3 *) Prevent silly maxProcesses and processSlack combinations.
4 Dmitry Dorofeev [dima@yasp.com]
6 *) Properly handle the killing of idle processes when one takes a long time
7 to exit once signaled down (or the config is funky).
8 Dmitry Dorofeev [dima@yasp.com]
10 *) Always kill the youngest instance of an application. Suggested by
11 Dmitry Dorofeev [dima@yasp.com]
15 *) Delay the logging of write errors to the pm to account for shutdown/restart.
17 *) (Win32) An assortment of fixes.
19 *) Fix some broken casts that were likely the cause of an assert.
21 *) Win32. Eliminate forward slashes from the named pipe path name.
22 Gerald Richter [richter@ecos.de]
24 *) SIGUSR2 is no longer blocked in the process manager and the fastcgi
25 applications it spawns. [ryans@amazon.com]
27 *) Added support for the -flush argument to FastCgiConfig.
28 Eric Sit [esit@alum.mit.edu]
30 *) Change the "which call to module_init() is this" check to a more
31 reliable approach. Doru Petrescu [pdoru@kappa.ro]
33 *) Close the old pipe file descriptor in apache main on USR1/HUP
34 (elimnates a small leak). James E. Jurach Jr. jjurach@fundsxpress.com
36 *) Fix a bug in fcgi_config_set_authoritative_slot(). Tetsuya Furukawa
37 [tetsuya@secom-sis.co.jp]
39 *) Eliminate the use of locks to assist in the clean shutdown of
40 applications. Instead, it is assumed that applications handle
41 termination signals properly (this is now embedded in the C
44 *) Fix Win32 process termination. Proper operation requires the use of
45 an updated application lib (termination is now signalled with an
46 Event and handled by specialized thread).
48 *) Docs cleanup. Andrew Benham [adsb@bigfoot.com]
50 *) Added code so if the last instance of a dynamic application died without
51 provocation, then don't restart it if singleThreshold > 0 (i.e. if the
52 configuration allows the last instance to be killed, then allow it to die).
53 Andrew Benham [adsb@bigfoot.com]
55 *) Fix the loadFactor calculation used to determine when dyanmic
56 applications could be killed off due to low demand [adsb@bigfoot.com].
58 *) Fix a deadlock condition that could occur with Win32 named
61 *) Fix a potential deadlock condition when FastCGI application
62 sent responses while still reading the client request (POST data).
66 *) Allow absolute pathnames in the -socket argument. Suggested by
67 Christian Jaeger [christian.jaeger@sl.ethz.ch].
69 *) Don't invoke suexec when the user/group for the fastcgi application
70 is the same as the apache main server. This is consistent with
71 apache's suexec handling. Suggested by Nikolaus Rath [Nikolaus@rath.org].
73 *) Reset the apache drop dead timer upon successful read or writes
74 to/from the client. This eliminates timeouts that were occuring
75 during the large file transfers to/from slow clients.
77 *) Support generic wrappers such as cgiwrap by eliminating dependencies
78 on Apache's SUEXEC, renaming the FastCgiSuexec directive
79 FastCgiWrapper and eliminating any checks regarding the target
80 application (this is the repsonibility of the wrapper).
82 *) Fix a nasty bug that occurred when a client aborted a POST request
83 before the connection to a dynamic FastCGI application was opened.
84 The application's lock file descriptor wasn't setup but was being
85 closed which resulted in FD0 being closed. Normally this is open to
86 /dev/null and should pose no problems except that because the FD
87 was available it was being returned by Apache's accept(). This
88 caused it to be registered for pool cleanup. mod_cgi though moves
89 the CGI stdin pipe to FD0 and thus it was getting waxed during pool
90 cleanup. Problem identified by checksum@163.net.
92 Changes with mod_fastcgi 2.2.8
94 *) Eliminate the concept of disabled applications. If a failure
95 occurs trying to setup an application (e.g. bind() error) its
96 tried repeatedly every init-start-delay seconds.
