libxml_disable_entity_loader() just in case...
[mediawiki.git] / maintenance / postgres / compare_schemas.pl
blob53aeb147fc9e4ebc8e0906cae2e029382c082848
1 #!/usr/bin/perl
3 ## Rough check that the base and postgres "tables.sql" are in sync
4 ## Should be run from maintenance/postgres
5 ## Checks a few other things as well...
7 use strict;
8 use warnings;
9 use Data::Dumper;
10 use Cwd;
12 #check_valid_sql();
14 my @old = ('../tables.sql');
15 my $new = 'tables.sql';
16 my @xfile;
18 ## Read in exceptions and other metadata
19 my %ok;
20 while (<DATA>) {
21 next unless /^(\w+)\s*:\s*([^#]+)/;
22 my ($name,$val) = ($1,$2);
23 chomp $val;
24 if ($name eq 'RENAME') {
25 die "Invalid rename\n" unless $val =~ /(\w+)\s+(\w+)/;
26 $ok{OLD}{$1} = $2;
27 $ok{NEW}{$2} = $1;
28 next;
30 if ($name eq 'XFILE') {
31 push @xfile, $val;
32 next;
34 for (split /\s+/ => $val) {
35 $ok{$name}{$_} = 0;
39 my $datatype = join '|' => qw(
40 bool
41 tinyint smallint int bigint real float
42 tinytext mediumtext text char varchar varbinary binary
43 timestamp datetime
44 tinyblob mediumblob blob
46 $datatype .= q{|ENUM\([\"\w\', ]+\)};
47 $datatype = qr{($datatype)};
49 my $typeval = qr{(\(\d+\))?};
51 my $typeval2 = qr{ signed| unsigned| binary| NOT NULL| NULL| PRIMARY KEY| AUTO_INCREMENT| default ['\-\d\w"]+| REFERENCES .+CASCADE};
53 my $indextype = join '|' => qw(INDEX KEY FULLTEXT), 'PRIMARY KEY', 'UNIQUE INDEX', 'UNIQUE KEY';
54 $indextype = qr{$indextype};
56 my $engine = qr{TYPE|ENGINE};
58 my $tabletype = qr{InnoDB|MyISAM|HEAP|HEAP MAX_ROWS=\d+|InnoDB MAX_ROWS=\d+ AVG_ROW_LENGTH=\d+};
60 my $charset = qr{utf8|binary};
62 open my $newfh, '<', $new or die qq{Could not open $new: $!\n};
65 my ($table,%old);
67 ## Read in the xfiles
68 my %xinfo;
69 for my $xfile (@xfile) {
70 print "Loading $xfile\n";
71 my $info = parse_sql($xfile);
72 for (keys %$info) {
73 $xinfo{$_} = $info->{$_};
77 for my $oldfile (@old) {
78 print "Loading $oldfile\n";
79 my $info = parse_sql($oldfile);
80 for (keys %xinfo) {
81 $info->{$_} = $xinfo{$_};
83 $old{$oldfile} = $info;
86 sub parse_sql {
88 my $oldfile = shift;
90 open my $oldfh, '<', $oldfile or die qq{Could not open $oldfile: $!\n};
92 my %info;
93 while (<$oldfh>) {
94 next if /^\s*\-\-/ or /^\s+$/;
95 s/\s*\-\- [\w ]+$//;
96 chomp;
98 if (/CREATE\s*TABLE/i) {
99 if (m{^CREATE TABLE /\*_\*/(\w+) \($}) {
100 $table = $1;
102 elsif (m{^CREATE TABLE /\*\$wgDBprefix\*/(\w+) \($}) {
103 $table = $1;
105 else {
106 die qq{Invalid CREATE TABLE at line $. of $oldfile\n};
108 $info{$table}{name}=$table;
110 elsif (m{^\) /\*\$wgDBTableOptions\*/}) {
111 $info{$table}{engine} = 'ENGINE';
112 $info{$table}{type} = 'variable';
114 elsif (/^\) ($engine)=($tabletype);$/) {
115 $info{$table}{engine}=$1;
116 $info{$table}{type}=$2;
118 elsif (/^\) ($engine)=($tabletype), DEFAULT CHARSET=($charset);$/) {
