pipe handling (problems with large responses).
88 *) (Win32) Remove the "can exec" check.
90 Changes with mod_fastcgi 2.2.4
92 *) Beta WinNT support. David Allen [djallen@raleigh.ibm.com] and
93 Rob Saccoccio [robs@chelsea.net]
95 *) Remove request type restriction (GET and POST only) in order to support
96 Web DAV requests. Scott Robertson [sroberts@codeit.com]
98 *) Allow requests for URLs such as /server/some/other/qualifier to match a
99 FastCGI server defined as /server. This was done primarily for Java
100 Servlets, but is generically useful.
102 *) Change the comm between the PM and the request handlers from a regular
103 file to a pipe and an assortment of other dynamic fixes.
105 *) Log dynamic process termination scheduling and the resulting
106 process exit notification.
108 *) Change the default singleThreshhold from 10 to 0 to prevent the last dynamic
109 application process from being killed off due to low demand.
111 *) Clean up FastCGI application pathnames (e.g. remove duplicate slashes).
113 *) Fix bugs that prevented dynamic processes from being shutdown when
114 the load subsided. Lars Heete [heete@do.isst.fhg.de]
116 *) Prevent dynamic processes from being scheduled to be started before
117 the init-start-delay or restart-delay period expires. This prevents
118 the queuing of process start requests while an application which has
119 a long initialization period starts up.
121 *) When suexec is enabled, allow the process manager to become root again in
122 order to signal applications it spawned.
124 *) Change the "server started" log message level from INFO to WARN.
126 *) Fix dynamic server "Connection Refused" handling.
128 *) Fix demand-based dynamic process spawing (uses a more portable approach).
130 *) Add -idle-timeout arg to FastCgiServer, FastCgiExternalServer, and
131 FastCgiConfig. mod_fastcgi will now abort a connection if inactive for
132 longer than this period. It applies to the initiation of connections as
133 well (and thus is similar to appConnTimeout). Default is 30 seconds.
135 *) appConnTimeout is 0 by default now resulting in blocking connect()s. This
136 is more platform portable/predicable.
138 *) Leave STDOUT and STDERR open to the main server error_log. This should
139 eliminate a pile of protocol errors and having to track down 3rd party
140 libs writing to stderr inadvertantly. This doesn't mean these shouldn't
141 be fixed in the application to use the FastCGI I/O, it just allows these
142 applications to run when previously they'd crash and burn.
144 *) Add the pass-header arg to server directives.
146 *) Allow supplementary groups to be passed to spawned FastCGI applications (as
147 is done for CGI) - at least until PR2580 is resolved.
149 *) Initialize the default (empty) environment in a more socially acceptable
152 *) Miscellaneous doc improvements based on notes I've collected up over the
155 *) Fix a bug in the stderr handling introduced in 2.2.2. There were conditions
156 that resulted in not or improperly terminated strings.
158 *) Fix the call to setsockopt() to disable Nagle. [based on a bug report by
159 Johannes Plassmann <johannes.plassmann@pdb.siemens.de>]
161 Changes with mod_fastcgi 2.2.2
163 *) Added support for blocking connect()s by setting appConnTimeout to 0.
164 Non-blocking connect()s (the default) can be troublesome on some platforms.
165 Its expected that blocking connect()s will become the default (and
166 non-blocking connect()s will become optional) in the next release.
168 *) Dump a compile time error if the version of Apache is too old.
170 *) Wrap the SIGPIPE handler manipulation code such that it is only applicable
171 to Apache releases prior to 1.3.6.
173 *) Minor tweaks for RUSSIAN_APACHE. [Sergey Gershtein <sg@mplik.ru>]
177 Always restart a failed dynamic application if it is the last instance.
178 This means there once started there will always be at least one process
179 instance of a dynamic application.
181 Send PLEASE_START to the PM when a connect() results in an ECONNREFUSED.
182 ECONNREFUSED means the listen queue is full (or there isn't one). Asking
183 the PM to start (another) application instance may help empty it faster.
185 Change two sleep() calls to select() based snoozes because alarm() is in
186 effect and sleep() and alarm() don't always play nice together.
188 Fixed a couple of error messages.
190 *) Fix -listen-queue-depth arg on FastCgiConfig (dynamic). Previously it
191 was ignored and the default was always used.
193 *) Allow the -initial-env argument to be used to pass variables from the
194 Apache process environment to the FastCGI server (by specifying a
195 variable name without the "=" or a value). [suggested by
196 Martin Lichtin <lichtin@oanda.com>]
198 *) Cleanup some debug macros.