98 *) Tweak to Makefile.tmpl to support DSOs. Dave Hill [ddhill@zk3.dec.com]
100 Changes with mod_fastcgi 2.2.6
102 *) Shutdown the PM when Apache appears to have disappeared.
104 *) seteuid() tweak for HP-UX 11. Milton L. Hankins [mlh@swl.msd.ray.com]
106 *) (Win32) More dynamic fixes.
108 *) (Win32) Eliminate the per application named pipe mutex. Its now
109 per process to keep the lib happy (the use _FCGI_MUTEX_ should
110 removed from the lib). This allows pipe-based applications to
111 now handle multiple simoultanous requests (one per process).
113 *) Increase the number of open FDs we look to close when spawning apps.
115 *) Prevent an assert from popping unnecessarily.
117 *) (Win32) Add support for interpretter scripts.
119 *) (Win32) Fix named pipe handling (problems with large responses).
121 *) (Win32) Remove the "can exec" check.
123 Changes with mod_fastcgi 2.2.4
125 *) Beta WinNT support. David Allen [djallen@raleigh.ibm.com] and
126 Rob Saccoccio [robs@chelsea.net]
128 *) Remove request type restriction (GET and POST only) in order to support
129 Web DAV requests. Scott Robertson [sroberts@codeit.com]
131 *) Allow requests for URLs such as /server/some/other/qualifier to match a
132 FastCGI server defined as /server. This was done primarily for Java
133 Servlets, but is generically useful.
135 *) Change the comm between the PM and the request handlers from a regular
136 file to a pipe and an assortment of other dynamic fixes.
138 *) Log dynamic process termination scheduling and the resulting
139 process exit notification.
141 *) Change the default singleThreshhold from 10 to 0 to prevent the last dynamic
142 application process from being killed off due to low demand.
144 *) Clean up FastCGI application pathnames (e.g. remove duplicate slashes).
146 *) Fix bugs that prevented dynamic processes from being shutdown when
147 the load subsided. Lars Heete [heete@do.isst.fhg.de]
149 *) Prevent dynamic processes from being scheduled to be started before
150 the init-start-delay or restart-delay period expires. This prevents
151 the queuing of process start requests while an application which has
152 a long initialization period starts up.
154 *) When suexec is enabled, allow the process manager to become root again in
155 order to signal applications it spawned.
157 *) Change the "server started" log message level from INFO to WARN.
159 *) Fix dynamic server "Connection Refused" handling.
161 *) Fix demand-based dynamic process spawing (uses a more portable approach).
163 *) Add -idle-timeout arg to FastCgiServer, FastCgiExternalServer, and
164 FastCgiConfig. mod_fastcgi will now abort a connection if inactive for
165 longer than this period. It applies to the initiation of connections as
166 well (and thus is similar to appConnTimeout). Default is 30 seconds.
168 *) appConnTimeout is 0 by default now resulting in blocking connect()s. This
169 is more platform portable/predicable.
171 *) Leave STDOUT and STDERR open to the main server error_log. This should
172 eliminate a pile of protocol errors and having to track down 3rd party
173 libs writing to stderr inadvertantly. This doesn't mean these shouldn't
174 be fixed in the application to use the FastCGI I/O, it just allows these
175 applications to run when previously they'd crash and burn.
177 *) Add the pass-header arg to server directives.
179 *) Allow supplementary groups to be passed to spawned FastCGI applications (as
180 is done for CGI) - at least until PR2580 is resolved.
182 *) Initialize the default (empty) environment in a more socially acceptable
185 *) Miscellaneous doc improvements based on notes I've collected up over the
188 *) Fix a bug in the stderr handling introduced in 2.2.2. There were conditions
189 that resulted in not or improperly terminated strings.
191 *) Fix the call to setsockopt() to disable Nagle. [based on a bug report by
192 Johannes Plassmann <johannes.plassmann@pdb.siemens.de>]
194 Changes with mod_fastcgi 2.2.2
196 *) Added support for blocking connect()s by setting appConnTimeout to 0.
