3 *) [WIN32] Don't set the OVERLAPPED_IO flag on NamedPipe listen HANDLEs -
4 setting it was just plain broken.
6 *) [WIN32] Fix the accept mutex - all applications were sharing one!?
8 *) Fix 'FastCgiConfig -autoUpdate'.
10 *) Fix 'FastCgiConfig -flush'.
12 *) Prevent silly maxProcesses and processSlack combinations.
13 Dmitry Dorofeev [dima@yasp.com]
15 *) Properly handle the killing of idle processes when one takes a long time
16 to exit once signaled down (or the config is funky).
17 Dmitry Dorofeev [dima@yasp.com]
19 *) Always kill the youngest instance of an application. Suggested by
20 Dmitry Dorofeev [dima@yasp.com]
24 *) Delay the logging of write errors to the pm to account for shutdown/restart.
26 *) (Win32) An assortment of fixes.
28 *) Fix some broken casts that were likely the cause of an assert.
30 *) Win32. Eliminate forward slashes from the named pipe path name.
31 Gerald Richter [richter@ecos.de]
33 *) SIGUSR2 is no longer blocked in the process manager and the fastcgi
34 applications it spawns. [ryans@amazon.com]
36 *) Added support for the -flush argument to FastCgiConfig.
37 Eric Sit [esit@alum.mit.edu]
39 *) Change the "which call to module_init() is this" check to a more
40 reliable approach. Doru Petrescu [pdoru@kappa.ro]
42 *) Close the old pipe file descriptor in apache main on USR1/HUP
43 (elimnates a small leak). James E. Jurach Jr. jjurach@fundsxpress.com
45 *) Fix a bug in fcgi_config_set_authoritative_slot(). Tetsuya Furukawa
46 [tetsuya@secom-sis.co.jp]
48 *) Eliminate the use of locks to assist in the clean shutdown of
49 applications. Instead, it is assumed that applications handle
50 termination signals properly (this is now embedded in the C
53 *) Fix Win32 process termination. Proper operation requires the use of
54 an updated application lib (termination is now signalled with an
55 Event and handled by specialized thread).
57 *) Docs cleanup. Andrew Benham [adsb@bigfoot.com]
59 *) Added code so if the last instance of a dynamic application died without
60 provocation, then don't restart it if singleThreshold > 0 (i.e. if the
61 configuration allows the last instance to be killed, then allow it to die).
62 Andrew Benham [adsb@bigfoot.com]
64 *) Fix the loadFactor calculation used to determine when dyanmic
65 applications could be killed off due to low demand [adsb@bigfoot.com].
67 *) Fix a deadlock condition that could occur with Win32 named
70 *) Fix a potential deadlock condition when FastCGI application
71 sent responses while still reading the client request (POST data).
75 *) Allow absolute pathnames in the -socket argument. Suggested by
76 Christian Jaeger [christian.jaeger@sl.ethz.ch].
78 *) Don't invoke suexec when the user/group for the fastcgi application
79 is the same as the apache main server. This is consistent with
80 apache's suexec handling. Suggested by Nikolaus Rath [Nikolaus@rath.org].
82 *) Reset the apache drop dead timer upon successful read or writes
83 to/from the client. This eliminates timeouts that were occuring
84 during the large file transfers to/from slow clients.
86 *) Support generic wrappers such as cgiwrap by eliminating dependencies
87 on Apache's SUEXEC, renaming the FastCgiSuexec directive
88 FastCgiWrapper and eliminating any checks regarding the target
89 application (this is the repsonibility of the wrapper).
91 *) Fix a nasty bug that occurred when a client aborted a POST request
92 before the connection to a dynamic FastCGI application was opened.
93 The application's lock file descriptor wasn't setup but was being
94 closed which resulted in FD0 being closed. Normally this is open to
95 /dev/null and should pose no problems except that because the FD
96 was available it was being returned by Apache's accept(). This
97 caused it to be registered for pool cleanup. mod_cgi though moves
98 the CGI stdin pipe to FD0 and thus it was getting waxed during pool
99 cleanup. Problem identified by checksum@163.net.
