Replaced the former date parser with a rewrite. No more yacc/bison needed.

CHANGES

@@ -10,8 +10,8 @@ Daniel (11 September 2004)
 - Added parsedate.[ch] that contains a rewrite of the date parser currently
   provided by getdate.y. The new one is MUCH smaller and will allow us to run
   away from the yacc/bison jungle. It is also slightly lacking in features
-  compared to the old one, but it should still support parsing of all date
-  formats HTTP might involve (and a fair bunch of others).
+  compared to the old one, but it supports parsing of all date formats HTTP
+  involves (and a fair bunch of others).
 
 Daniel (10 September 2004)
 - As found out by Jonas Forsman, curl didn't allow -F to set Content-Type on

configure.ac

@@ -1282,22 +1282,6 @@ then
   USE_MANUAL="no";
 fi
 
-AC_PROG_YACC
-
-if test -z "$YACC"
-then
-  AC_MSG_CHECKING([if OK to build without bison/yacc])
-  dnl no yacc is a big deal if we have no pre-fixed getdate.y
-  if test -r "$srcdir/lib/getdate.c"
-  then
-    dnl all is well, we don't have to generate it!
-    AC_MSG_RESULT(yes)
-  else
-    AC_MSG_RESULT(no)
-    AC_MSG_ERROR([no yacc or bison found, can't build libcurl!])
-  fi
-fi
-
 dnl *************************************************************************
 dnl If the manual variable still is set, then we go with providing a built-in
 dnl manual

docs/libcurl/curl_getdate.3

@@ -4,35 +4,36 @@
 .\"
 .TH curl_getdate 3 "5 March 2001" "libcurl 7.0" "libcurl Manual"
 .SH NAME
-curl_getdate - Convert an date in a ASCII string to number of seconds since
-January 1, 1970
+curl_getdate - Convert an date string to number of seconds since January 1,
+1970
 .SH SYNOPSIS
 .B #include <curl/curl.h>
 .sp
-.BI "time_t curl_getdate(char *" datestring ", time_t *"now" );
+.BI "time_t curl_getdate(char *" datestring ", time_t *"now" );"
 .ad
 .SH DESCRIPTION
-This function returns the number of seconds since January 1st 1970, for the
-date and time that the
-.I datestring
-parameter specifies. The 
-.I now
-parameter is there and should hold the current time to allow the datestring to
-specify relative dates/times. Read further in the date string parser section
-below.
+This function returns the number of seconds since January 1st 1970 in the UTC
+time zone, for the date and time that the \fIdatestring\fP parameter
+specifies. The \fInow\fP parameter is not used, pass a NULL there.
+
+\fBNOTE:\fP This function was rewritten for the 7.12.2 release and this
+documentation covers the functionality of the new one. The new one is not
+feature-complete with the old one, but most of the formats supported by the
+new one was supported by the old too.
 .SH PARSING DATES AND TIMES
-A "date" is a string, possibly empty, containing many items separated by
-whitespace.  The whitespace may be omitted when no ambiguity arises.  The
-empty string means the beginning of today (i.e., midnight).  Order of the
-items is immaterial.  A date string may contain many flavors of items:
+A "date" is a string containing several items separated by whitespace. The
+order of the items is immaterial.  A date string may contain many flavors of
+items:
 .TP 0.8i
 .B calendar date items
-This can be specified in a number of different ways. Including 1970-09-17, 70-9-17, 70-09-17, 9/17/72, 24 September 1972, 24 Sept 72, 24 Sep 72, Sep 24, 1972, 24-sep-72, 24sep72.
-The year can also be omitted, for example: 9/17 or "sep 17".
+Can be specified several ways. Month names can only be three-letter
+abbrivations, numbers can be zero-prefixed and the year may use 2 or 4 digits.
+Examples: 06 Nov 1994, 06-Nov-94 and Nov-94 6.
 .TP
 .B time of the day items
-This string specifies the time on a given day. Syntax supported includes:
-18:19:0, 18:19, 6:19pm, 18:19-0500 (for specifying the time zone as well).
+This string specifies the time on a given day. You must specify it with 6
+digits with two colons: HH:MM:SS. To not include the time in a date string,
+will make the function assume 00:00:00. Example: 18:19:21.
 .TP
 .B time zone items
 Specifies international time zone. There are a few acronyms supported, but in
@@ -40,41 +41,52 @@ general you should instead use the specific relative time compared to
 UTC. Supported formats include: -1200, MST, +0100.
 .TP
 .B day of the week items
-Specifies a day of the week. If this is mentioned alone it means that day of
-the week in the future.
-
-Days of the week may be spelled out in full: `Sunday', `Monday', etc or they
-may be abbreviated to their first three letters, optionally followed by a
-period.  The special abbreviations `Tues' for `Tuesday', `Wednes' for
-`Wednesday' and `Thur' or `Thurs' for `Thursday' are also allowed.
-
-A number may precede a day of the week item to move forward supplementary
-weeks.  It is best used in expression like `third monday'.  In this context,
-`last DAY' or `next DAY' is also acceptable; they move one week before or
-after the day that DAY by itself would represent.
-.TP
-.B relative items
-A relative item adjusts a date (or the current date if none) forward or
-backward. Example syntax includes: "1 year", "1 year ago", "2 days", "4
-weeks".
-
-The string `tomorrow' is worth one day in the future (equivalent to `day'),
-the string `yesterday' is worth one day in the past (equivalent to `day ago').
+Specifies a day of the week. Days of the week may be spelled out in full:
+`Sunday', `Monday', etc or they may be abbreviated to their first three
+letters. This is usually not info that adds anything.
 .TP
 .B pure numbers
-If the decimal number is of the form YYYYMMDD and no other calendar date item
-appears before it in the date string, then YYYY is read as the year, MM as the
-month number and DD as the day of the month, for the specified calendar date.
+If a decimal number of the form YYYYMMDD appears, then YYYY is read as the
+year, MM as the month number and DD as the day of the month, for the specified
+calendar date.
 .PP
+.SH EXAMPLES
+.nf
+Sun, 06 Nov 1994 08:49:37 GMT
+Sunday, 06-Nov-94 08:49:37 GMT
+Sun Nov  6 08:49:37 1994
+06 Nov 1994 08:49:37 GMT
+06-Nov-94 08:49:37 GMT
+Nov  6 08:49:37 1994
+06 Nov 1994 08:49:37
+06-Nov-94 08:49:37
+1994 Nov 6 08:49:37
+GMT 08:49:37 06-Nov-94 Sunday
+94 6 Nov 08:49:37
+1994 Nov 6
+06-Nov-94
+Sun Nov 6 94
+1994.Nov.6
+Sun/Nov/6/94/GMT
+Sun, 06 Nov 1994 08:49:37 CET
+06 Nov 1994 08:49:37 EST
+Sun, 12 Sep 2004 15:05:58 -0700
+Sat, 11 Sep 2004 21:32:11 +0200
+20040912 15:05:58 -0700
+20040911 +0200
+.fi
+.SH STANDARDS
+This parser was written to handle date formats specified in RFC 822 (including
+the update in RFC 1123) using time zone name or time zone delta and RFC 850
+(obsoleted by RFC 1036) and ANSI C's asctime() format. These formats are the
+only ones RFC2616 says HTTP applications may use.
 .SH RETURN VALUE
-This function returns zero when it fails to parse the date string. Otherwise
-it returns the number of seconds as described.
-.SH AUTHORS
-Originally written by Steven M. Bellovin <smb@research.att.com> while at the
-University of North Carolina at Chapel Hill.  Later tweaked by a couple of
-people on Usenet.  Completely overhauled by Rich $alz <rsalz@bbn.com> and Jim
-Berets <jberets@bbn.com> in August, 1990.
+This function returns -1 when it fails to parse the date string. Otherwise it
+returns the number of seconds as described.
+.SH REWRITE
+The former version of this function was built with yacc and was not only very
+large, it was also never quite understood and it wasn't possible to build with
+non-GNU tools since only Bison could make it thread-safe!
 