197 Non-blocking connect()s (the default) can be troublesome on some platforms.
198 Its expected that blocking connect()s will become the default (and
199 non-blocking connect()s will become optional) in the next release.
201 *) Dump a compile time error if the version of Apache is too old.
203 *) Wrap the SIGPIPE handler manipulation code such that it is only applicable
204 to Apache releases prior to 1.3.6.
206 *) Minor tweaks for RUSSIAN_APACHE. [Sergey Gershtein <sg@mplik.ru>]
210 Always restart a failed dynamic application if it is the last instance.
211 This means there once started there will always be at least one process
212 instance of a dynamic application.
214 Send PLEASE_START to the PM when a connect() results in an ECONNREFUSED.
215 ECONNREFUSED means the listen queue is full (or there isn't one). Asking
216 the PM to start (another) application instance may help empty it faster.
218 Change two sleep() calls to select() based snoozes because alarm() is in
219 effect and sleep() and alarm() don't always play nice together.
221 Fixed a couple of error messages.
223 *) Fix -listen-queue-depth arg on FastCgiConfig (dynamic). Previously it
224 was ignored and the default was always used.
226 *) Allow the -initial-env argument to be used to pass variables from the
227 Apache process environment to the FastCGI server (by specifying a
228 variable name without the "=" or a value). [suggested by
229 Martin Lichtin <lichtin@oanda.com>]
231 *) Cleanup some debug macros.
233 *) Improved script stderr handling. [based on suggestions by
234 David Birnbaum <davidb@chelsea.net>]
236 *) Added IRIX and FreeBSD to the list of supported platforms and other minor
237 updates to the INSTALL doc.
239 *) Changed the default listen-queue-depth (FCGI_DEFAULT_LISTEN_Q) from
240 5 to 100. Its still configurable with the -listen-queue-depth option.
241 This should help eliminate the FAQ - Why do I see "Connection Refused"
244 *) Fix a bug in FastCgiExternalServer that broke support for external servers
245 on other hosts. [Dave Neuer <dneuer@futuristics.net>]
247 Changes with mod_fastcgi 2.2.1
249 *) Updates to the INSTALL doc to describe building as a DSO.
251 *) If the FastCgiIpcDir directive was in use and "httpd -t" was issued while
252 Apache was running ("apachectl restart" does this), the contents of the
253 dynamic directory (dynamic sockets and the mbox) would be blown away
256 *) Add an extern declaration for ap_sys_siglist (its exposed but not in any
257 Apache header file) to prevent a compile error on systems without
260 *) During auth requests, the subprocess_env table was left holding variables
261 sent to the auth FastCGI server (including REMOTE_PASSWD!) which means they
262 were passed to other processing phases (such as CGI/FastCGI). The
263 subprocess_env is now restored to its pre-auth condition.
265 *) Added a FastCGI Authorizer Role compatibility mode which implements the
266 specification to the tee. Use -compat arg with any of the Auth server
267 directives. This is intended for new and existing FastCGI authorizer
268 applications that require compatibility with other server implementations.
270 *) Fix logging from the process manager. Previously, all calls to the
271 logging routines used NULL for the server_rec. This works fine in a dev
272 environment, but results in most of these messages being tossed by
273 log_error_core() because DEFAULT_LOGLEVEL is too low. Odd logic.
275 *) Always setup fr->header so that in the event the FastCGI server
276 doesn't get a valid FastCGI protocol header over the wire, we
277 can still print our error message to the log from do_work().
279 *) Fix a compile error in open_connection_to_fs() on systems with TCP_NODELAY
280 defined (Don Locrasto [locrasto@iceminer.com]).
282 Changes with mod_fastcgi 2.2.0
284 *) Serious rewrite of mod_fastcgi.html. An example conf will have
285 to wait until another time - I'm burned out on docs updates.
