path: root/www-misc/mergelog/files/mergelog-4.5-splitlog.patch

Index: mergelog-4.5/configure.in
===================================================================
--- mergelog-4.5.orig/configure.in
+++ mergelog-4.5/configure.in
@@ -2,7 +2,7 @@ dnl Process this file with autoconf to p
 AC_INIT(src/mergelog.c)
 
 PACKAGE=mergelog
-VERSION=4.5
+VERSION=4.5-split
 AM_INIT_AUTOMAKE($PACKAGE,$VERSION,nosubst)
 
 AC_ARG_PROGRAM
Index: mergelog-4.5/man/mergelog.1
===================================================================
--- mergelog-4.5.orig/man/mergelog.1
+++ mergelog-4.5/man/mergelog.1
@@ -1,8 +1,9 @@
-.TH MERGELOG 1 "22 Jan 2001"
+.TH MERGELOG 1 "20 Jun 2001"
 .SH NAME
 mergelog \- a fast tool to merge http log files by date
 .SH SYNOPSIS
 .B mergelog
+.IR [-vh]\ [-o\ outfmt]
 .IR logfile1
 .IR logfile2 \ ...
 .SH DESCRIPTION
Index: mergelog-4.5/man/zmergelog.1
===================================================================
--- mergelog-4.5.orig/man/zmergelog.1
+++ mergelog-4.5/man/zmergelog.1
@@ -1,8 +1,9 @@
-.TH ZMERGELOG 1 "22 Jan 2001"
+.TH ZMERGELOG 1 "20 Jun 2001"
 .SH NAME
 zmergelog \- a fast tool to merge gzipped http log files by date
 .SH SYNOPSIS
 .B zmergelog
+.IR [-vh]\ [-o\ outfmt]
 .IR logfile1
 .IR logfile2 \ ...
 .SH DESCRIPTION
Index: mergelog-4.5/src/mergelog.c
===================================================================
--- mergelog-4.5.orig/src/mergelog.c
+++ mergelog-4.5/src/mergelog.c
@@ -45,12 +45,16 @@
 #define mygets(a,b,c,d) fast_gzgets(a,b,c,d)
 #define myrewind gzrewind
 #define myclose gzclose
+#define mywrite(f,b,l) gzwrite((f),(b),(l))
+#define MYSUFFIX ".gz"
 #else
 #define myFH FILE
 #define myopen fopen
 #define mygets(a,b,c,d) fgets(a,b,c)
 #define myrewind rewind
 #define myclose fclose
+#define mywrite(f,b,l) fwrite((b),1,(l),(f))
+#define MYSUFFIX
 #endif
 
 
@@ -112,20 +116,141 @@ int main (int argc, char *argv[]) {
   char *trans_digits[60];
   char *trans_year[200];
   char months[24]="anebarprayunulugepctovec";
+  int goc;
+  int goUsage = 0, goVerbose = 0;
+  const char* goOutput = NULL;
+  const char* argv0 = *argv;
+  struct tm gotmFlags;
+  const char* goOutPtr;
+  myFH *outFile = NULL;
+  char outFileName[512];
+  struct tm outFileTM;
+  char outFileNewName[512];
+  size_t outBytes;
+
+  while((goc=getopt(argc,argv,"vho:"))!=-1) {
+    switch(goc) {
+      case 'h': /* -h print usage */
+        goUsage++;
+        break;
+      case 'v': /* -v verbose */
+        goVerbose++;
+        break;
+      case 'o': /* -o fmt output specification */
+        goOutput = optarg;
+        break;
+      default:  /* whatever.. but we better print out usage */
+        goUsage++;
+        break;
+    }
+  }
+  argv = &argv[optind];
+  argc -= optind;
 
