LEFT | RIGHT |
1 # This file is part of the Adblock Plus web scripts, | 1 # This file is part of the Adblock Plus web scripts, |
2 # Copyright (C) 2006-2014 Eyeo GmbH | 2 # Copyright (C) 2006-2014 Eyeo GmbH |
3 # | 3 # |
4 # Adblock Plus is free software: you can redistribute it and/or modify | 4 # Adblock Plus is free software: you can redistribute it and/or modify |
5 # it under the terms of the GNU General Public License version 3 as | 5 # it under the terms of the GNU General Public License version 3 as |
6 # published by the Free Software Foundation. | 6 # published by the Free Software Foundation. |
7 # | 7 # |
8 # Adblock Plus is distributed in the hope that it will be useful, | 8 # Adblock Plus is distributed in the hope that it will be useful, |
9 # but WITHOUT ANY WARRANTY; without even the implied warranty of | 9 # but WITHOUT ANY WARRANTY; without even the implied warranty of |
10 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | 10 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
11 # GNU General Public License for more details. | 11 # GNU General Public License for more details. |
12 # | 12 # |
13 # You should have received a copy of the GNU General Public License | 13 # You should have received a copy of the GNU General Public License |
14 # along with Adblock Plus. If not, see <http://www.gnu.org/licenses/>. | 14 # along with Adblock Plus. If not, see <http://www.gnu.org/licenses/>. |
15 | 15 |
16 import re, warnings | 16 import itertools |
17 | 17 import warnings |
18 try: | 18 import re |
19 import libxml2 | 19 import urllib2 |
20 except ImportError: | 20 from xml.dom import minidom |
21 libxml2 = None | |
22 warnings.warn('libxml2 is not installed, can not validate PAD files') | |
23 | 21 |
24 FIELDS = [ | 22 FIELDS = [ |
25 ('/XML_DIZ_INFO/MASTER_PAD_VERSION_INFO/MASTER_PAD_VERSION', r'^4\.0\Z'), | 23 (['XML_DIZ_INFO', 'MASTER_PAD_VERSION_INFO', 'MASTER_PAD_VERSION'], r'^4\.0\Z'
), |
26 ('/XML_DIZ_INFO/MASTER_PAD_VERSION_INFO/MASTER_PAD_EDITOR', r'^[^<\x09]{0,100}
\Z'), | 24 (['XML_DIZ_INFO', 'MASTER_PAD_VERSION_INFO', 'MASTER_PAD_EDITOR'], r'^[^<\x09]
{0,100}\Z'), |
27 ('/XML_DIZ_INFO/MASTER_PAD_VERSION_INFO/MASTER_PAD_INFO', r'^[^<\x09]{0,1000}\
Z'), | 25 (['XML_DIZ_INFO', 'MASTER_PAD_VERSION_INFO', 'MASTER_PAD_INFO'], r'^[^<\x09]{0
,1000}\Z'), |
28 ('/XML_DIZ_INFO/Company_Info/Company_Name', r'^[^<\x09]{2,40}\Z'), | 26 (['XML_DIZ_INFO', 'MASTER_PAD_VERSION_INFO', 'CERTIFIED'], r'^[YyNn]\Z'), |