200 *) Improved script stderr handling. [based on suggestions by
201 David Birnbaum <davidb@chelsea.net>]
203 *) Added IRIX and FreeBSD to the list of supported platforms and other minor
204 updates to the INSTALL doc.
206 *) Changed the default listen-queue-depth (FCGI_DEFAULT_LISTEN_Q) from
207 5 to 100. Its still configurable with the -listen-queue-depth option.
208 This should help eliminate the FAQ - Why do I see "Connection Refused"
211 *) Fix a bug in FastCgiExternalServer that broke support for external servers
212 on other hosts. [Dave Neuer <dneuer@futuristics.net>]
214 Changes with mod_fastcgi 2.2.1
216 *) Updates to the INSTALL doc to describe building as a DSO.
218 *) If the FastCgiIpcDir directive was in use and "httpd -t" was issued while
219 Apache was running ("apachectl restart" does this), the contents of the
220 dynamic directory (dynamic sockets and the mbox) would be blown away
223 *) Add an extern declaration for ap_sys_siglist (its exposed but not in any
224 Apache header file) to prevent a compile error on systems without
227 *) During auth requests, the subprocess_env table was left holding variables
228 sent to the auth FastCGI server (including REMOTE_PASSWD!) which means they
229 were passed to other processing phases (such as CGI/FastCGI). The
230 subprocess_env is now restored to its pre-auth condition.
232 *) Added a FastCGI Authorizer Role compatibility mode which implements the
233 specification to the tee. Use -compat arg with any of the Auth server
234 directives. This is intended for new and existing FastCGI authorizer
235 applications that require compatibility with other server implementations.
237 *) Fix logging from the process manager. Previously, all calls to the
238 logging routines used NULL for the server_rec. This works fine in a dev
239 environment, but results in most of these messages being tossed by
240 log_error_core() because DEFAULT_LOGLEVEL is too low. Odd logic.
242 *) Always setup fr->header so that in the event the FastCGI server
243 doesn't get a valid FastCGI protocol header over the wire, we
244 can still print our error message to the log from do_work().
246 *) Fix a compile error in open_connection_to_fs() on systems with TCP_NODELAY
247 defined (Don Locrasto [locrasto@iceminer.com]).
249 Changes with mod_fastcgi 2.2.0
251 *) Serious rewrite of mod_fastcgi.html. An example conf will have
252 to wait until another time - I'm burned out on docs updates.
254 *) Add to the mod_fastcgi to the server version string.
256 *) SCO doesn't like the const in the arg to inet_addr() and
259 *) Use Apache's NET_SIZE_T to get the correct arg type for getsockopt().
261 *) Deleted redundant header files from fcgi.h.
263 *) Full path names are no longer required for directives which take
264 file names as arguments. Like other Apache directives you can
265 specify paths from server_root (don't start with a "/").
267 *) The fcgiKillMgr is gone. This intermediate process sat between
268 the Apache parent and the FastCGI Process Manager under Apache
269 1.2. The process manager tries to name itself fcgi_pm (it used
272 *) Authentication, Authorization, and Access phases are now
273 supported. I've bent the FastCgi spec a bit: as many of the standard
274 environment variables (that were easy) are sent (the spec says don't
275 send some by name), all headers (except Status) sent by the auth
276 FastCgi server are passed to subprocesses (CGI/FastCGI invocations)
277 as environment variables rather than just those prefixed by
278 "Variable-". Custom responses for auth failures aren't supported
279 (yet.. we'll see what the demand for this is like). In addition to
280 the FastCGI protocol defined environment variable "FCGI_ROLE" being
281 set to "AUTHORIZER" an environment variable "FCGI_APACHE_ROLE" is
282 set indicating which of the three phases is being processed. See
283 the docs for info on the new per-directory directives:
285 FastCgiAuthenticator, FastCgiAuthenticatorAuthoritative,
286 FastCgiAuthorizer, FastCgiAuthorizerAuthoritative,
287 FastCgiAccessChecker, FastCgiAccessCheckerAuthoritative
289 *) The code has been broken into logical chunks making figuring it
290 out much easier. All of the #defines a user may want to change are
291 now in mod_fastcgi.h.
293 *) All the logging has been converted to the "supported" Apache
294 logging routines, ap_log_error() and ap_log_rerror(). Log entries
295 for request specific errors should now be directed to the correct
296 server errorlog (if your running more than one). Some attempt has
297 been made at setting appropriate log levels (comments of course are
298 welcome) and printing errno info when its of value. All of log
299 messages now have "FastCGI" in them (nice for grep).