101 Changes with mod_fastcgi 2.2.8
103 *) Eliminate the concept of disabled applications. If a failure
104 occurs trying to setup an application (e.g. bind() error) its
105 tried repeatedly every init-start-delay seconds.
107 *) Tweak to Makefile.tmpl to support DSOs. Dave Hill [ddhill@zk3.dec.com]
109 Changes with mod_fastcgi 2.2.6
111 *) Shutdown the PM when Apache appears to have disappeared.
113 *) seteuid() tweak for HP-UX 11. Milton L. Hankins [mlh@swl.msd.ray.com]
115 *) (Win32) More dynamic fixes.
117 *) (Win32) Eliminate the per application named pipe mutex. Its now
118 per process to keep the lib happy (the use _FCGI_MUTEX_ should
119 removed from the lib). This allows pipe-based applications to
120 now handle multiple simoultanous requests (one per process).
122 *) Increase the number of open FDs we look to close when spawning apps.
124 *) Prevent an assert from popping unnecessarily.
126 *) (Win32) Add support for interpretter scripts.
128 *) (Win32) Fix named pipe handling (problems with large responses).
130 *) (Win32) Remove the "can exec" check.
132 Changes with mod_fastcgi 2.2.4
134 *) Beta WinNT support. David Allen [djallen@raleigh.ibm.com] and
135 Rob Saccoccio [robs@chelsea.net]
137 *) Remove request type restriction (GET and POST only) in order to support
138 Web DAV requests. Scott Robertson [sroberts@codeit.com]
140 *) Allow requests for URLs such as /server/some/other/qualifier to match a
141 FastCGI server defined as /server. This was done primarily for Java
142 Servlets, but is generically useful.
144 *) Change the comm between the PM and the request handlers from a regular
145 file to a pipe and an assortment of other dynamic fixes.
147 *) Log dynamic process termination scheduling and the resulting
148 process exit notification.
150 *) Change the default singleThreshhold from 10 to 0 to prevent the last dynamic
151 application process from being killed off due to low demand.
153 *) Clean up FastCGI application pathnames (e.g. remove duplicate slashes).
155 *) Fix bugs that prevented dynamic processes from being shutdown when
156 the load subsided. Lars Heete [heete@do.isst.fhg.de]
158 *) Prevent dynamic processes from being scheduled to be started before
159 the init-start-delay or restart-delay period expires. This prevents
160 the queuing of process start requests while an application which has
161 a long initialization period starts up.
163 *) When suexec is enabled, allow the process manager to become root again in
164 order to signal applications it spawned.
166 *) Change the "server started" log message level from INFO to WARN.
168 *) Fix dynamic server "Connection Refused" handling.
170 *) Fix demand-based dynamic process spawing (uses a more portable approach).
172 *) Add -idle-timeout arg to FastCgiServer, FastCgiExternalServer, and
173 FastCgiConfig. mod_fastcgi will now abort a connection if inactive for
174 longer than this period. It applies to the initiation of connections as
175 well (and thus is similar to appConnTimeout). Default is 30 seconds.
177 *) appConnTimeout is 0 by default now resulting in blocking connect()s. This
178 is more platform portable/predicable.
180 *) Leave STDOUT and STDERR open to the main server error_log. This should
181 eliminate a pile of protocol errors and having to track down 3rd party
182 libs writing to stderr inadvertantly. This doesn't mean these shouldn't
183 be fixed in the application to use the FastCGI I/O, it just allows these
184 applications to run when previously they'd crash and burn.
186 *) Add the pass-header arg to server directives.
188 *) Allow supplementary groups to be passed to spawned FastCGI applications (as
189 is done for CGI) - at least until PR2580 is resolved.
191 *) Initialize the default (empty) environment in a more socially acceptable
194 *) Miscellaneous doc improvements based on notes I've collected up over the
197 *) Fix a bug in the stderr handling introduced in 2.2.2. There were conditions
198 that resulted in not or improperly terminated strings.