119 $info{$table}{engine}=$1;
120 $info{$table}{type}=$2;
121 $info{$table}{charset}=$3;
123 elsif (/^ (\w+) $datatype$typeval$typeval2{0,4},?$/) {
124 $info{$table}{column}{$1} = $2;
125 my $extra = $3 || '';
126 $info{$table}{columnfull}{$1} = "$2$extra";
128 elsif (m{^ UNIQUE KEY (\w+) \((.+?)\)}) {
130 elsif (m{^CREATE (?:UNIQUE )?(?:FULLTEXT )?INDEX /\*i\*/(\w+) ON /\*_\*/(\w+) \((.+?)\);}) {
132 elsif (m{^\s*PRIMARY KEY \([\w,]+\)}) {
134 else {
135 die "Cannot parse line $. of $oldfile:\n$_\n";
139 close $oldfh or die qq{Could not close "$oldfile": $!\n};
141 return \%info;
143 } ## end of parse_sql
145 for my $oldfile (@old) {
147 ## Begin non-standard indent
149 ## MySQL sanity checks
150 for my $table (sort keys %{$old{$oldfile}}) {
151 my $t = $old{$oldfile}{$table};
152 if ($t->{engine} eq 'TYPE') {
153 die "Invalid engine for $oldfile: $t->{engine}\n" unless $t->{name} eq 'profiling';
155 my $charset = $t->{charset} || '';
156 if ($oldfile !~ /binary/ and $charset eq 'binary') {
157 die "Invalid charset for $oldfile: $charset\n";
161 my $dtypelist = join '|' => qw(
162 SMALLINT INTEGER BIGINT NUMERIC SERIAL
163 TEXT CHAR VARCHAR
164 BYTEA
165 TIMESTAMPTZ
166 CIDR
168 my $dtype = qr{($dtypelist)};
169 my %new;
170 my ($infunction,$inview,$inrule,$lastcomma) = (0,0,0,0);
171 my %custom_type;
172 seek $newfh, 0, 0;
173 while (<$newfh>) {
174 next if /^\s*\-\-/ or /^\s*$/;
175 s/\s*\-\- [\w ']+$//;
176 next if /^BEGIN;/ or /^SET / or /^COMMIT;/;
177 next if /^CREATE SEQUENCE/;
178 next if /^CREATE(?: UNIQUE)? INDEX/;
179 next if /^CREATE FUNCTION/;
180 next if /^CREATE TRIGGER/ or /^ FOR EACH ROW/;
181 next if /^INSERT INTO/ or /^ VALUES \(/;
182 next if /^ALTER TABLE/;
183 next if /^DROP SEQUENCE/;
184 next if /^DROP FUNCTION/;
186 if (/^CREATE TYPE (\w+)/) {
187 die "Type $1 declared more than once!\n" if $custom_type{$1}++;
188 $dtype = qr{($dtypelist|$1)};
189 next;
192 chomp;
194 if (/^\$mw\$;?$/) {
195 $infunction = $infunction ? 0 : 1;
196 next;
198 next if $infunction;
200 next if /^CREATE VIEW/ and $inview = 1;
201 if ($inview) {
202 /;$/ and $inview = 0;
203 next;
206 next if /^CREATE RULE/ and $inrule = 1;
207 if ($inrule) {
208 /;$/ and $inrule = 0;
209 next;
212 if (/^CREATE TABLE "?(\w+)"? \($/) {
213 $table = $1;
214 $new{$table}{name}=$table;
215 $lastcomma = 1;
217 elsif (/^\);$/) {
218 if ($lastcomma) {
219 warn "Stray comma before line $.\n";
222 elsif (/^ (\w+) +$dtype.*?(,?)(?: --.*)?$/) {
223 $new{$table}{column}{$1} = $2;
224 if (!$lastcomma) {
225 print "Missing comma before line $. of $new\n";
227 $lastcomma = $3 ? 1 : 0;
229 elsif (m{^\s*PRIMARY KEY \([\w,]+\)}) {
230 $lastcomma = 0;
232 else {
233 die "Cannot parse line $. of $new:\n$_\n";
237 ## Which column types are okay to map from mysql to postgres?