29 ('/XML_DIZ_INFO/Company_Info/Address_1', r'^[a-zA-Z0-9\xbc-\xff .\-,#\/\x27]{0
,40}\Z'), | 27 (['XML_DIZ_INFO', 'MASTER_PAD_VERSION_INFO', 'CERTIFICATE_ID'], r'^(crt\-[0-9A
-Z]{12}|)\Z'), |
30 ('/XML_DIZ_INFO/Company_Info/Address_2', r'^[a-zA-Z\xbc-\xff0-9 .\-,#\/\x27]{0
,40}\Z'), | 28 (['XML_DIZ_INFO', 'MASTER_PAD_VERSION_INFO', 'CERTIFICATE_LICENSE'], r'^(http\
:\/\/repository\.appvisor\.com\/crt\-[0-9a-z]{12}|)\Z'), |
31 ('/XML_DIZ_INFO/Company_Info/City_Town', r'^[a-zA-Z\xbc-\xff0-9 .\-,#\/\x27]{2
,40}\Z'), | 29 (['XML_DIZ_INFO', 'Company_Info', 'PublisherID'], r'^(pid-[0-9a-z]{12}|)\Z'), |
32 ('/XML_DIZ_INFO/Company_Info/State_Province', r'^[a-zA-Z\xbc-\xff0-9 .\-,\/]{0
,30}\Z'), | 30 (['XML_DIZ_INFO', 'Company_Info', 'Company_Name'], r'^[^<\x09]{2,40}\Z'), |
33 ('/XML_DIZ_INFO/Company_Info/Zip_Postal_Code', r'^[^<\x09]{0,20}\Z'), | 31 (['XML_DIZ_INFO', 'Company_Info', 'Address_1'], r'^[a-zA-Z0-9\xbc-\xff .\-,#\/
\x27]{0,40}\Z'), |
34 ('/XML_DIZ_INFO/Company_Info/Country', r'^[a-z A-Z\xbc-\xff\x27-]{2,40}\Z'), | 32 (['XML_DIZ_INFO', 'Company_Info', 'Address_2'], r'^[a-zA-Z\xbc-\xff0-9 .\-,#\/
\x27]{0,40}\Z'), |
35 ('/XML_DIZ_INFO/Company_Info/Company_WebSite_URL', r'^http:\/\/.{2,120}\Z'), | 33 (['XML_DIZ_INFO', 'Company_Info', 'City_Town'], r'^[a-zA-Z\xbc-\xff0-9 .\-,#\/
\x27]{2,40}\Z'), |
36 ('/XML_DIZ_INFO/Company_Info/Contact_Info/Author_First_Name', r'^[^<\x09]{2,30
}\Z'), | 34 (['XML_DIZ_INFO', 'Company_Info', 'State_Province'], r'^[a-zA-Z\xbc-\xff0-9 .\
-,\/]{0,30}\Z'), |
37 ('/XML_DIZ_INFO/Company_Info/Contact_Info/Author_Last_Name', r'^[^<\x09]{2,30}
\Z'), | 35 (['XML_DIZ_INFO', 'Company_Info', 'Zip_Postal_Code'], r'^[^<\x09]{0,20}\Z'), |
38 ('/XML_DIZ_INFO/Company_Info/Contact_Info/Author_Email', r'^.{2,30}\@.{2,63}\.
.{2,20}\Z'), | 36 (['XML_DIZ_INFO', 'Company_Info', 'Country'], r'^[a-z A-Z\xbc-\xff\x27-]{2,40}
\Z'), |
39 ('/XML_DIZ_INFO/Company_Info/Contact_Info/Contact_First_Name', r'^[^<\x09]{2,3
0}\Z'), | 37 (['XML_DIZ_INFO', 'Company_Info', 'Company_WebSite_URL'], r'^(http|https):\/\/
.{2,120}\Z'), |
40 ('/XML_DIZ_INFO/Company_Info/Contact_Info/Contact_Last_Name', r'^[^<\x09]{2,30
}\Z'), | 38 (['XML_DIZ_INFO', 'Company_Info', 'Contact_Info', 'Author_First_Name'], r'^[^<
\x09]{2,30}\Z'), |
41 ('/XML_DIZ_INFO/Company_Info/Contact_Info/Contact_Email', r'^.{2,30}\@.{2,63}\
..{2,20}\Z'), | 39 (['XML_DIZ_INFO', 'Company_Info', 'Contact_Info', 'Author_Last_Name'], r'^[^<\
x09]{2,30}\Z'), |
42 ('/XML_DIZ_INFO/Company_Info/Support_Info/Sales_Email', r'^.{2,30}\@.{2,63}\..