287 *) Add to the mod_fastcgi to the server version string.
289 *) SCO doesn't like the const in the arg to inet_addr() and
292 *) Use Apache's NET_SIZE_T to get the correct arg type for getsockopt().
294 *) Deleted redundant header files from fcgi.h.
296 *) Full path names are no longer required for directives which take
297 file names as arguments. Like other Apache directives you can
298 specify paths from server_root (don't start with a "/").
300 *) The fcgiKillMgr is gone. This intermediate process sat between
301 the Apache parent and the FastCGI Process Manager under Apache
302 1.2. The process manager tries to name itself fcgi_pm (it used
305 *) Authentication, Authorization, and Access phases are now
306 supported. I've bent the FastCgi spec a bit: as many of the standard
307 environment variables (that were easy) are sent (the spec says don't
308 send some by name), all headers (except Status) sent by the auth
309 FastCgi server are passed to subprocesses (CGI/FastCGI invocations)
310 as environment variables rather than just those prefixed by
311 "Variable-". Custom responses for auth failures aren't supported
312 (yet.. we'll see what the demand for this is like). In addition to
313 the FastCGI protocol defined environment variable "FCGI_ROLE" being
314 set to "AUTHORIZER" an environment variable "FCGI_APACHE_ROLE" is
315 set indicating which of the three phases is being processed. See
316 the docs for info on the new per-directory directives:
318 FastCgiAuthenticator, FastCgiAuthenticatorAuthoritative,
319 FastCgiAuthorizer, FastCgiAuthorizerAuthoritative,
320 FastCgiAccessChecker, FastCgiAccessCheckerAuthoritative
322 *) The code has been broken into logical chunks making figuring it
323 out much easier. All of the #defines a user may want to change are
324 now in mod_fastcgi.h.
326 *) All the logging has been converted to the "supported" Apache
327 logging routines, ap_log_error() and ap_log_rerror(). Log entries
328 for request specific errors should now be directed to the correct
329 server errorlog (if your running more than one). Some attempt has
330 been made at setting appropriate log levels (comments of course are
331 welcome) and printing errno info when its of value. All of log
332 messages now have "FastCGI" in them (nice for grep).
334 *) All of the calls to Apache routines have been updated to the ap_
335 convention, eliminating the need for the compat header.
337 *) By default we no longer flush every piece of data we get from the
338 FastCGI server to the client. This allows the FastCGI server to
339 release without having to wait for the client flush to complete.
340 The old behaviour can revived with the -flush argument to AppClass.
342 *) I renamed some of the directives for consistency (the old names
345 AppClass -> FastCgiServer,
346 FCGIConfig -> FastCgiConfig,
347 ExternalAppClass -> FastCgiExternalServer
349 *) Directive arguments are no longer case sensitive.
351 *) The module now uses hard_timeout() rather than soft_timeout. This
352 means that the module will longjump out of the request if the
353 drop-dead timeout expires (set with Apache's Timeout directive, the
354 default is 5min) or if the client closes the connection (SIGPIPE).
355 This is more typical for a Apache modules and I'm not convinced we
356 properly handle all of the error cases well enough to use
357 soft_timeout (I can think of one place we don't anyway). This means
358 your FastCGI application can see SIGPIPE.
360 *) SIGPIPE is now ignored by default in FastCGI servers spawned by
361 Apache. Without this the default behaviour is exit().
363 *) It should (I think) run now under DSO now.
365 *) The bug which prevented sending of binary data in 2.1b1 is fixed.
367 *) The broken -initial-env bug ins 2.1b1 is fixed.
369 *) Maybe some other stuff.
373 *** Well there's a big gap here. Maybe I'll go back and extract the CVS
374 commit notes and stick 'em in here.. when I get some free time. ;)
379 *** Originally from docs/README..