238 my $COLMAP = q{
239 ## INTS:
240 tinyint SMALLINT
241 int INTEGER SERIAL
242 smallint SMALLINT
243 bigint BIGINT
244 real NUMERIC
245 float NUMERIC
247 ## TEXT:
248 varchar(15) TEXT
249 varchar(32) TEXT
250 varchar(70) TEXT
251 varchar(255) TEXT
252 varchar TEXT
253 text TEXT
254 tinytext TEXT
255 ENUM TEXT
257 ## TIMESTAMPS:
258 varbinary(14) TIMESTAMPTZ
259 binary(14) TIMESTAMPTZ
260 datetime TIMESTAMPTZ
261 timestamp TIMESTAMPTZ
263 ## BYTEA:
264 mediumblob BYTEA
266 ## OTHER:
267 bool SMALLINT # Sigh
270 ## Allow specific exceptions to the above
271 my $COLMAPOK = q{
272 ## User inputted text strings:
273 ar_comment tinyblob TEXT
274 fa_description tinyblob TEXT
275 img_description tinyblob TEXT
276 ipb_reason tinyblob TEXT
277 log_action varbinary(32) TEXT
278 log_type varbinary(32) TEXT
279 oi_description tinyblob TEXT
280 rev_comment tinyblob TEXT
281 rc_log_action varbinary(255) TEXT
282 rc_log_type varbinary(255) TEXT
284 ## Simple text-only strings:
285 ar_flags tinyblob TEXT
286 cf_name varbinary(255) TEXT
287 cf_value blob TEXT
288 ar_sha1 varbinary(32) TEXT
289 cl_collation varbinary(32) TEXT
290 cl_sortkey varbinary(230) TEXT
291 ct_params blob TEXT
292 fa_minor_mime varbinary(100) TEXT
293 fa_storage_group varbinary(16) TEXT # Just 'deleted' for now, should stay plain text
294 fa_storage_key varbinary(64) TEXT # sha1 plus text extension
295 ipb_address tinyblob TEXT # IP address or username
296 ipb_range_end tinyblob TEXT # hexadecimal
297 ipb_range_start tinyblob TEXT # hexadecimal
298 img_minor_mime varbinary(100) TEXT
299 lc_lang varbinary(32) TEXT
300 lc_value varbinary(32) TEXT
301 img_sha1 varbinary(32) TEXT
302 iw_wikiid varchar(64) TEXT
303 job_cmd varbinary(60) TEXT # Should we limit to 60 as well?
304 keyname varbinary(255) TEXT # No tablename prefix (objectcache)
305 ll_lang varbinary(20) TEXT # Language code
306 lc_value mediumblob TEXT
307 log_params blob TEXT # LF separated list of args
308 log_type varbinary(10) TEXT
309 ls_field varbinary(32) TEXT
310 md_deps mediumblob TEXT # JSON
311 md_module varbinary(255) TEXT
312 md_skin varbinary(32) TEXT
313 mr_blob mediumblob TEXT # JSON
314 mr_lang varbinary(32) TEXT
315 mr_resource varbinary(255) TEXT
316 mrl_message varbinary(255) TEXT
317 mrl_resource varbinary(255) TEXT
318 oi_minor_mime varbinary(100) TEXT
319 oi_sha1 varbinary(32) TEXT
320 old_flags tinyblob TEXT
321 old_text mediumblob TEXT
322 pp_propname varbinary(60) TEXT
323 pp_value blob TEXT
324 page_restrictions tinyblob TEXT # CSV string
325 pf_server varchar(30) TEXT
326 pr_level varbinary(60) TEXT
327 pr_type varbinary(60) TEXT
328 pt_create_perm varbinary(60) TEXT
329 pt_reason tinyblob TEXT
330 qc_type varbinary(32) TEXT
331 qcc_type varbinary(32) TEXT
332 qci_type varbinary(32) TEXT
333 rc_params blob TEXT
334 rev_sha1 varbinary(32) TEXT
335 rlc_to_blob blob TEXT
336 ts_tags blob TEXT
337 ufg_group varbinary(32) TEXT
338 ug_group varbinary(32) TEXT
339 ul_value blob TEXT
340 up_property varbinary(255) TEXT
341 up_value blob TEXT