-  /*
-    print usage if necessary
-   */
-  if (argc == 1) {
-    fprintf(stderr,"usage: %s logfile1 logfile2 ...\nmergelog %s Copyright (C) 2000-2001 Bertrand Demiddelaer\n",argv[0],VERSION);
+  if(argc<1)
+    goUsage++;
+  if(goUsage) {
+    printf(
+        "mergelog " VERSION " Copyright (c) 2000-2001 Bertrand Demiddelaer\n\n"
+        "I have at least %d reason(s) for lending you a helping hand on using the program.\n\n"
+        "Usage: %s [-vh] [-o fmt] logfile1 logfile2 ..\n\n"
+        " -h    Print this message\n"
+        " -v    Increase verbosity\n"
+        " -o    Give template for output file name to write log to\n"
+        "        instead of standard output in strftime(3) form.\n"
+        "        ( eg. -o %%Y-%%m-%%d-access.log" MYSUFFIX ")\n",
+                goUsage,argv0);
     exit(1);
   }
 
+  /* Analyze output format if given. It's better than trying to
+   * regenerate file name and compare it to the filename currently being
+   * written each time we want to dump one line. */
+  if(goOutput) {
+    memset(&gotmFlags,0,sizeof(gotmFlags));
+    for(goOutPtr=goOutput;*goOutPtr;goOutPtr++) {
+      if( (*goOutPtr=='%') && goOutPtr[1]) {
+        switch(*(++goOutPtr)) {
+          case 'a': /* abbreviated weekday */
+          case 'A': /* full weekday */
+          case 'u': /* day of week number */
+          case 'w': /* zerobased weekday number */
+            gotmFlags.tm_wday = 1;
+            break;
+          case 'b': /* abbreviated month */
+          case 'h': /* %b */
+          case 'B': /* full month */
+          case 'm': /* month number */
+            gotmFlags.tm_mon = 1;
+            break;
+          case 'C': /* century */
+          case 'y': /* centuryless year */
+          case 'Y': /* year with a century */
+            gotmFlags.tm_year = 1;
+            break;
+          case 'd': /* day of month */
+          case 'e': /* day of month space padded */
+            gotmFlags.tm_mday = 1;
+            break;
+          case 'D': /* %m/%d/%y americanism */
+            gotmFlags.tm_mon = gotmFlags.tm_mday = gotmFlags.tm_year = 1;
+            break;
+          case 'G': /* ISO 8601 year repsesentation. depends on week */
+          case 'g': /* ISO 8601 year repsesentation without century. depends on week */
+          case 'U': /* sunday-based week number */
+          case 'V': /* ISO 8601 week number */
+          case 'W': /* monday-based week number */
+          case 'x': /* preferred date representation for locale without the time */
+            gotmFlags.tm_yday = gotmFlags.tm_year = 1;
+            break;
+          case 'H': /* 24h format hour */
+          case 'I': /* 12h format hour */
+          case 'k': /* 24h format hour space padded */
+          case 'l': /* 12h format hour space padded */
+          case 'p': /* AM/PM indicator */
+          case 'P': /* am/pm indicator */
+            gotmFlags.tm_hour = 1;
+            break;
+          case 'j': /* day of year */
+            gotmFlags.tm_yday = 1;
+            break;
+          case 'M': /* minute */
+            gotmFlags.tm_min = 1;
+            break;
+          case 'n': /* newline character */
+          case 't': /* tab character */
+          case 'z': /* time zone as an offset from GMT */
+          case 'Z': /* timezone name or abbreviation */
+          case '%': /* % character */
+            break;
+          case 'r': /* %I:%M:%S %p time in am/pm notation*/
+          case 'T': /* %H:%M:%S time in 24h notation */
+          case 'X': /* preferred time representation for locale without the date */
+            gotmFlags.tm_hour = gotmFlags.tm_min = gotmFlags.tm_sec = 1;
+            break;
+          case 'R': /* %H:%M time in 24h notation without seconds */
+            gotmFlags.tm_hour = gotmFlags.tm_min = 1;
+            break;
+          case 'S': /* second */
+            gotmFlags.tm_sec = 1;
+            break;
+          case 'c': /* preferred date and time representation for locale */
+          case 'E': /* weird modifier */
+          case 'O': /* weird modifier */
+          case 's': /* number of seconds since epoch */
+          case '+': /* date and time in date(1) format */
+            memset(&gotmFlags,0xFF,sizeof(gotmFlags));
+            break;
+        }
+      }
+    }
+  }
+
 #ifdef USE_ZLIB
   /*
     check if there are enough gunzip buffers
    */
-  if(argc>MAX_FILES) {
+  if(argc>=MAX_FILES) {
     fputs("too many gzipped log files, aborting\n",stderr);
     exit(1);
   }
@@ -134,9 +259,9 @@ int main (int argc, char *argv[]) {
   /*
     open log files
   */
-  for (i=1;i<argc;i++) {
-    log_file[i-1]=myopen(argv[i],"r");
-    if (log_file[i-1] == NULL) {
+  for (i=0;i<argc;i++) {
+    log_file[i]=myopen(argv[i],"r");
+    if (log_file[i] == NULL) {
       fprintf(stderr,"can't open %s, aborting\n",argv[i]);
       exit(1);
     }
@@ -174,8 +299,8 @@ int main (int argc, char *argv[]) {
   /*
     init things for each log file and get the older date to start with
    */
-  nb_files=argc-1;
-  for (i=0;i<argc-1;i++) {
+  nb_files=argc;
+  for (i=0;i<argc;i++) {
 