-It has been modified extensively since imported to curl.
-.SH "SEE ALSO"
-.BR GNU date(1)
+The rewrite was done for 7.12.2. The new one is much smaller and use simpler
+code.

lib/Makefile.am

@@ -24,7 +24,7 @@ AUTOMAKE_OPTIONS = foreign nostdinc
 
 DSP = curllib.dsp
 
-EXTRA_DIST = getdate.y Makefile.b32 Makefile.m32	\
+EXTRA_DIST = Makefile.b32 Makefile.m32	\
   Makefile.vc6 Makefile.riscos libcurl.def $(DSP) curllib.dsw		\
   config-vms.h config-win32.h config-riscos.h config-mac.h config.h.in	\
   ca-bundle.crt README.encoding README.memoryleak README.ares		\
@@ -33,7 +33,7 @@ EXTRA_DIST = getdate.y Makefile.b32 Makefile.m32	\
   makefile.amiga Makefile.netware nwlib.c libcurl.imp	\
   msvcproj.head msvcproj.foot
 
-CLEANFILES = $(DSP) getdate.c
+CLEANFILES = $(DSP)
 
 lib_LTLIBRARIES = libcurl.la
 
@@ -93,12 +93,7 @@ libcurl_la_SOURCES = $(CSOURCES) $(HHEADERS)
 WIN32SOURCES = $(CSOURCES) libcurl.def
 WIN32HEADERS = $(HHEADERS) config-win32.h
 