381 Apache[X1.03.01]/mod_fastcgi[02.00.05] 19970909 unsupported
382 From: Stanley Gambarin <stanleyg@cs.bu.edu>
384 *) Yet more changes to the source distribution. Separated out
385 the Tcl dynamic string and buffer libraries into separate files.
386 Did the same for the OS library, however, it is not really an
387 abstraction... more like a bunch of wrappers. Added to DEVNOTES
388 a note about the problem of Apache not allowing including header
391 Apache[X1.03.01]/mod_fastcgi[02.00.04] 19970908 unsupported
392 From: Stanley Gambarin <stanleyg@cs.bu.edu>
394 *) More changes to the installation scripts. Theoretically we
395 should be able to to use the same Makefile/installation script
396 for both Apache 1.2.x and 1.3.x. Need some sort of the abstraction
397 layer to use the same source code for both 1.2.x and 1.3.x sources.
399 Apache[X1.03.01]/mod_fastcgi[02.00.03] 19970905 unsupported
400 From: Stanley Gambarin <stanleyg@cs.bu.edu>
402 *) Created Makefile to compile the module into the archive library,
403 just like the proxy module. Also initial draft of the installation
404 script, which should work for both 1.2 and 1.3 sources of Apache.
406 Apache[01.02.04]/mod_fastcgi[02.00.02] 19970903 unsupported
407 From: Stanley Gambarin <stanleyg@cs.bu.edu>
409 *) More source reorg. Separated out some header files. Need much
410 more work on this. Updated the TODO file. Added fcgivers.h and
411 CHANGES file to track the history.
413 Apache[01.02.04]/mod_fastcgi[02.00.01] 19970902 unsupported
414 From: Stanley Gambarin <stanleyg@cs.bu.edu>
416 *) Source reorganization. This is done to provide a basis for later
417 functionality implementation. See TODO file for more information
418 of what possible future enhancements are possible/desired.
420 Apache[01.02.04]/mod_fastcgi[02.00.00] 19970902 unsupported
421 From: David MacKenzie <djm@va.pubnix.com>
423 *) Create the "dynamic" dir and "mbox" as the user specified in the
424 User and Group directives instead of as root. They are created with
425 restrictive permissions, and then the module checks to see whether the
426 user specified in the User and Group directives has read, write, and
427 execute permissions on them. Those permissions had been explicitly
428 denied by creating them as "root".
430 *) Some documentation was garbled or incomplete.
432 *) When a FastCGI application can't be execl'd, the code used a
433 value of errno that may have been stomped on by intervening system
434 calls. That happened on BSD/OS 2.1, where the error_log was
435 reporting errno=25 (NOTTY) instead of the correct errno=13 (EACCES).
436 An intervening stdio call was setting errno as a side-effect, which
437 it has the right to do.
439 *) Erroneous fprintf arguments, a missing return value, an unused
440 function, missing declarations, and other problems detected by gcc
443 *) Duplicated code to create the lock file name merged into a single
444 function, closing several memory leaks.
446 *) FCGIConfig -minProcesses didn't allow a value of 0, so the last
447 FCGI app in a given dynamic class would live forever, even if it
448 hadn't been requested for weeks.
450 *) The test for whether to keep looking for dynamic app victims to
451 kill was backwards, so no dynamic apps were ever selected as
454 *) The killInterval and updateInterval were being ignored; the
455 sigsuspend() forced recalculations only when a child died,
456 instead of at the intervals specified.
458 *) When a dynamic app couldn't be started, the program SEGV'd when
459 trying to free the ipcAddrPtr twice.
461 *) Remove the dead lock file and socket when cleaning up after a
462 server whose last child has been killed, so the FastCGIHandler
463 isn't fooled into thinking there is still a process serving
466 *) Make file locking robust in the presence of signals.
468 *) Rename some badly named variables.
470 *) Fix typos in many comments.
472 *) Fix some memory and file descriptor leaks.
474 *) Make the blocking kill of a server closer to working.
476 *) More fixes to calculations in dynamic application management.