200 *) Fix the call to setsockopt() to disable Nagle. [based on a bug report by
201 Johannes Plassmann <johannes.plassmann@pdb.siemens.de>]
203 Changes with mod_fastcgi 2.2.2
205 *) Added support for blocking connect()s by setting appConnTimeout to 0.
206 Non-blocking connect()s (the default) can be troublesome on some platforms.
207 Its expected that blocking connect()s will become the default (and
208 non-blocking connect()s will become optional) in the next release.
210 *) Dump a compile time error if the version of Apache is too old.
212 *) Wrap the SIGPIPE handler manipulation code such that it is only applicable
213 to Apache releases prior to 1.3.6.
215 *) Minor tweaks for RUSSIAN_APACHE. [Sergey Gershtein <sg@mplik.ru>]
219 Always restart a failed dynamic application if it is the last instance.
220 This means there once started there will always be at least one process
221 instance of a dynamic application.
223 Send PLEASE_START to the PM when a connect() results in an ECONNREFUSED.
224 ECONNREFUSED means the listen queue is full (or there isn't one). Asking
225 the PM to start (another) application instance may help empty it faster.
227 Change two sleep() calls to select() based snoozes because alarm() is in
228 effect and sleep() and alarm() don't always play nice together.
230 Fixed a couple of error messages.
232 *) Fix -listen-queue-depth arg on FastCgiConfig (dynamic). Previously it
233 was ignored and the default was always used.
235 *) Allow the -initial-env argument to be used to pass variables from the
236 Apache process environment to the FastCGI server (by specifying a
237 variable name without the "=" or a value). [suggested by
238 Martin Lichtin <lichtin@oanda.com>]
240 *) Cleanup some debug macros.
242 *) Improved script stderr handling. [based on suggestions by
243 David Birnbaum <davidb@chelsea.net>]
245 *) Added IRIX and FreeBSD to the list of supported platforms and other minor
246 updates to the INSTALL doc.
248 *) Changed the default listen-queue-depth (FCGI_DEFAULT_LISTEN_Q) from
249 5 to 100. Its still configurable with the -listen-queue-depth option.
250 This should help eliminate the FAQ - Why do I see "Connection Refused"
253 *) Fix a bug in FastCgiExternalServer that broke support for external servers
254 on other hosts. [Dave Neuer <dneuer@futuristics.net>]
256 Changes with mod_fastcgi 2.2.1
258 *) Updates to the INSTALL doc to describe building as a DSO.
260 *) If the FastCgiIpcDir directive was in use and "httpd -t" was issued while
261 Apache was running ("apachectl restart" does this), the contents of the
262 dynamic directory (dynamic sockets and the mbox) would be blown away
265 *) Add an extern declaration for ap_sys_siglist (its exposed but not in any
266 Apache header file) to prevent a compile error on systems without
269 *) During auth requests, the subprocess_env table was left holding variables
270 sent to the auth FastCGI server (including REMOTE_PASSWD!) which means they
271 were passed to other processing phases (such as CGI/FastCGI). The
272 subprocess_env is now restored to its pre-auth condition.
274 *) Added a FastCGI Authorizer Role compatibility mode which implements the
275 specification to the tee. Use -compat arg with any of the Auth server
276 directives. This is intended for new and existing FastCGI authorizer
277 applications that require compatibility with other server implementations.
279 *) Fix logging from the process manager. Previously, all calls to the
280 logging routines used NULL for the server_rec. This works fine in a dev
281 environment, but results in most of these messages being tossed by
282 log_error_core() because DEFAULT_LOGLEVEL is too low. Odd logic.
284 *) Always setup fr->header so that in the event the FastCGI server
285 doesn't get a valid FastCGI protocol header over the wire, we
286 can still print our error message to the log from do_work().
288 *) Fix a compile error in open_connection_to_fs() on systems with TCP_NODELAY
289 defined (Don Locrasto [locrasto@iceminer.com]).