{2,20}\Z'), | 40 (['XML_DIZ_INFO', 'Company_Info', 'Contact_Info', 'Author_Email'], r'^.{2,30}\
@.{2,63}\..{2,20}\Z'), |
43 ('/XML_DIZ_INFO/Company_Info/Support_Info/Support_Email', r'^.{2,30}\@.{2,63}\
..{2,20}\Z'), | 41 (['XML_DIZ_INFO', 'Company_Info', 'Contact_Info', 'Contact_First_Name'], r'^[^
<\x09]{2,30}\Z'), |
44 ('/XML_DIZ_INFO/Company_Info/Support_Info/General_Email', r'^.{2,30}\@.{2,63}\
..{2,20}\Z'), | 42 (['XML_DIZ_INFO', 'Company_Info', 'Contact_Info', 'Contact_Last_Name'], r'^[^<
\x09]{2,30}\Z'), |
45 ('/XML_DIZ_INFO/Company_Info/Support_Info/Sales_Phone', r'^\+{0,2}(([0-9#*()-\
/_] *){7,40})?\Z'), | 43 (['XML_DIZ_INFO', 'Company_Info', 'Contact_Info', 'Contact_Email'], r'^.{2,30}
\@.{2,63}\..{2,20}\Z'), |
46 ('/XML_DIZ_INFO/Company_Info/Support_Info/Support_Phone', r'^\+{0,2}(([0-9#*()
-\/_] *){7,40})?\Z'), | 44 (['XML_DIZ_INFO', 'Company_Info', 'Support_Info', 'Sales_Email'], r'^.{2,30}\@
.{2,63}\..{2,20}\Z'), |
47 ('/XML_DIZ_INFO/Company_Info/Support_Info/General_Phone', r'^\+{0,2}(([0-9#*()
-\/_] *){7,40})?\Z'), | 45 (['XML_DIZ_INFO', 'Company_Info', 'Support_Info', 'Support_Email'], r'^.{2,30}
\@.{2,63}\..{2,20}\Z'), |
48 ('/XML_DIZ_INFO/Company_Info/Support_Info/Fax_Phone', r'^\+{0,2}(([0-9#*()-\/_
] *){7,40})?\Z'), | 46 (['XML_DIZ_INFO', 'Company_Info', 'Support_Info', 'General_Email'], r'^.{2,30}
\@.{2,63}\..{2,20}\Z'), |
49 ('/XML_DIZ_INFO/Program_Info/Program_Name', r'^[^<\x09]{1,40}\Z'), | 47 (['XML_DIZ_INFO', 'Company_Info', 'Support_Info', 'Sales_Phone'], r'^\+{0,2}((
[0-9#*()-\/_] *){7,40})?\Z'), |
50 ('/XML_DIZ_INFO/Program_Info/Program_Version', r'^[a-zA-Z0-9_.\-]{1,15}\Z'), | 48 (['XML_DIZ_INFO', 'Company_Info', 'Support_Info', 'Support_Phone'], r'^\+{0,2}
(([0-9#*()-\/_] *){7,40})?\Z'), |
51 ('/XML_DIZ_INFO/Program_Info/Program_Release_Month', r'^(0[1-9]|1[0-2])\Z'), | 49 (['XML_DIZ_INFO', 'Company_Info', 'Support_Info', 'General_Phone'], r'^\+{0,2}
(([0-9#*()-\/_] *){7,40})?\Z'), |
52 ('/XML_DIZ_INFO/Program_Info/Program_Release_Day', r'^(0[1-9]|[12][0-9]|3[01])
\Z'), | 50 (['XML_DIZ_INFO', 'Company_Info', 'Support_Info', 'Fax_Phone'], r'^\+{0,2}(([0
-9#*()-\/_] *){7,40})?\Z'), |
53 ('/XML_DIZ_INFO/Program_Info/Program_Release_Year', r'^(19|20|21)[0-9]{2}\Z'), | 51 (['XML_DIZ_INFO', 'Company_Info', 'GooglePlusPage'], r'^((http|https):\/\/.{4,
120}|)\Z'), |
54 ('/XML_DIZ_INFO/Program_Info/Program_Cost_Dollars', r'^([0-9]+(\.[0-9]{2})?)?\
Z'), | 52 (['XML_DIZ_INFO', 'Company_Info', 'LinkedinPage'], r'^((http|https):\/\/.{4,12
0}|)\Z'), |
