]> git.ipfire.org Git - thirdparty/sarg.git/blame - topsites.c
Rename configure.in as configure.ac
[thirdparty/sarg.git] / topsites.c
CommitLineData
25697a35 1/*
94ff9470 2 * SARG Squid Analysis Report Generator http://sarg.sourceforge.net
110ce984 3 * 1998, 2015
25697a35
GS
4 *
5 * SARG donations:
6 * please look at http://sarg.sourceforge.net/donations.php
1164c474
FM
7 * Support:
8 * http://sourceforge.net/projects/sarg/forums/forum/363374
25697a35
GS
9 * ---------------------------------------------------------------------
10 *
11 * This program is free software; you can redistribute it and/or modify
12 * it under the terms of the GNU General Public License as published by
13 * the Free Software Foundation; either version 2 of the License, or
14 * (at your option) any later version.
15 *
16 * This program is distributed in the hope that it will be useful,
17 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19 * GNU General Public License for more details.
20 *
21 * You should have received a copy of the GNU General Public License
22 * along with this program; if not, write to the Free Software
23 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
24 *
25 */
26
27#include "include/conf.h"
5f3cfd1d 28#include "include/defs.h"
25697a35 29
9dc20988
FM
30#ifdef ENABLE_DOUBLE_CHECK_DATA
31extern struct globalstatstruct globstat;
32#endif
33
32e71fa4 34void topsites(void)
25697a35 35{
9bd92830
FM
36 FILE *fp_in, *fp_ou;
37
38 char *buf;
39 char *url;
40 char *ourl=NULL;
997a4328 41 char csort[4096];
9bd92830
FM
42 char general[MAXLEN];
43 char general2[MAXLEN];
44 char general3[MAXLEN];
45 char sites[MAXLEN];
46 char report[MAXLEN];
95092e58 47 char ouser[MAX_USER_LEN]="";
9bd92830
FM
48 const char *sortf;
49 const char *sortt;
50 long long int nacc;
51 long long int nbytes;
52 long long int ntime;
53 long long int tnacc=0;
54 long long int tnbytes=0;
55 long long int tntime=0;
56 long long int twork1=0, twork2=0, twork3=0;
9dc20988
FM
57#ifdef ENABLE_DOUBLE_CHECK_DATA
58 long long int ttnacc=0;
59 long long int ttnbytes=0;
60 long long int ttntime=0;
61#endif
95092e58 62 int nusers=0;
9bd92830
FM
63 int regs=0;
64 int cstatus;
65 int url_len;
66 int ourl_size=0;
67 struct getwordstruct gwarea;
68 longline line;
69 struct generalitemstruct item;
70
9f93fec3 71 if(Privacy) {
af961877 72 if (debugz>=LogLevel_Process) debugaz(__FILE__,__LINE__,_("Top sites report not produced because privacy option is on\n"));
9bd92830 73 return;
9f93fec3 74 }
b7413c4c 75 if (debugz>=LogLevel_Process)
af961877 76 debuga(__FILE__,__LINE__,_("Creating top sites report...\n"));
9bd92830
FM
77
78 sprintf(general,"%s/sarg-general",outdirname);
79 sprintf(sites,"%s/sarg-sites",outdirname);
80 sprintf(general2,"%s/sarg-general2",outdirname);
81 sprintf(general3,"%s/sarg-general3",outdirname);
82
124ec0b0 83 sprintf(report,"%s/topsites.html",outdirname);
9bd92830 84
95092e58 85 if (snprintf(csort,sizeof(csort),"sort -t \"\t\" -k 4,4 -k 1,1 -o \"%s\" \"%s\"",general2,general)>=sizeof(csort)) {