-BUILT_SOURCES = $(srcdir)/getdate.c $(top_builddir)/lib/ca-bundle.h
-
-# Say $(srcdir), so GNU make does not report an ambiguity with the .y.c rule.
-$(srcdir)/getdate.c: getdate.y
-	cd $(srcdir) && $(YACC) $(YFLAGS) getdate.y && \
-	sed -e 's:YYSTYPE yylval;:YYSTYPE yylval = {0}; /* post-bison fix */:' < y.tab.c > getdate.c && rm -f y.tab.c
+BUILT_SOURCES = $(top_builddir)/lib/ca-bundle.h
 
 $(top_builddir)/lib/ca-bundle.h: Makefile.in Makefile
 	@if test -f $@; then \

lib/Makefile.inc

@@ -2,21 +2,18 @@
  
 CSOURCES = file.c timeval.c base64.c hostip.c progress.c formdata.c	\
   cookie.c http.c sendf.c ftp.c url.c dict.c if2ip.c speedcheck.c	\
-  getdate.c ldap.c ssluse.c version.c getenv.c escape.c mprintf.c	\
-  telnet.c netrc.c getinfo.c transfer.c strequal.c easy.c security.c	\
-  krb4.c memdebug.c http_chunks.c strtok.c connect.c llist.c hash.c	\
-  multi.c content_encoding.c share.c http_digest.c md5.c		\
-  http_negotiate.c http_ntlm.c inet_pton.c strtoofft.c strerror.c	\
-  hostares.c hostasyn.c hostip4.c hostip6.c hostsyn.c hostthre.c	\
-  inet_ntop.c
+  ldap.c ssluse.c version.c getenv.c escape.c mprintf.c telnet.c	\
+  netrc.c getinfo.c transfer.c strequal.c easy.c security.c krb4.c	\
+  memdebug.c http_chunks.c strtok.c connect.c llist.c hash.c multi.c	\
+  content_encoding.c share.c http_digest.c md5.c http_negotiate.c	\
+  http_ntlm.c inet_pton.c strtoofft.c strerror.c hostares.c hostasyn.c	\
+  hostip4.c hostip6.c hostsyn.c hostthre.c inet_ntop.c parsedate.c
 
 HHEADERS = arpa_telnet.h netrc.h file.h timeval.h base64.h hostip.h	\
   progress.h formdata.h cookie.h http.h sendf.h ftp.h url.h dict.h	\
-  if2ip.h speedcheck.h urldata.h getdate.h ldap.h ssluse.h escape.h	\
-  telnet.h getinfo.h strequal.h security.h krb4.h memdebug.h		\
-  inet_ntoa_r.h http_chunks.h strtok.h connect.h llist.h hash.h		\
-  content_encoding.h share.h md5.h http_digest.h http_negotiate.h	\
-  http_ntlm.h ca-bundle.h inet_pton.h strtoofft.h strerror.h		\
-  inet_ntop.h curlx.h memory.h setup.h transfer.h
-
-
+  if2ip.h speedcheck.h urldata.h ldap.h ssluse.h escape.h telnet.h	\
+  getinfo.h strequal.h security.h krb4.h memdebug.h inet_ntoa_r.h	\
+  http_chunks.h strtok.h connect.h llist.h hash.h content_encoding.h	\
+  share.h md5.h http_digest.h http_negotiate.h http_ntlm.h ca-bundle.h	\
+  inet_pton.h strtoofft.h strerror.h inet_ntop.h curlx.h memory.h	\
+  setup.h transfer.h

lib/getdate.h

-/*
-**  Originally written by Steven M. Bellovin <smb@research.att.com> while
-**  at the University of North Carolina at Chapel Hill.  Later tweaked by
-**  a couple of people on Usenet.  Completely overhauled by Rich $alz
-**  <rsalz@bbn.com> and Jim Berets <jberets@bbn.com> in August, 1990.
-**
-**  This code is in the public domain and has no copyright.
-*/
-
-# include "setup.h"
-
-#ifndef PARAMS
-# if defined PROTOTYPES || (defined __STDC__ && __STDC__)
-#  define PARAMS(Args) Args
-# else
-#  define PARAMS(Args) ()
-# endif
-#endif
-
-#ifdef vms
-# include <types.h>
-# include <time.h>
-#else
-# include <sys/types.h>
-# if TIME_WITH_SYS_TIME
-#  include <sys/time.h>
-#  include <time.h>
-# else
-#  if HAVE_SYS_TIME_H
-#   include <sys/time.h>
-#  else
-#   include <time.h>
-#  endif
-# endif
-#endif /* defined (vms) */
-
-time_t curl_getdate PARAMS ((const char *p, const time_t *now));

lib/parsedate.c

@@ -67,7 +67,7 @@
   Sun, 12 Sep 2004 15:05:58 -0700
   Sat, 11 Sep 2004 21:32:11 +0200
 
-  compact numerical date strings ISO846-style:
+  compact numerical date strings:
 
   20040912 15:05:58 -0700
   20040911 +0200