479 *** Originally from docs/README.OMI..
481 What's New: Version 2.0b1, 16 Apr 1997
483 *) Implemented a mechanism by which FastCGI applications are started by
484 the web server on the first request and continue running. Also, a
485 heuristics have been implemented to allow for dynamic killing of the
486 running FastCGI apps. The configuration options are supplied via
487 the new FCGIConfig directive, that is described in the
490 *) When performing internal redirect, since the original request's body
491 has already been read, do not allow the redirected request to think
492 that is has one [body].
494 *) More conditional compilation for OS/2.
496 *) Fixed occassional "Assertion failed: len > 0, file mod_fastcgi.c"
498 *) Fixed "failed assertion `count >= 0 && count <= bufPtr->length'"
500 *) Added bflush() call after bwrite() in DrainReqOutBuf function to
501 force some output to be written in the period of inactivity. This
502 circumvents the Apache's buffering during the server-push.
504 *) Added a header-parser function placeholder for compatibility with an
507 *) Additional checks have been placed with regard to permissions of the
508 FastCGI processes. These include the existence of the file,
509 ExecCGI and IncludesNOEXEC checks, disallowing of nph- scripts, etc.
511 *) Implemented a restart cleanup, so that no parentlesses FastCGI
512 applications are left after Apache is restarted and/or terminated.
513 This was accomplished via two independent mechanisms, where the
514 first one prevents a start of the process manager on the first
515 reading of the configuration files. The second mechanism deals away
516 from Apache's implementation of fork() (via spawn_child()) and
517 implements its own forking, thus removing any dependencies on the
518 Apache to cleanup processes during termination phase.
520 *) Removed the definition of the Sigfunc, as Apache 2.0b10 defines it
521 itself in the file conf.h
523 What's New: Version 1.4.3, 15 Jan 1997
525 *) Fixed compilation warnings for various platforms, as well as
526 conditional compilation for OS/2.
528 What's New: Version 1.4.2, 12 Dec 1996
530 *) mod_fastcgi.c is ported to Apache 1.2b. Any further development
531 will proceed under this version of Apache web server.
533 *) As the result of porting, the "include virtual" construct of SSI
534 will now work correctly using either <Location> or <Directory>
537 What's New: Version 1.4.1, 4 Dec 1996
539 *) Checks have been removed from the ScanCGIHeaders that provided
540 for the presence of both Status and Location headers as being
541 an error. Contradictory to CGI/1.1 Internet Draft, both of
542 these headers are used by the current CGI applications.
544 What's New: Version 1.4, 22 Nov 1996
546 *) Added the -port option to AppClass, allowing TCP/IP communication.
547 Added the -socket option to AppClass, allowing Unix domain
548 communication via a configurable pathname.
550 *) Added the ExternalAppClass directive, allowing TCP/IP
551 communication with remote FastCGI applications.
553 *) The handler had its own code for generating HTTP response headers;
554 now it uses Apache's. This reduces the size of the module.
555 More importantly, it fixes the bug in which "include virtual"
556 sees the HTTP response headers.
558 *) The response header parser performed very little checking.
559 Now the parser enforces the guidelines in the CGI/1.1
560 Internet-Draft: Status and Location are mutually exclusive,
561 Location can only be a response to GET or HEAD,
562 CGI response headers can't be repeated, etc. (The CGI response
563 headers are the ones the handler interprets: Status, Location,
566 *) The response header parser used to miss CGI headers
567 with no whitespace after the colon, e.g. "Status:200 OK".
569 *) The response header parser sometimes interpreted the first line
570 of content as an RFC822 continuation line.
572 *) The handler implemented a nonstandard version of Location
573 which never used internal redirects. The handler
574 now attempts to implement Location as specified in the
575 CGI/1.1 Internet-Draft. The module documentation explains
578 *) Error log entries from the response header parser were
579 pretty uninformative; they are better now. If a header is
580 malformed, the log entry includes it. If the headers are
581 unterminated rather than malformed, the log entry says that,
582 and says how many bytes were received from the app.