af961877 86 debuga(__FILE__,__LINE__,_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),general,general2);
78eeb33f
FM
87 exit(EXIT_FAILURE);
88 }
9bd92830
FM
89 cstatus=system(csort);
90 if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
af961877
FM
91 debuga(__FILE__,__LINE__,_("sort command return status %d\n"),WEXITSTATUS(cstatus));
92 debuga(__FILE__,__LINE__,_("sort command: %s\n"),csort);
9bd92830
FM
93 exit(EXIT_FAILURE);
94 }
95
96 if((fp_in=fopen(general2,"r"))==NULL) {
af961877
FM
97 debuga(__FILE__,__LINE__,_("Cannot open file \"%s\": %s\n"),general2,strerror(errno));
98 debuga(__FILE__,__LINE__,_("sort command: %s\n"),csort);
007905af 99 exit(EXIT_FAILURE);
9bd92830
FM
100 }
101
102 if((fp_ou=fopen(general3,"w"))==NULL) {
af961877 103 debuga(__FILE__,__LINE__,_("Cannot open file \"%s\": %s\n"),general3,strerror(errno));
007905af 104 exit(EXIT_FAILURE);
9bd92830
FM
105 }
106
107 if ((line=longline_create())==NULL) {
af961877 108 debuga(__FILE__,__LINE__,_("Not enough memory to read file \"%s\"\n"),general2);
9bd92830
FM
109 exit(EXIT_FAILURE);
110 }
111
112 while((buf=longline_read(fp_in,line))!=NULL) {
113 ger_read(buf,&item,general2);
114 if(item.total) continue;
115
116 if(!regs) {
117 url_len=strlen(item.url);
118 if (!ourl || url_len>=ourl_size) {
119 ourl_size=url_len+1;
120 ourl=realloc(ourl,ourl_size);
121 if (!ourl) {
af961877 122 debuga(__FILE__,__LINE__,_("Not enough memory to store the url\n"));
9bd92830
FM
123 exit(EXIT_FAILURE);
124 }
125 }
126 strcpy(ourl,item.url);
127 regs++;
128 }
129
130 if(strcmp(item.url,ourl) != 0) {
131 /*
132 This complicated printf is due to Microsoft's inability to comply with any standard. Msvcrt is unable
133 to print a long long int unless it is exactly 64-bits long.
134 */
95092e58 135 fprintf(fp_ou,"%"PRIu64"\t%"PRIu64"\t%"PRIu64"\t%d\t%s\n",(uint64_t)tnacc,(uint64_t)tnbytes,(uint64_t)tntime,nusers,ourl);
9bd92830
FM
136 url_len=strlen(item.url);
137 if (url_len>=ourl_size) {
138 ourl_size=url_len+1;
139 ourl=realloc(ourl,ourl_size);
140 if (!ourl) {
af961877 141 debuga(__FILE__,__LINE__,_("Not enough memory to store the url\n"));
9bd92830
FM
142 exit(EXIT_FAILURE);
143 }
144 }
145 strcpy(ourl,item.url);
95092e58 146 strcpy(ouser,item.user);
9bd92830
FM
147 tnacc=0;
148 tnbytes=0;
149 tntime=0;
95092e58
FM
150 nusers=1;
151 } else if (strcmp(item.user,ouser)!=0) {
152 strcpy(ouser,item.user);
153 nusers++;
9bd92830
FM
154 }
155
156 tnacc+=item.nacc;
157 tnbytes+=item.nbytes;
158 tntime+=item.nelap;
9dc20988
FM
159#ifdef ENABLE_DOUBLE_CHECK_DATA
160 ttnacc+=item.nacc;
161 ttnbytes+=item.nbytes;
162 ttntime+=item.nelap;
163#endif
9bd92830 164 }
204781f4 165 if (fclose(fp_in)==EOF) {
af961877 166 debuga(__FILE__,__LINE__,_("Read error in \"%s\": %s\n"),general2,strerror(errno));
204781f4
FM
167 exit(EXIT_FAILURE);
168 }
9bd92830
FM
169 longline_destroy(&line);
170
171 if (ourl) {
172 /*
173 This complicated printf is due to Microsoft's inability to comply with any standard. Msvcrt is unable
174 to print a long long int unless it is exactly 64-bits long.