291 Changes with mod_fastcgi 2.2.0
293 *) Serious rewrite of mod_fastcgi.html. An example conf will have
294 to wait until another time - I'm burned out on docs updates.
296 *) Add to the mod_fastcgi to the server version string.
298 *) SCO doesn't like the const in the arg to inet_addr() and
301 *) Use Apache's NET_SIZE_T to get the correct arg type for getsockopt().
303 *) Deleted redundant header files from fcgi.h.
305 *) Full path names are no longer required for directives which take
306 file names as arguments. Like other Apache directives you can
307 specify paths from server_root (don't start with a "/").
309 *) The fcgiKillMgr is gone. This intermediate process sat between
310 the Apache parent and the FastCGI Process Manager under Apache
311 1.2. The process manager tries to name itself fcgi_pm (it used
314 *) Authentication, Authorization, and Access phases are now
315 supported. I've bent the FastCgi spec a bit: as many of the standard
316 environment variables (that were easy) are sent (the spec says don't
317 send some by name), all headers (except Status) sent by the auth
318 FastCgi server are passed to subprocesses (CGI/FastCGI invocations)
319 as environment variables rather than just those prefixed by
320 "Variable-". Custom responses for auth failures aren't supported
321 (yet.. we'll see what the demand for this is like). In addition to
322 the FastCGI protocol defined environment variable "FCGI_ROLE" being
323 set to "AUTHORIZER" an environment variable "FCGI_APACHE_ROLE" is
324 set indicating which of the three phases is being processed. See
325 the docs for info on the new per-directory directives:
327 FastCgiAuthenticator, FastCgiAuthenticatorAuthoritative,
328 FastCgiAuthorizer, FastCgiAuthorizerAuthoritative,
329 FastCgiAccessChecker, FastCgiAccessCheckerAuthoritative
331 *) The code has been broken into logical chunks making figuring it
332 out much easier. All of the #defines a user may want to change are
333 now in mod_fastcgi.h.
335 *) All the logging has been converted to the "supported" Apache
336 logging routines, ap_log_error() and ap_log_rerror(). Log entries
337 for request specific errors should now be directed to the correct
338 server errorlog (if your running more than one). Some attempt has
339 been made at setting appropriate log levels (comments of course are
340 welcome) and printing errno info when its of value. All of log
341 messages now have "FastCGI" in them (nice for grep).
343 *) All of the calls to Apache routines have been updated to the ap_
344 convention, eliminating the need for the compat header.
346 *) By default we no longer flush every piece of data we get from the
347 FastCGI server to the client. This allows the FastCGI server to
348 release without having to wait for the client flush to complete.
349 The old behaviour can revived with the -flush argument to AppClass.
351 *) I renamed some of the directives for consistency (the old names
354 AppClass -> FastCgiServer,
355 FCGIConfig -> FastCgiConfig,
356 ExternalAppClass -> FastCgiExternalServer
358 *) Directive arguments are no longer case sensitive.
360 *) The module now uses hard_timeout() rather than soft_timeout. This
361 means that the module will longjump out of the request if the
362 drop-dead timeout expires (set with Apache's Timeout directive, the
363 default is 5min) or if the client closes the connection (SIGPIPE).
364 This is more typical for a Apache modules and I'm not convinced we
365 properly handle all of the error cases well enough to use
366 soft_timeout (I can think of one place we don't anyway). This means
367 your FastCGI application can see SIGPIPE.
369 *) SIGPIPE is now ignored by default in FastCGI servers spawned by
370 Apache. Without this the default behaviour is exit().
372 *) It should (I think) run now under DSO now.
374 *) The bug which prevented sending of binary data in 2.1b1 is fixed.
376 *) The broken -initial-env bug ins 2.1b1 is fixed.
378 *) Maybe some other stuff.
382 *** Well there's a big gap here. Maybe I'll go back and extract the CVS
383 commit notes and stick 'em in here.. when I get some free time. ;)
388 *** Originally from docs/README..
