Re: cvs commit: apache-1.3/src/ap ap_base64.c
Eh. I am not quite sure why I only now see this one going in. And why it says 'disk full'. But I supped a comletely fresh copy of CVS and CVS certainly seems as correct as it should be. Dw. On 8 Aug 1999 [EMAIL PROTECTED] wrote: dirkx 99/08/08 09:12:58 Added: src/ap ap_base64.c Log: disk full Revision ChangesPath 1.1 apache-1.3/src/ap/ap_base64.c Index: ap_base64.c === ... /* * base64 encoder/decoder. Origianly part of main/util.c * but moved here so that support/ab and ap_sha1.c could * use it. This mean (re)moving the ap_palloc()s. And adding * ugly 'len' functions. Which is quite a nasty cost. */ ..
Re: cvs commit: apache-1.3/src/modules/standard mod_auth.c mod_auth_db.c mod_auth_dbm.c
On 2 Aug 1999 [EMAIL PROTECTED] wrote: src/include ap.h ap_md5.h ap_sha1.h src/main http_main.c src/modules/standard mod_auth.c mod_auth_db.c mod_auth_dbm.c Removed: src/include ap_checkpass.h I am struggling with the same now; moving a few base64 encoders into one ap_base64.c file. What is the cut-off point for moving all declarations into ap.h as opposed to adding a few individual ap_*.h file ? Or was the current ap_md5.h simply wrong to begin with ? And caused by its legacy start; as an import from elsewhere ? Seeing your change I kind of prefer the ap.h collation; but I wonder a bit, as the base64 and validate_passwd functions are only used in just one or two files. ap.h is not getting any smaller. Or is that a needless worry ? Dw. RCS file: /home/cvs/apache-1.3/src/include/ap.h,v retrieving revision 1.21 retrieving revision 1.22 diff -u -r1.21 -r1.22 --- ap.h1999/05/31 17:09:31 1.21 +++ ap.h1999/08/02 20:50:14 1.22 @@ -159,6 +159,7 @@ __attribute__((format(printf,3,4))); API_EXPORT(int) ap_vsnprintf(char *buf, size_t len, const char *format, va_list ap); +API_EXPORT(char *) ap_validate_password(const char *passwd, const char *hash); #ifdef __cplusplus }
Re: cvs commit: apache-1.3/src/support httpd.exp
Good stuff. And a usefull extra function in the API as well. Dw On 28 Jul 1999 [EMAIL PROTECTED] wrote: coar99/07/28 10:37:22 Modified:.STATUS src CHANGES src/include alloc.h src/main alloc.c src/modules/standard mod_headers.c mod_negotiation.c mod_rewrite.c src/support httpd.exp Log: Treat the Vary response header field specially; change the modules that touch it to use a new routine that only adds a token if it isn't already present. O(n^2) behaviour as Dean points out, but absent set/atom operations it seems a reasonable stopgap for 1.3.7. PR: 4118 (previously closed with 'use force-no-vary') Reviewed by:Ken Coar Revision ChangesPath 1.727 +1 -8 apache-1.3/STATUS Index: STATUS === RCS file: /home/cvs/apache-1.3/STATUS,v retrieving revision 1.726 retrieving revision 1.727 diff -u -r1.726 -r1.727 --- STATUS 1999/07/28 14:06:20 1.726 +++ STATUS 1999/07/28 17:37:05 1.727 @@ -1,5 +1,5 @@ 1.3 STATUS: - Last modified at [$Date: 1999/07/28 14:06:20 $] + Last modified at [$Date: 1999/07/28 17:37:05 $] Release: @@ -67,13 +67,6 @@ RELEASE SHOWSTOPPERS: -* The Vary header field stuff is still broken (multiple - entries occur, etc.). The result is that some browsers (AFAIK at least - MSIE) are horribly confused by the responses. - Status: It should be fixed before 1.3.7 went out. For details - how it should be done, please look at new-httpd mailing list - archive: Ken, Ralf and Roy have already found consensus in the - past there. RELEASE NON-SHOWSTOPPERS BUT WOULD BE REAL NICE TO WRAP THESE UP: 1.1400+7 -0 apache-1.3/src/CHANGES Index: CHANGES === RCS file: /home/cvs/apache-1.3/src/CHANGES,v retrieving revision 1.1399 retrieving revision 1.1400 diff -u -r1.1399 -r1.1400 --- CHANGES 1999/07/24 18:48:20 1.1399 +++ CHANGES 1999/07/28 17:37:09 1.1400 @@ -1,5 +1,12 @@ Changes with Apache 1.3.7 + *) Sanitise Vary values by not adding duplicate keywords. A + separate routine needs to be used to do this, so any module + that frobs Vary needs to be changed. The standard modules + have all been modified. This solution is somewhat inelegant, + but it does the job for now. PR#4118 (better fix than before) + [Ken Coar, Roy Fielding] + *) Link DSO's with gcc -shared instead of ld -Bshareable at least on Linux and FreeBSD for now. [Rasmus Lerdorf] 1.69 +4 -0 apache-1.3/src/include/alloc.h Index: alloc.h === RCS file: /home/cvs/apache-1.3/src/include/alloc.h,v retrieving revision 1.68 retrieving revision 1.69 diff -u -r1.68 -r1.69 --- alloc.h 1999/05/13 19:44:14 1.68 +++ alloc.h 1999/07/28 17:37:14 1.69 @@ -225,6 +225,10 @@ API_EXPORT(const char *) ap_table_get(const table *, const char *); API_EXPORT(void) ap_table_set(table *, const char *name, const char *val); API_EXPORT(void) ap_table_setn(table *, const char *name, const char *val); +API_EXPORT(void) ap_table_merge_unique_token(table *t, const char *key, +const char *val); +API_EXPORT(void) ap_table_mergen_unique_token(table *t, const char *key, + const char *val); API_EXPORT(void) ap_table_merge(table *, const char *name, const char *more_val); API_EXPORT(void) ap_table_mergen(table *, const char *name, const char *more_val); API_EXPORT(void) ap_table_unset(table *, const char *key); 1.114 +26 -0 apache-1.3/src/main/alloc.c Index: alloc.c === RCS file: /home/cvs/apache-1.3/src/main/alloc.c,v retrieving revision 1.113 retrieving revision 1.114 diff -u -r1.113 -r1.114 --- alloc.c 1999/05/25 15:32:54 1.113 +++ alloc.c 1999/07/28 17:37:16 1.114 @@ -1346,6 +1346,32 @@ } } +/* + * Merge an HTTP token into a table entry IFF it isn't already in there. + * (Intended primarily to avoid Vary: host, host.) + */ +API_EXPORT(void) ap_table_merge_unique_token(table *t, const char *key, +const char *val) +{ +const char *curval; + +curval = ap_table_get(t, key); +if ((curval == NULL) || (!ap_find_token(t-a.pool, curval, val))) { +
Re: cvs commit: apache-1.3 STATUS
Did anyone else test the IPv6 patch ? Preferably on non solaris / non freebsd ? Dw. On 7 Jul 1999 [EMAIL PROTECTED] wrote: coar99/07/07 04:05:16 Modified:.STATUS Log: Veto EAPI for 1.3.7 (too much confusion and controversy, esp. with the KEAPI alternative); defer it until the next release. Also note some platforms I can build. Revision ChangesPath 1.724 +8 -30 apache-1.3/STATUS Index: STATUS === RCS file: /home/cvs/apache-1.3/STATUS,v retrieving revision 1.723 retrieving revision 1.724 diff -u -r1.723 -r1.724 --- STATUS 1999/06/25 08:12:18 1.723 +++ STATUS 1999/07/07 11:05:15 1.724 @@ -1,9 +1,11 @@ 1.3 STATUS: - Last modified at [$Date: 1999/06/25 08:12:18 $] + Last modified at [$Date: 1999/07/07 11:05:15 $] Release: 1.3.7-dev: current. Ken volunteers to be release manager. +Proposed dates: roll on Mon, 26 July 1999, test, and release +on Fri, 30 July 1999. 1.3.6. Tagged and rolled on Mar. 22. Released and announced on 24th. 1.3.5: Not released. @@ -20,7 +22,7 @@ Platform Avail. Volunteer -- alpha-dec-osf3.0 no Sameer Parekh - alpha-dec-osf4.0 yes Lars Eilebrecht + alpha-dec-osf4.0 yes Lars Eilebrecht, Ken Coar armv4l-whatever-linux2yes Rasmus Lerdorf hppa1.1-hp-hpux no Rob Hartill i386-slackware-linux(a.out) no Sameer Parekh @@ -34,9 +36,10 @@ i686-pc-freebsd3.1no Ralf S. Engelschall i586-unknown-linux2 yes Ralf S. Engelschall, Lars Eilebrecht i686-unknown-linux2 yes Lars Eilebrecht + i686-whatever-linux2 yes Ken Coar i386-unknown-linux(ELF) no Aram Mirzadeh, Michael Douglass i386-unknown-netBSD-1.2.1 N/A Lars Eilebrecht - i386-unknown-netBSD-1.3.2 yes Lars Eilebrecht + i386-unknown-netBSD-1.3.2 yes Lars Eilebrecht i386-unknown-sco3 no Ben Laurie i386-unknown-sco5 no Ben Laurie i386-sni-svr4 yes Martin Kraemer @@ -105,33 +108,8 @@ Status: Jim +1, Mark +1, Dean +1, BenH +1, Randy +1 (please choose name other than hook) Doug +1 on concept (untested), Lars +1 on concept, - Martin +1 (untested), - - Ken: -1 for pre-2.0 if it will: a) force a new release of - mod_perl or mod_php in order to maintain compatibility OR - b) require a version bump to 1.4.0 and a beta cycle -Ralf: It doesn't force a new release of any module (just a -recompilation), because it's a pure _ADDITION_ to the API and -doesn't make anything incompatible. The point b) I still do -not sunderstand, sorry. - -* H now we've to solve another problem for the EAPI -* idea: the existance of KEAPI. That's not my problem any -* longer. Others have to make decisions now. Additionally the -* group as a whole is far away from a point where it's clear -* what should be done in the future and also far away from -* the point where it's fun. -* -* So I finally withdraw my proposed addition of EAPI for -* 1.3.7/1.4.0 and for 2.0 I've still no strong opinion how it -* should be done. I'll try to stay out of all this and keep -* being happy to have EAPI myself and don't have to worry -* about all those useless politics and endless ping-pong -* decisions. -* -* IMHO the ASF should really quickly propose the -* revitalization of RST's old ENOTOWN symbolic errno value -* for use by its developers... -- RSE + Martin +1 (untested), Ken -1 for 1.3.7 (too controversial, + esp. w/KEAPI offered as well) * Brian Havard's patch to remove dependency of mod_auth_dbm on mod_auth. (PR#2598)
Re: cvs commit: apache-apr/pthreads/src/main acceptlock.c http_accept.c
On 9 May 1999 [EMAIL PROTECTED] wrote: Partial solution for the graceful restart during accept locking problem. Make sure that each thread checks the shutdown flag before attempting to get the cross-process accept lock, so that if one thread notices that it's time to die, the rest will soon after. Revision ChangesPath 1.5 +7 -2 apache-apr/pthreads/src/include/acceptlock.h Good stuff ! Seems to make a/the difference. Dw