 #ifdef USE_ZLIB
     /*
@@ -230,7 +355,7 @@ int main (int argc, char *argv[]) {
       */
       log_date=memchr(log_scan[i],'[',SCAN_SIZE);
       if (log_date == NULL) {
-	fprintf(stderr,"abort due to a problem with %s:\n%s\n",argv[i+1],log_buffer[i]);
+	fprintf(stderr,"abort due to a problem with %s:\n%s\n",argv[i],log_buffer[i]);
 	exit(1);
       }
       
@@ -239,7 +364,7 @@ int main (int argc, char *argv[]) {
        */
        for (j=0;((j == 12)&&(memcmp(months+2*j,log_date+5,2) != 0));j++);
        if (j == 12) {
-	 fprintf(stderr,"abort due to a problem with %s:\n%s\n",argv[i+1],log_buffer[i]);
+	 fprintf(stderr,"abort due to a problem with %s:\n%s\n",argv[i],log_buffer[i]);
 	 exit(1);
        }
        memcpy(log_month[i],trans_digits[j],2);
@@ -255,7 +380,7 @@ int main (int argc, char *argv[]) {
 	extract the date of this first line
       */
        if (sscanf(log_date+1,"%d/%3c/%d:%d:%d:%d",&day,month,&year,&hour,&minut,&second) < 6) {
-	 fprintf(stderr,"abort due to a problem with %s:\n%s\n",argv[i+1],log_buffer[i]);
+	 fprintf(stderr,"abort due to a problem with %s:\n%s\n",argv[i],log_buffer[i]);
 	 exit(1);
        }
       
@@ -270,7 +395,7 @@ int main (int argc, char *argv[]) {
       date->tm_isdst=-1;
       for (j=0;((j<12)&&(memcmp(months+2*j,month+1,2) != 0));j++);
       if (j == 12) {
-	fprintf(stderr,"abort due to a problem with %s:\n%s\n",argv[i+1],log_buffer[i]);
+	fprintf(stderr,"abort due to a problem with %s:\n%s\n",argv[i],log_buffer[i]);
 	exit(1);
       }
       date->tm_mon=j;
@@ -303,6 +428,8 @@ int main (int argc, char *argv[]) {
     exit if we have only empty files
    */
   if (nb_files == 0) {
+    if(outFile)
+      myclose(outFile);
     exit(0);
   }
 
@@ -323,7 +450,7 @@ int main (int argc, char *argv[]) {
   /*
     start to compute since this date
   */
-  nb_files_orig=argc-1;
+  nb_files_orig=argc;
   for(;;) {
 