390 Apache[X1.03.01]/mod_fastcgi[02.00.05] 19970909 unsupported
391 From: Stanley Gambarin <stanleyg@cs.bu.edu>
393 *) Yet more changes to the source distribution. Separated out
394 the Tcl dynamic string and buffer libraries into separate files.
395 Did the same for the OS library, however, it is not really an
396 abstraction... more like a bunch of wrappers. Added to DEVNOTES
397 a note about the problem of Apache not allowing including header
400 Apache[X1.03.01]/mod_fastcgi[02.00.04] 19970908 unsupported
401 From: Stanley Gambarin <stanleyg@cs.bu.edu>
403 *) More changes to the installation scripts. Theoretically we
404 should be able to to use the same Makefile/installation script
405 for both Apache 1.2.x and 1.3.x. Need some sort of the abstraction
406 layer to use the same source code for both 1.2.x and 1.3.x sources.
408 Apache[X1.03.01]/mod_fastcgi[02.00.03] 19970905 unsupported
409 From: Stanley Gambarin <stanleyg@cs.bu.edu>
411 *) Created Makefile to compile the module into the archive library,
412 just like the proxy module. Also initial draft of the installation
413 script, which should work for both 1.2 and 1.3 sources of Apache.
415 Apache[01.02.04]/mod_fastcgi[02.00.02] 19970903 unsupported
416 From: Stanley Gambarin <stanleyg@cs.bu.edu>
418 *) More source reorg. Separated out some header files. Need much
419 more work on this. Updated the TODO file. Added fcgivers.h and
420 CHANGES file to track the history.
422 Apache[01.02.04]/mod_fastcgi[02.00.01] 19970902 unsupported
423 From: Stanley Gambarin <stanleyg@cs.bu.edu>
425 *) Source reorganization. This is done to provide a basis for later
426 functionality implementation. See TODO file for more information
427 of what possible future enhancements are possible/desired.
429 Apache[01.02.04]/mod_fastcgi[02.00.00] 19970902 unsupported
430 From: David MacKenzie <djm@va.pubnix.com>
432 *) Create the "dynamic" dir and "mbox" as the user specified in the
433 User and Group directives instead of as root. They are created with
434 restrictive permissions, and then the module checks to see whether the
435 user specified in the User and Group directives has read, write, and
436 execute permissions on them. Those permissions had been explicitly
437 denied by creating them as "root".
439 *) Some documentation was garbled or incomplete.
441 *) When a FastCGI application can't be execl'd, the code used a
442 value of errno that may have been stomped on by intervening system
443 calls. That happened on BSD/OS 2.1, where the error_log was
444 reporting errno=25 (NOTTY) instead of the correct errno=13 (EACCES).
445 An intervening stdio call was setting errno as a side-effect, which
446 it has the right to do.
448 *) Erroneous fprintf arguments, a missing return value, an unused
449 function, missing declarations, and other problems detected by gcc
452 *) Duplicated code to create the lock file name merged into a single
453 function, closing several memory leaks.
455 *) FCGIConfig -minProcesses didn't allow a value of 0, so the last
456 FCGI app in a given dynamic class would live forever, even if it
457 hadn't been requested for weeks.
459 *) The test for whether to keep looking for dynamic app victims to
460 kill was backwards, so no dynamic apps were ever selected as
463 *) The killInterval and updateInterval were being ignored; the
464 sigsuspend() forced recalculations only when a child died,
465 instead of at the intervals specified.
467 *) When a dynamic app couldn't be started, the program SEGV'd when
468 trying to free the ipcAddrPtr twice.
470 *) Remove the dead lock file and socket when cleaning up after a
471 server whose last child has been killed, so the FastCGIHandler
472 isn't fooled into thinking there is still a process serving
475 *) Make file locking robust in the presence of signals.
477 *) Rename some badly named variables.
479 *) Fix typos in many comments.
481 *) Fix some memory and file descriptor leaks.