342 us_sha1 varchar(31) TEXT
343 us_source_type varchar(50) TEXT
344 us_status varchar(50) TEXT
345 user_email_token binary(32) TEXT
346 user_ip varbinary(40) TEXT
347 user_newpassword tinyblob TEXT
348 user_options blob TEXT
349 user_password tinyblob TEXT
350 user_token binary(32) TEXT
351 iwl_prefix varbinary(20) TEXT
353 ## Text URLs:
354 el_index blob TEXT
355 el_to blob TEXT
356 iw_api blob TEXT
357 iw_url blob TEXT
358 tb_url blob TEXT
359 tc_url varbinary(255) TEXT
361 ## Deprecated or not yet used:
362 ar_text mediumblob TEXT
363 job_params blob TEXT
364 log_deleted tinyint INTEGER # Not used yet, but keep it INTEGER for safety
365 rc_type tinyint CHAR
367 ## Number tweaking:
368 fa_bits int SMALLINT # bits per pixel
369 fa_height int SMALLINT
370 fa_width int SMALLINT # Hope we don't see an image this wide...
371 hc_id int BIGINT # Odd that site_stats is all bigint...
372 img_bits int SMALLINT # bits per image should stay sane
373 oi_bits int SMALLINT
375 ## True binary fields, usually due to gzdeflate and/or serialize:
376 math_inputhash varbinary(16) BYTEA
377 math_outputhash varbinary(16) BYTEA
379 ## Namespaces: not need for such a high range
380 ar_namespace int SMALLINT
381 job_namespace int SMALLINT
382 log_namespace int SMALLINT
383 page_namespace int SMALLINT
384 pl_namespace int SMALLINT
385 pt_namespace int SMALLINT
386 qc_namespace int SMALLINT
387 rc_namespace int SMALLINT
388 rd_namespace int SMALLINT
389 rlc_to_namespace int SMALLINT
390 tl_namespace int SMALLINT
391 wl_namespace int SMALLINT
393 ## Easy enough to change if a wiki ever does grow this big:
394 ss_active_users bigint INTEGER
395 ss_good_articles bigint INTEGER
396 ss_total_edits bigint INTEGER
397 ss_total_pages bigint INTEGER
398 ss_total_views bigint INTEGER
399 ss_users bigint INTEGER
401 ## True IP - keep an eye on these, coders tend to make textual assumptions
402 rc_ip varbinary(40) CIDR # Want to keep an eye on this
404 ## Others:
405 tc_time int TIMESTAMPTZ
410 my %colmap;
411 for (split /\n/ => $COLMAP) {
412 next unless /^\w/;
413 s/(.*?)#.*/$1/;
414 my ($col,@maps) = split / +/, $_;
415 for (@maps) {
416 $colmap{$col}{$_} = 1;
420 my %colmapok;
421 for (split /\n/ => $COLMAPOK) {
422 next unless /^\w/;
423 my ($col,$old,$new) = split / +/, $_;
424 $colmapok{$col}{$old}{$new} = 1;
427 ## Old but not new
428 for my $t (sort keys %{$old{$oldfile}}) {
429 if (!exists $new{$t} and !exists $ok{OLD}{$t}) {
430 print "Table not in $new: $t\n";
431 next;
433 next if exists $ok{OLD}{$t} and !$ok{OLD}{$t};
434 my $newt = exists $ok{OLD}{$t} ? $ok{OLD}{$t} : $t;
435 my $oldcol = $old{$oldfile}{$t}{column};
436 my $oldcolfull = $old{$oldfile}{$t}{columnfull};
437 my $newcol = $new{$newt}{column};
438 for my $c (keys %$oldcol) {
439 if (!exists $newcol->{$c}) {
440 print "Column $t.$c not in $new\n";
441 next;
444 for my $c (sort keys %$newcol) {
445 if (!exists $oldcol->{$c}) {
446 print "Column $t.$c not in $oldfile\n";
447 next;
449 ## Column types (roughly) match up?