301 *) All of the calls to Apache routines have been updated to the ap_
302 convention, eliminating the need for the compat header.
304 *) By default we no longer flush every piece of data we get from the
305 FastCGI server to the client. This allows the FastCGI server to
306 release without having to wait for the client flush to complete.
307 The old behaviour can revived with the -flush argument to AppClass.
309 *) I renamed some of the directives for consistency (the old names
312 AppClass -> FastCgiServer,
313 FCGIConfig -> FastCgiConfig,
314 ExternalAppClass -> FastCgiExternalServer
316 *) Directive arguments are no longer case sensitive.
318 *) The module now uses hard_timeout() rather than soft_timeout. This
319 means that the module will longjump out of the request if the
320 drop-dead timeout expires (set with Apache's Timeout directive, the
321 default is 5min) or if the client closes the connection (SIGPIPE).
322 This is more typical for a Apache modules and I'm not convinced we
323 properly handle all of the error cases well enough to use
324 soft_timeout (I can think of one place we don't anyway). This means
325 your FastCGI application can see SIGPIPE.
327 *) SIGPIPE is now ignored by default in FastCGI servers spawned by
328 Apache. Without this the default behaviour is exit().
330 *) It should (I think) run now under DSO now.
332 *) The bug which prevented sending of binary data in 2.1b1 is fixed.
334 *) The broken -initial-env bug ins 2.1b1 is fixed.
336 *) Maybe some other stuff.
340 *** Well there's a big gap here. Maybe I'll go back and extract the CVS
341 commit notes and stick 'em in here.. when I get some free time. ;)
346 *** Originally from docs/README..
348 Apache[X1.03.01]/mod_fastcgi[02.00.05] 19970909 unsupported
349 From: Stanley Gambarin <stanleyg@cs.bu.edu>
351 *) Yet more changes to the source distribution. Separated out
352 the Tcl dynamic string and buffer libraries into separate files.
353 Did the same for the OS library, however, it is not really an
354 abstraction... more like a bunch of wrappers. Added to DEVNOTES
355 a note about the problem of Apache not allowing including header
358 Apache[X1.03.01]/mod_fastcgi[02.00.04] 19970908 unsupported
359 From: Stanley Gambarin <stanleyg@cs.bu.edu>
361 *) More changes to the installation scripts. Theoretically we
362 should be able to to use the same Makefile/installation script
363 for both Apache 1.2.x and 1.3.x. Need some sort of the abstraction
364 layer to use the same source code for both 1.2.x and 1.3.x sources.
366 Apache[X1.03.01]/mod_fastcgi[02.00.03] 19970905 unsupported
367 From: Stanley Gambarin <stanleyg@cs.bu.edu>
369 *) Created Makefile to compile the module into the archive library,
370 just like the proxy module. Also initial draft of the installation
371 script, which should work for both 1.2 and 1.3 sources of Apache.
373 Apache[01.02.04]/mod_fastcgi[02.00.02] 19970903 unsupported
374 From: Stanley Gambarin <stanleyg@cs.bu.edu>
376 *) More source reorg. Separated out some header files. Need much
377 more work on this. Updated the TODO file. Added fcgivers.h and
378 CHANGES file to track the history.
380 Apache[01.02.04]/mod_fastcgi[02.00.01] 19970902 unsupported
381 From: Stanley Gambarin <stanleyg@cs.bu.edu>
383 *) Source reorganization. This is done to provide a basis for later
384 functionality implementation. See TODO file for more information
385 of what possible future enhancements are possible/desired.
387 Apache[01.02.04]/mod_fastcgi[02.00.00] 19970902 unsupported
388 From: David MacKenzie <djm@va.pubnix.com>
390 *) Create the "dynamic" dir and "mbox" as the user specified in the
391 User and Group directives instead of as root. They are created with
392 restrictive permissions, and then the module checks to see whether the
393 user specified in the User and Group directives has read, write, and
394 execute permissions on them. Those permissions had been explicitly
395 denied by creating them as "root".
397 *) Some documentation was garbled or incomplete.
399 *) When a FastCGI application can't be execl'd, the code used a
400 value of errno that may have been stomped on by intervening system
401 calls. That happened on BSD/OS 2.1, where the error_log was
402 reporting errno=25 (NOTTY) instead of the correct errno=13 (EACCES).
403 An intervening stdio call was setting errno as a side-effect, which
404 it has the right to do.