584 *) When the application manager forked a new process, and that
585 process ran into trouble before executing the first instruction
586 of application code, the process used to exit with status =
587 errno. This made certain configuration problems (e.g. incorrect
588 file permissions when server parent is root) quite difficult
589 to pin down. Now the failing child process opens up the error log
590 and writes an informative entry before exiting.
592 *) The module now correctly handles a slash at the end of the
593 DocumentRoot directive. This was a one-line fix.
595 What's New: Version 1.3.3, 17 Oct 1996
597 *) The module now registers its request handler under the name
598 fastcgi-script in addition to the name application/x-httpd-fcgi.
599 This was a one-line addition to the module, but had quite
600 a large impact on the documentation and sample configuration.
602 What's New: Version 1.3.2, 27 Sept 1996
604 *) On some systems (SunOS, Linux?), fopen for append has a bug
605 that strikes when two processes append to the same file.
606 This bug causes the process manager to corrupt the error log.
607 Work around the bug by calling open, then fdopen. A patch
608 to Apache 1.1.1 is also required, as described in mod_fastcgi.html.
609 (Reported by Bob Ramstad.)
611 What's New: Version 1.3.1, 17 Sept 1996
613 *) On Linux and SunOS, the Apache default user_id (-1) is not a
614 legal value of uid_t, so cast it. (Reported by Bob Ramstad,
615 Scott Langley, others.)
617 *) On some systems (Linux, some Solaris?, Irix?), connect requires
618 write access to a Unix Domain socket, so provide it. (Reported
619 by Scott Langley, freeform@wired.com, others.)
621 *) If you hit a .fcg file, but there's no AppClass defined, the
622 error should be NOT_FOUND instead of SERVER_ERROR. (Reported
625 What's New: Version 1.3, 4 Sept 1996
627 *) Module sometimes busy-waited in FastCgiDoWork. Fixed.
629 *) Module violated the Apache buff abstraction.
630 Fixed. Might now work with SSL (not tested).
631 As part of the fix, eliminated support for
632 Apache 1.0x versions.
634 *) Module failed to chmod the Unix domain listening sockets
635 it created, so protections were set according to the current
638 *) Module forked too many process manager processes: One
639 per AppClass. When the Apache parent ran as root, these
640 process manager processes ran as root. New process
641 manager is a single process (only started if AppClass is
642 used) and runs with same privileges as other children.
643 The new process manager doesn't do any polling, so
644 there's less system overhead than before.
646 *) AppClass insisted on getting at least two arguments. Fixed.
648 *) Module used setjmp/longjmp, causing compiler warnings
649 on some platforms (e.g. Linux.) The module no longer
652 *) Module created listening sockets in /tmp, where they were
653 sometimes wiped out by cleanup scripts. Added FastCgiIpcDir
654 directive to give control over the location of listening
657 *) Module wrote error log entries without a timestamp. Fixed.
659 *) AppClass directive wrote error messages to stderr in addition
660 to returning a char * message to the Apache core. Fixed.
662 What's New: Version 1.2, 3 June 1996
664 *) Ported from Apache-1.0-based code to Apache-1.1b2 internals by
667 Add version string: APACHEVERSION macro to mod_fastcgi.c
669 *) chown FastCGI socket to user_id and setuid to user_id for app
672 *) Modify GetFromStream() by having it call an OS dependent
673 function GetStreamSize(FILE *) which uses FILE internal data member.
674 Linux users might need to modify GetStreamSize().
676 What's New: Version 1.1, 10 May 1996
678 *) If you specify a non-existent executable in the AppClass directive,
679 or if the file exists and it does not have execute permission you
680 get a constant stream of error messages telling you that "program
681 terminated due to a signal".
683 *) The mod_fastcgi module should use the standard Apache error logging
684 facility instead of writing to stderr.
686 Version 1.0, 30 April 1996