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