406 *) Erroneous fprintf arguments, a missing return value, an unused
407 function, missing declarations, and other problems detected by gcc
410 *) Duplicated code to create the lock file name merged into a single
411 function, closing several memory leaks.
413 *) FCGIConfig -minProcesses didn't allow a value of 0, so the last
414 FCGI app in a given dynamic class would live forever, even if it
415 hadn't been requested for weeks.
417 *) The test for whether to keep looking for dynamic app victims to
418 kill was backwards, so no dynamic apps were ever selected as
421 *) The killInterval and updateInterval were being ignored; the
422 sigsuspend() forced recalculations only when a child died,
423 instead of at the intervals specified.
425 *) When a dynamic app couldn't be started, the program SEGV'd when
426 trying to free the ipcAddrPtr twice.
428 *) Remove the dead lock file and socket when cleaning up after a
429 server whose last child has been killed, so the FastCGIHandler
430 isn't fooled into thinking there is still a process serving
433 *) Make file locking robust in the presence of signals.
435 *) Rename some badly named variables.
437 *) Fix typos in many comments.
439 *) Fix some memory and file descriptor leaks.
441 *) Make the blocking kill of a server closer to working.
443 *) More fixes to calculations in dynamic application management.
446 *** Originally from docs/README.OMI..
448 What's New: Version 2.0b1, 16 Apr 1997
450 *) Implemented a mechanism by which FastCGI applications are started by
451 the web server on the first request and continue running. Also, a
452 heuristics have been implemented to allow for dynamic killing of the
453 running FastCGI apps. The configuration options are supplied via
454 the new FCGIConfig directive, that is described in the
457 *) When performing internal redirect, since the original request's body
458 has already been read, do not allow the redirected request to think
459 that is has one [body].
461 *) More conditional compilation for OS/2.
463 *) Fixed occassional "Assertion failed: len > 0, file mod_fastcgi.c"
465 *) Fixed "failed assertion `count >= 0 && count <= bufPtr->length'"
467 *) Added bflush() call after bwrite() in DrainReqOutBuf function to
468 force some output to be written in the period of inactivity. This
469 circumvents the Apache's buffering during the server-push.
471 *) Added a header-parser function placeholder for compatibility with an
474 *) Additional checks have been placed with regard to permissions of the
475 FastCGI processes. These include the existence of the file,
476 ExecCGI and IncludesNOEXEC checks, disallowing of nph- scripts, etc.
478 *) Implemented a restart cleanup, so that no parentlesses FastCGI
479 applications are left after Apache is restarted and/or terminated.
480 This was accomplished via two independent mechanisms, where the
481 first one prevents a start of the process manager on the first
482 reading of the configuration files. The second mechanism deals away
483 from Apache's implementation of fork() (via spawn_child()) and
484 implements its own forking, thus removing any dependencies on the
485 Apache to cleanup processes during termination phase.
487 *) Removed the definition of the Sigfunc, as Apache 2.0b10 defines it
488 itself in the file conf.h
490 What's New: Version 1.4.3, 15 Jan 1997
492 *) Fixed compilation warnings for various platforms, as well as
493 conditional compilation for OS/2.
495 What's New: Version 1.4.2, 12 Dec 1996
497 *) mod_fastcgi.c is ported to Apache 1.2b. Any further development
498 will proceed under this version of Apache web server.
500 *) As the result of porting, the "include virtual" construct of SSI
501 will now work correctly using either <Location> or <Directory>
504 What's New: Version 1.4.1, 4 Dec 1996
506 *) Checks have been removed from the ScanCGIHeaders that provided
507 for the presence of both Status and Location headers as being
508 an error. Contradictory to CGI/1.1 Internet Draft, both of
509 these headers are used by the current CGI applications.
511 What's New: Version 1.4, 22 Nov 1996
513 *) Added the -port option to AppClass, allowing TCP/IP communication.
514 Added the -socket option to AppClass, allowing Unix domain
515 communication via a configurable pathname.
517 *) Added the ExternalAppClass directive, allowing TCP/IP
518 communication with remote FastCGI applications.
520 *) The handler had its own code for generating HTTP response headers;
521 now it uses Apache's. This reduces the size of the module.
522 More importantly, it fixes the bug in which "include virtual"
523 sees the HTTP response headers.
525 *) The response header parser performed very little checking.
526 Now the parser enforces the guidelines in the CGI/1.1
527 Internet-Draft: Status and Location are mutually exclusive,
528 Location can only be a response to GET or HEAD,
529 CGI response headers can't be repeated, etc. (The CGI response
530 headers are the ones the handler interprets: Status, Location,
533 *) The response header parser used to miss CGI headers
534 with no whitespace after the colon, e.g. "Status:200 OK".