175 */
95092e58 176 fprintf(fp_ou,"%"PRIu64"\t%"PRIu64"\t%"PRIu64"\t%d\t%s\n",(uint64_t)tnacc,(uint64_t)tnbytes,(uint64_t)tntime,nusers,ourl);
9bd92830
FM
177 free(ourl);
178 }
179
507460ae 180 if (fclose(fp_ou)==EOF) {
af961877 181 debuga(__FILE__,__LINE__,_("Write error in \"%s\": %s\n"),general3,strerror(errno));
507460ae
FM
182 exit(EXIT_FAILURE);
183 }
9bd92830 184
9dc20988
FM
185#ifdef ENABLE_DOUBLE_CHECK_DATA
186 if (ttnacc!=globstat.nacc || ttnbytes!=globstat.nbytes || ttntime!=globstat.elap) {
af961877 187 debuga(__FILE__,__LINE__,_("Total statistics mismatch when reading \"%s\" to produce the top sites\n"),general2);
9dc20988
FM
188 exit(EXIT_FAILURE);
189 }
190#endif
191
11767c6a 192 if (!KeepTempLog && unlink(general2)) {
af961877 193 debuga(__FILE__,__LINE__,_("Cannot delete \"%s\": %s\n"),general2,strerror(errno));
9dc20988
FM
194 exit(EXIT_FAILURE);
195 }
196
9bd92830
FM
197 if((TopsitesSort & TOPSITE_SORT_CONNECT) != 0) {
198 sortf="-k 1,1 -k 2,2";
199 } else if((TopsitesSort & TOPSITE_SORT_BYTES) != 0) {
200 sortf="-k 2,2 -k 1,1";
201 } else if((TopsitesSort & TOPSITE_SORT_TIME) != 0) {
202 sortf="-k 3,3";
7d82ea1a
FM
203 } else if((TopsitesSort & TOPSITE_SORT_USER) != 0) {
204 sortf="-k 4,4 -k 1,1 -k 2,2";
9bd92830
FM
205 } else {
206 sortf="-k 2,2 -k 1,1"; //default is BYTES
207 }
208 if((TopsitesSort & TOPSITE_SORT_REVERSE) != 0) {
209 sortt="-r";
210 } else {
211 sortt="";
212 }
213
78eeb33f 214 if (snprintf(csort,sizeof(csort),"sort -t \"\t\" %s -n %s -o \"%s\" \"%s\"",sortt,sortf,sites,general3)>=sizeof(csort)) {
af961877 215 debuga(__FILE__,__LINE__,_("Sort command too long when sorting file \"%s\" to \"%s\"\n"),general3,sites);
78eeb33f
FM
216 exit(EXIT_FAILURE);
217 }
9bd92830
FM
218 cstatus=system(csort);
219 if (!WIFEXITED(cstatus) || WEXITSTATUS(cstatus)) {
af961877
FM
220 debuga(__FILE__,__LINE__,_("sort command return status %d\n"),WEXITSTATUS(cstatus));
221 debuga(__FILE__,__LINE__,_("sort command: %s\n"),csort);
9bd92830
FM
222 exit(EXIT_FAILURE);
223 }
224 if((fp_in=fopen(sites,"r"))==NULL) {
af961877
FM
225 debuga(__FILE__,__LINE__,_("Cannot open file \"%s\": %s\n"),sites,strerror(errno));
226 debuga(__FILE__,__LINE__,_("sort command: %s\n"),csort);
9bd92830
FM
227 exit(EXIT_FAILURE);
228 }
229
11767c6a 230 if (!KeepTempLog && unlink(general3)) {
af961877 231 debuga(__FILE__,__LINE__,_("Cannot delete \"%s\": %s\n"),general3,strerror(errno));
08f9b029
FM
232 exit(EXIT_FAILURE);
233 }
9bd92830
FM
234
235 if((fp_ou=fopen(report,"w"))==NULL) {
af961877 236 debuga(__FILE__,__LINE__,_("Cannot open file \"%s\": %s\n"),report,strerror(errno));
007905af 237 exit(EXIT_FAILURE);
9bd92830
FM
238 }
239
240 write_html_header(fp_ou,(IndexTree == INDEX_TREE_DATE) ? 3 : 1,_("Top sites"),HTML_JS_SORTTABLE);
241 fputs("<tr><td class=\"header_c\">",fp_ou);
242 fprintf(fp_ou,_("Period: %s"),period.html);
243 fputs("</td></tr>\n",fp_ou);
244 fputs("<tr><th class=\"header_c\">",fp_ou);
245 fprintf(fp_ou,_("Top %d sites"),TopSitesNum);
246 fputs("</th></tr>\n",fp_ou);
247 close_html_header(fp_ou);
248
249 fputs("<div class=\"report\"><table cellpadding=\"1\" cellspacing=\"2\"",fp_ou);
250 if (SortTableJs[0]) fputs(" class=\"sortable\"",fp_ou);
251 fputs(">\n",fp_ou);
87e8625b
FM
252 fprintf(fp_ou,"<thead><tr><th class=\"header_l\">%s</th><th class=\"header_l",
253 /* TRANSLATORS: This is a column header showing the position of the entry in the sorted list. */
254 _("NUM"));
9bd92830 255 if (SortTableJs[0]) fputs(" sorttable_alpha",fp_ou);
87e8625b
FM
