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