536 *) The response header parser sometimes interpreted the first line
537 of content as an RFC822 continuation line.
539 *) The handler implemented a nonstandard version of Location
540 which never used internal redirects. The handler
541 now attempts to implement Location as specified in the
542 CGI/1.1 Internet-Draft. The module documentation explains
545 *) Error log entries from the response header parser were
546 pretty uninformative; they are better now. If a header is
547 malformed, the log entry includes it. If the headers are
548 unterminated rather than malformed, the log entry says that,
549 and says how many bytes were received from the app.
551 *) When the application manager forked a new process, and that
552 process ran into trouble before executing the first instruction
553 of application code, the process used to exit with status =
554 errno. This made certain configuration problems (e.g. incorrect
555 file permissions when server parent is root) quite difficult
556 to pin down. Now the failing child process opens up the error log
557 and writes an informative entry before exiting.
559 *) The module now correctly handles a slash at the end of the
560 DocumentRoot directive. This was a one-line fix.
562 What's New: Version 1.3.3, 17 Oct 1996
564 *) The module now registers its request handler under the name
565 fastcgi-script in addition to the name application/x-httpd-fcgi.
566 This was a one-line addition to the module, but had quite
567 a large impact on the documentation and sample configuration.
569 What's New: Version 1.3.2, 27 Sept 1996
571 *) On some systems (SunOS, Linux?), fopen for append has a bug
572 that strikes when two processes append to the same file.
573 This bug causes the process manager to corrupt the error log.
574 Work around the bug by calling open, then fdopen. A patch
575 to Apache 1.1.1 is also required, as described in mod_fastcgi.html.
576 (Reported by Bob Ramstad.)
578 What's New: Version 1.3.1, 17 Sept 1996
580 *) On Linux and SunOS, the Apache default user_id (-1) is not a
581 legal value of uid_t, so cast it. (Reported by Bob Ramstad,
582 Scott Langley, others.)
584 *) On some systems (Linux, some Solaris?, Irix?), connect requires
585 write access to a Unix Domain socket, so provide it. (Reported
586 by Scott Langley, freeform@wired.com, others.)
588 *) If you hit a .fcg file, but there's no AppClass defined, the
589 error should be NOT_FOUND instead of SERVER_ERROR. (Reported
592 What's New: Version 1.3, 4 Sept 1996
594 *) Module sometimes busy-waited in FastCgiDoWork. Fixed.
596 *) Module violated the Apache buff abstraction.
597 Fixed. Might now work with SSL (not tested).
598 As part of the fix, eliminated support for
599 Apache 1.0x versions.
601 *) Module failed to chmod the Unix domain listening sockets
602 it created, so protections were set according to the current
605 *) Module forked too many process manager processes: One
606 per AppClass. When the Apache parent ran as root, these
607 process manager processes ran as root. New process
608 manager is a single process (only started if AppClass is
609 used) and runs with same privileges as other children.
610 The new process manager doesn't do any polling, so
611 there's less system overhead than before.
613 *) AppClass insisted on getting at least two arguments. Fixed.
615 *) Module used setjmp/longjmp, causing compiler warnings
616 on some platforms (e.g. Linux.) The module no longer
619 *) Module created listening sockets in /tmp, where they were
620 sometimes wiped out by cleanup scripts. Added FastCgiIpcDir
621 directive to give control over the location of listening
624 *) Module wrote error log entries without a timestamp. Fixed.
626 *) AppClass directive wrote error messages to stderr in addition
627 to returning a char * message to the Apache core. Fixed.
629 What's New: Version 1.2, 3 June 1996
631 *) Ported from Apache-1.0-based code to Apache-1.1b2 internals by
634 Add version string: APACHEVERSION macro to mod_fastcgi.c
636 *) chown FastCGI socket to user_id and setuid to user_id for app
639 *) Modify GetFromStream() by having it call an OS dependent
640 function GetStreamSize(FILE *) which uses FILE internal data member.
641 Linux users might need to modify GetStreamSize().
643 What's New: Version 1.1, 10 May 1996
645 *) If you specify a non-existent executable in the AppClass directive,
646 or if the file exists and it does not have execute permission you
647 get a constant stream of error messages telling you that "program
648 terminated due to a signal".
650 *) The mod_fastcgi module should use the standard Apache error logging
651 facility instead of writing to stderr.
653 Version 1.0, 30 April 1996