256 fprintf(fp_ou,"\">%s</th><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th><th class=\"header_l\">%s</th></tr></thead>\n",
257 /* TRANSLATORS: This is a column header showing the URL of the visited sites. */
258 _("ACCESSED SITE"),
259 /* TRANSLATORS: This is a column header showing the number of connections to a visited site. */
260 _("CONNECT"),
261 /* TRANSLATORS: This is a column header showing the number of transfered bytes. */
262 _("BYTES"),
263 /* TRANSLATORS: This is a column header showing the time spent by the proxy processing the requests. */
eb7ab620 264 pgettext("duration","TIME"),
87e8625b
FM
265 /* TRANSLATORS: This is a column header showing the number of users who visited a sites. */
266 _("USERS"));
9bd92830
FM
267
268 regs=0;
269 ntopsites = 0;
270
271 if ((line=longline_create())==NULL) {
af961877 272 debuga(__FILE__,__LINE__,_("Not enough memory to read file \"%s\"\n"),sites);
9bd92830
FM
273 exit(EXIT_FAILURE);
274 }
275
276 while(regs<TopSitesNum && (buf=longline_read(fp_in,line))!=NULL) {
277 getword_start(&gwarea,buf);
278 if (getword_atoll(&nacc,&gwarea,'\t')<0) {
af961877 279 debuga(__FILE__,__LINE__,_("Invalid record in file \"%s\"\n"),sites);
9bd92830
FM
280 exit(EXIT_FAILURE);
281 }
282 if (nacc == 0) continue;
283 if (getword_atoll(&nbytes,&gwarea,'\t')<0 || getword_atoll(&ntime,&gwarea,'\t')<0) {
af961877 284 debuga(__FILE__,__LINE__,_("Invalid record in file \"%s\"\n"),sites);
9bd92830
FM
285 exit(EXIT_FAILURE);
286 }
95092e58 287 if (getword_atoi(&nusers,&gwarea,'\t')<0) {
af961877 288 debuga(__FILE__,__LINE__,_("Invalid number of users in file \"%s\"\n"),sites);
95092e58
FM
289 exit(EXIT_FAILURE);
290 }
9bd92830 291 if (getword_ptr(buf,&url,&gwarea,'\t')<0) {
af961877 292 debuga(__FILE__,__LINE__,_("Invalid url in file \"%s\"\n"),sites);
9bd92830
FM
293 exit(EXIT_FAILURE);
294 }
295
296 twork1=nacc;
297 twork2=nbytes;
298 twork3=ntime;
299
67a93701 300 fprintf(fp_ou,"<tr><td class=\"data\">%d</td><td class=\"data2\">",++regs);
9bd92830 301
67a93701 302 if(BlockIt[0] != '\0' && url[0]!=ALIAS_PREFIX) {
9bd92830
FM
303 fprintf(fp_ou,"<a href=\"%s%s?url=\"",wwwDocumentRoot,BlockIt);
304 output_html_url(fp_ou,url);
305 fputs("\"><img src=\"../images/sarg-squidguard-block.png\"></a>&nbsp;",fp_ou);
306 }
307
6fa33a32 308 output_html_link(fp_ou,url,100);
67a93701 309 fputs("</td><td class=\"data\"",fp_ou);
9bd92830
FM
310 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork1);
311 fprintf(fp_ou,">%s</td>",fixnum(twork1,1));
312 fputs("<td class=\"data\"",fp_ou);
313 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork2);
314 fprintf(fp_ou,">%s</td>",fixnum(twork2,1));
315 fputs("<td class=\"data\"",fp_ou);
316 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%"PRId64"\"",(uint64_t)twork3);
95092e58
FM
317 fprintf(fp_ou,">%s</td>",fixtime(twork3));
318 fputs("<td class=\"data\"",fp_ou);
319 if (SortTableJs[0]) fprintf(fp_ou," sorttable_customkey=\"%d\"",nusers);
320 fprintf(fp_ou,">%s</td></tr>\n",fixnum(nusers,1));
9bd92830 321 }
204781f4 322 if (fclose(fp_in)==EOF) {
af961877 323 debuga(__FILE__,__LINE__,_("Read error in \"%s\": %s\n"),sites,strerror(errno));
204781f4
FM
324 exit(EXIT_FAILURE);
325 }
9bd92830
FM
326 longline_destroy(&line);
327
328 fputs("</table></div>\n",fp_ou);
342bd723 329 write_html_trailer(fp_ou);
507460ae 330 if (fclose(fp_ou)==EOF) {
af961877 331 debuga(__FILE__,__LINE__,_("Write error in \"%s\": %s\n"),report,strerror(errno));
507460ae
FM
332 exit(EXIT_FAILURE);
333 }
9bd92830
FM
334
335 return;
25697a35 336}