483 *) Make the blocking kill of a server closer to working.
485 *) More fixes to calculations in dynamic application management.
488 *** Originally from docs/README.OMI..
490 What's New: Version 2.0b1, 16 Apr 1997
492 *) Implemented a mechanism by which FastCGI applications are started by
493 the web server on the first request and continue running. Also, a
494 heuristics have been implemented to allow for dynamic killing of the
495 running FastCGI apps. The configuration options are supplied via
496 the new FCGIConfig directive, that is described in the
499 *) When performing internal redirect, since the original request's body
500 has already been read, do not allow the redirected request to think
501 that is has one [body].
503 *) More conditional compilation for OS/2.
505 *) Fixed occassional "Assertion failed: len > 0, file mod_fastcgi.c"
507 *) Fixed "failed assertion `count >= 0 && count <= bufPtr->length'"
509 *) Added bflush() call after bwrite() in DrainReqOutBuf function to
510 force some output to be written in the period of inactivity. This
511 circumvents the Apache's buffering during the server-push.
513 *) Added a header-parser function placeholder for compatibility with an
516 *) Additional checks have been placed with regard to permissions of the
517 FastCGI processes. These include the existence of the file,
518 ExecCGI and IncludesNOEXEC checks, disallowing of nph- scripts, etc.
520 *) Implemented a restart cleanup, so that no parentlesses FastCGI
521 applications are left after Apache is restarted and/or terminated.
522 This was accomplished via two independent mechanisms, where the
523 first one prevents a start of the process manager on the first
524 reading of the configuration files. The second mechanism deals away
525 from Apache's implementation of fork() (via spawn_child()) and
526 implements its own forking, thus removing any dependencies on the
527 Apache to cleanup processes during termination phase.
529 *) Removed the definition of the Sigfunc, as Apache 2.0b10 defines it
530 itself in the file conf.h
532 What's New: Version 1.4.3, 15 Jan 1997
534 *) Fixed compilation warnings for various platforms, as well as
535 conditional compilation for OS/2.
537 What's New: Version 1.4.2, 12 Dec 1996
539 *) mod_fastcgi.c is ported to Apache 1.2b. Any further development
540 will proceed under this version of Apache web server.
542 *) As the result of porting, the "include virtual" construct of SSI
543 will now work correctly using either <Location> or <Directory>
546 What's New: Version 1.4.1, 4 Dec 1996
548 *) Checks have been removed from the ScanCGIHeaders that provided
549 for the presence of both Status and Location headers as being
550 an error. Contradictory to CGI/1.1 Internet Draft, both of
551 these headers are used by the current CGI applications.
553 What's New: Version 1.4, 22 Nov 1996
555 *) Added the -port option to AppClass, allowing TCP/IP communication.
556 Added the -socket option to AppClass, allowing Unix domain
557 communication via a configurable pathname.
559 *) Added the ExternalAppClass directive, allowing TCP/IP
560 communication with remote FastCGI applications.
562 *) The handler had its own code for generating HTTP response headers;
563 now it uses Apache's. This reduces the size of the module.
564 More importantly, it fixes the bug in which "include virtual"
565 sees the HTTP response headers.
567 *) The response header parser performed very little checking.
568 Now the parser enforces the guidelines in the CGI/1.1
569 Internet-Draft: Status and Location are mutually exclusive,
570 Location can only be a response to GET or HEAD,
571 CGI response headers can't be repeated, etc. (The CGI response
572 headers are the ones the handler interprets: Status, Location,
575 *) The response header parser used to miss CGI headers
576 with no whitespace after the colon, e.g. "Status:200 OK".
578 *) The response header parser sometimes interpreted the first line
579 of content as an RFC822 continuation line.
581 *) The handler implemented a nonstandard version of Location
582 which never used internal redirects. The handler
583 now attempts to implement Location as specified in the
584 CGI/1.1 Internet-Draft. The module documentation explains
587 *) Error log entries from the response header parser were
588 pretty uninformative; they are better now. If a header is
589 malformed, the log entry includes it. If the headers are
590 unterminated rather than malformed, the log entry says that,
591 and says how many bytes were received from the app.