450 my $new = $newcol->{$c};
451 my $old = $oldcolfull->{$c};
453 ## Known exceptions:
454 next if exists $colmapok{$c}{$old}{$new};
456 $old =~ s/ENUM.*/ENUM/;
458 next if $old eq 'ENUM' and $new eq 'media_type';
460 if (! exists $colmap{$old}{$new}) {
461 print "Column types for $t.$c do not match: $old does not map to $new\n";
465 ## New but not old:
466 for (sort keys %new) {
467 if (!exists $old{$oldfile}{$_} and !exists $ok{NEW}{$_}) {
468 print "Not in $oldfile: $_\n";
469 next;
474 } ## end each file to be parsed
477 sub check_valid_sql {
479 ## Check for a few common problems in most php files
481 my $olddir = getcwd();
482 chdir("../..");
483 for my $basedir (qw/includes extensions/) {
484 scan_dir($basedir);
486 chdir $olddir;
488 return;
490 } ## end of check_valid_sql
493 sub scan_dir {
495 my $dir = shift;
497 opendir my $dh, $dir or die qq{Could not opendir $dir: $!\n};
498 #print "Scanning $dir...\n";
499 for my $file (grep { -f "$dir/$_" and /\.php$/ } readdir $dh) {
500 find_problems("$dir/$file");
502 rewinddir $dh;
503 for my $subdir (grep { -d "$dir/$_" and ! /\./ } readdir $dh) {
504 scan_dir("$dir/$subdir");
506 closedir $dh or die qq{Closedir failed: $!\n};
507 return;
509 } ## end of scan_dir
511 sub find_problems {
513 my $file = shift;
514 open my $fh, '<', $file or die qq{Could not open "$file": $!\n};
515 my $lastline = '';
516 my $inarray = 0;
517 while (<$fh>) {
518 if (/FORCE INDEX/ and $file !~ /Database\w*\.php/) {
519 warn "Found FORCE INDEX string at line $. of $file\n";
521 if (/REPLACE INTO/ and $file !~ /Database\w*\.php/) {
522 warn "Found REPLACE INTO string at line $. of $file\n";
524 if (/\bIF\s*\(/ and $file !~ /DatabaseMySQL\.php/) {
525 warn "Found IF string at line $. of $file\n";
527 if (/\bCONCAT\b/ and $file !~ /Database\w*\.php/) {
528 warn "Found CONCAT string at line $. of $file\n";
530 if (/\bGROUP\s+BY\s*\d\b/i and $file !~ /Database\w*\.php/) {
531 warn "Found GROUP BY # at line $. of $file\n";
533 if (/wfGetDB\s*\(\s+\)/io) {
534 warn "wfGETDB is missing parameters at line $. of $file\n";
536 if (/=\s*array\s*\(\s*$/) {
537 $inarray = 1;
538 next;
540 if ($inarray) {
541 if (/\s*\);\s*$/) {
542 $inarray = 0;
543 next;
545 next if ! /\w/ or /array\(\s*$/ or /^\s*#/ or m{^\s*//};
546 if (! /,/) {
547 my $nextline = <$fh>;
548 last if ! defined $nextline;
549 if ($nextline =~ /^\s*\)[;,]/) {
550 $inarray = 0;
551 next;
553 #warn "Array is missing a comma? Line $. of $file\n";
557 close $fh or die qq{Could not close "$file": $!\n};
558 return;
560 } ## end of find_problems
563 __DATA__
564 ## Known exceptions
565 OLD: searchindex ## We use tsearch2 directly on the page table instead
566 RENAME: user mwuser ## Reserved word causing lots of problems
567 RENAME: text pagecontent ## Reserved word
568 XFILE: ../archives/patch-profiling.sql