     /*
@@ -377,7 +504,55 @@ int main (int argc, char *argv[]) {
 	  write the log line
 	  faster than a puts and we are sure to find a '\0' in log_buffer[i]
 	 */
-	write(1,log_buffer[i],(size_t)((char *)memchr(log_buffer[i],0,BUFFER_SIZE)-log_buffer[i]));
+        if(goOutput) {
+          /* Check whether any of the date/time components concerned
+           * have changes or if we have no output file opened. It
+           * doesn't make much sense to do all the comparinsons in case
+           * we have no output file opened, but it makes even less sense
+           * to check for file handle each time we evaluate date. */
+          if(
+#         define TMCOMPARE(tmpart) ( gotmFlags.tm_##tmpart && (outFileTM.tm_##tmpart!=date->tm_##tmpart) )
+            TMCOMPARE(sec) || TMCOMPARE(min) || TMCOMPARE(hour)
+            || TMCOMPARE(mday) || TMCOMPARE(mon) || TMCOMPARE(year)
+            || TMCOMPARE(wday) || TMCOMPARE(yday)
+            || !outFile
+#         undef TMCOMPARE
+          ) {
+            if(!strftime(outFileNewName,sizeof(outFileNewName)-1,goOutput,date)) {
+              fprintf(stderr,"abort due to a failed attempt to generate output file name\n");
+              if(outFile)
+                myclose(outFile);
+              exit(1);
+            }
+            if( (!outFile) || strcmp(outFileName,outFileNewName) ) {
+              /* okay, we're about to change output file or start
+               * writing first file */
+              if(outFile) {
+                myclose(outFile);
+              }
+              if(goVerbose)
+                fprintf(stderr, "Writing to \"%s\"..\n",outFileNewName);
+              outFile = myopen(outFileNewName,"a");
+              if(!outFile) {
+                fprintf(stderr, "abort due to a failed attempt to open/create output file %s: %m\n",outFileNewName);
+                exit(1);
+              }
+              memmove(outFileName,outFileNewName,sizeof(outFileName));
+              memmove(&outFileTM,date,sizeof(outFileTM));
+            }
+          }
+          /*  %Y-%m-%d_%H-%M-%S-access.log */
+          outBytes = (size_t)((char *)memchr(log_buffer[i],0,BUFFER_SIZE)-log_buffer[i]);
+          if( mywrite(outFile,log_buffer[i],outBytes) != outBytes) {
+            fprintf(stderr,"abort due to a failed write operation on output file %s: %m\n",outFileName);
+            if(outFile)
+              myclose(outFile);
+            exit(1);
+          }
+        }else{
+          /* do it the way we would do without */
+          write(1,log_buffer[i],(size_t)((char *)memchr(log_buffer[i],0,BUFFER_SIZE)-log_buffer[i]));
+        }
 
 	/*
 	  is it an end of file ?
@@ -388,7 +563,9 @@ int main (int argc, char *argv[]) {
 	    close all log files and exit if all end of files are reached
 	   */
 	  if (--nb_files == 0) {
-	    for (j=0;j<argc-1;j++) {
+            if(outFile)
+              myclose(outFile);
+	    for (j=0;j<argc;j++) {
 	      myclose(log_file[j]);
 	    }
 	    exit(0);
@@ -425,7 +602,7 @@ int main (int argc, char *argv[]) {
 	    } else {
 	      for (j=0;((j<12)&&(memcmp(months+2*j,log_date+5,2) != 0));j++);
 	      if (j == 12) {
-		fprintf(stderr,"problem with %s:\n%s\ncontinuing...\n",argv[i+1],log_buffer[i]);
+		fprintf(stderr,"problem with %s:\n%s\ncontinuing...\n",argv[i],log_buffer[i]);
 	      } else {
 		memcpy(log_month[i],trans_digits[j],2);
 		memcpy(log_month[i]+2,months+2*j,2);
@@ -439,7 +616,7 @@ int main (int argc, char *argv[]) {
 	      }
 	    }
 	  } else {
-	    fprintf(stderr,"problem with %s:\n%s\ncontinuing...\n",argv[i+1],log_buffer[i]);
+	    fprintf(stderr,"problem with %s:\n%s\ncontinuing...\n",argv[i],log_buffer[i]);
 	  }
 	}
       }
@@ -451,3 +628,7 @@ int main (int argc, char *argv[]) {
    */
   exit(1);
 }
+/* vim:set textwidth=72:				*/
+/* vim:set cindent smartindent:				*/
+/* vim:set formatoptions-=t formatoptions+=croql:	*/
+/* vim:set expandtab shiftwidth=2:			*/