593 *) When the application manager forked a new process, and that
594 process ran into trouble before executing the first instruction
595 of application code, the process used to exit with status =
596 errno. This made certain configuration problems (e.g. incorrect
597 file permissions when server parent is root) quite difficult
598 to pin down. Now the failing child process opens up the error log
599 and writes an informative entry before exiting.
601 *) The module now correctly handles a slash at the end of the
602 DocumentRoot directive. This was a one-line fix.
604 What's New: Version 1.3.3, 17 Oct 1996
606 *) The module now registers its request handler under the name
607 fastcgi-script in addition to the name application/x-httpd-fcgi.
608 This was a one-line addition to the module, but had quite
609 a large impact on the documentation and sample configuration.
611 What's New: Version 1.3.2, 27 Sept 1996
613 *) On some systems (SunOS, Linux?), fopen for append has a bug
614 that strikes when two processes append to the same file.
615 This bug causes the process manager to corrupt the error log.
616 Work around the bug by calling open, then fdopen. A patch
617 to Apache 1.1.1 is also required, as described in mod_fastcgi.html.
618 (Reported by Bob Ramstad.)
620 What's New: Version 1.3.1, 17 Sept 1996
622 *) On Linux and SunOS, the Apache default user_id (-1) is not a
623 legal value of uid_t, so cast it. (Reported by Bob Ramstad,
624 Scott Langley, others.)
626 *) On some systems (Linux, some Solaris?, Irix?), connect requires
627 write access to a Unix Domain socket, so provide it. (Reported
628 by Scott Langley, freeform@wired.com, others.)
630 *) If you hit a .fcg file, but there's no AppClass defined, the
631 error should be NOT_FOUND instead of SERVER_ERROR. (Reported
634 What's New: Version 1.3, 4 Sept 1996
636 *) Module sometimes busy-waited in FastCgiDoWork. Fixed.
638 *) Module violated the Apache buff abstraction.
639 Fixed. Might now work with SSL (not tested).
640 As part of the fix, eliminated support for
641 Apache 1.0x versions.
643 *) Module failed to chmod the Unix domain listening sockets
644 it created, so protections were set according to the current
647 *) Module forked too many process manager processes: One
648 per AppClass. When the Apache parent ran as root, these
649 process manager processes ran as root. New process
650 manager is a single process (only started if AppClass is
651 used) and runs with same privileges as other children.
652 The new process manager doesn't do any polling, so
653 there's less system overhead than before.
655 *) AppClass insisted on getting at least two arguments. Fixed.
657 *) Module used setjmp/longjmp, causing compiler warnings
658 on some platforms (e.g. Linux.) The module no longer
661 *) Module created listening sockets in /tmp, where they were
662 sometimes wiped out by cleanup scripts. Added FastCgiIpcDir
663 directive to give control over the location of listening
666 *) Module wrote error log entries without a timestamp. Fixed.
668 *) AppClass directive wrote error messages to stderr in addition
669 to returning a char * message to the Apache core. Fixed.
671 What's New: Version 1.2, 3 June 1996
673 *) Ported from Apache-1.0-based code to Apache-1.1b2 internals by
676 Add version string: APACHEVERSION macro to mod_fastcgi.c
678 *) chown FastCGI socket to user_id and setuid to user_id for app
681 *) Modify GetFromStream() by having it call an OS dependent
682 function GetStreamSize(FILE *) which uses FILE internal data member.
683 Linux users might need to modify GetStreamSize().
685 What's New: Version 1.1, 10 May 1996
687 *) If you specify a non-existent executable in the AppClass directive,
688 or if the file exists and it does not have execute permission you
689 get a constant stream of error messages telling you that "program
690 terminated due to a signal".
692 *) The mod_fastcgi module should use the standard Apache error logging
693 facility instead of writing to stderr.
695 Version 1.0, 30 April 1996