55 ('/XML_DIZ_INFO/Program_Info/Program_Cost_Other_Code', r'^(AED|AFN|ALL|AMD|ANG
|AOA|ARS|AUD|AWG|AZM|BAM|BBD|BDT|BGN|BHD|BIF|BMD|BND|BOB|BRL|BSD|BTN|BWP|BYR|BZD
|CAD|CDF|CHF|CLP|CNY|COP|COU|CRC|CSD|CZK|CUP|CVE|CYP|DJF|DKK|DOP|DZD|EEK|EGP|ERN
|ETB|EUR|FJD|FKP|GBP|GEL|GHC|GIP|GMD|GNF|GTQ|GYD|HKD|HNL|HRK|HTG|HUF|IDR|ILS|INR
|IQD|IRR|ISK|JMD|JOD|JPY|KES|KGS|KHR|KMF|KPW|KRW|KWD|KYD|KZT|LAK|LBP|LKR|LRD|LSL
|LTL|LVL|LYD|MAD|MDL|MGA|MKD|MMK|MNT|MOP|MRO|MTL|MUR|MVR|MWK|MXN|MYR|MZN|NAD|NGN
|NIO|NOK|NPR|NZD|OMR|PAB|PEN|PGK|PHP|PKR|PLN|PYG|QAR|RON|RUB|RWF|SAR|SBD|SCR|SDD
|SEK|SGD|SHP|SIT|SKK|SLL|SOS|SRD|STD|SYP|SZL|THB|TJS|TMM|TND|TOP|TRY|TTD|TWD|TZS
|UAH|UGX|USD|UYU|UZS|VEB|VND|VUV|WST|XAF|XCD|XOF|XPF|YER|ZAR|ZMK|ZWD)?\Z'), | 53 (['XML_DIZ_INFO', 'Company_Info', 'TwitterCompanyPage'], r'^((http|https):\/\/
.{4,120}|)\Z'), |
56 ('/XML_DIZ_INFO/Program_Info/Program_Cost_Other', r'^([0-9]+(\.[0-9]{2})?)?\Z'
), | 54 (['XML_DIZ_INFO', 'Company_Info', 'FacebookCompanyPage'], r'^((http|https):\/\
/.{4,120}|)\Z'), |
57 ('/XML_DIZ_INFO/Program_Info/Program_Type', r'^(Shareware|Freeware|Adware|Demo
|Commercial|Data Only)\Z'), | 55 (['XML_DIZ_INFO', 'Company_Info', 'CompanyStorePage'], r'^((http|https):\/\/.{
4,120}|)\Z'), |
58 ('/XML_DIZ_INFO/Program_Info/Program_Release_Status', r'^(Major Update|Minor U
pdate|New Release|Beta|Alpha|Media Only)\Z'), | 56 (['XML_DIZ_INFO', 'Program_Info', 'AppID'], r'^(app-[0-9a-z]{12}|)\Z'), |
59 ('/XML_DIZ_INFO/Program_Info/Program_Install_Support', r'^(Install and Uninsta
ll|Install Only|No InstallSupport|Uninstall Only)\Z'), | 57 (['XML_DIZ_INFO', 'Program_Info', 'Program_Name'], r'^[^<\x09]{1,40}\Z'), |
60 ('/XML_DIZ_INFO/Program_Info/Program_OS_Support', r'^((Android|BlackBerry|Hand
held\/Mobile Other|iPhone|iPod|iTouch|Java|Linux|Linux Console|Linux Gnome|Linux
GPL|Linux Open Source|Mac OS X|Mac Other|MS-DOS|Netware|OpenVMS|Palm|Pocket PC|
Symbian|Unix|Win2000|Win7 x32|Win7 x64|Win98|WinMobile|WinOther|WinServer|WinVis
ta|WinVista x64|WinXP|Other|Not Applicable)[, ]*)+\Z'), | 58 (['XML_DIZ_INFO', 'Program_Info', 'Program_Version'], r'^[a-zA-Z0-9_.\-]{1,15}
\Z'), |
61 ('/XML_DIZ_INFO/Program_Info/Program_Language', r'^(Abkhazian|Afar|Afrikaans|A
lbanian|Amharic|Arabic|Armenian|Assamese|Aymara|Azerbaijani|Bashkir|Basque|Benga
li|Bhutani|Bihari|Bislama|Breton|Bulgarian|Burmese|Byelorussian|Cambodian|Catala
n|Chinese|ChineseSimplified|ChineseTraditional|Corsican|Croatian|Czech|Danish|Du
tch|English|Esperanto|Estonian|Faeroese|Fiji|Finnish|French|Frisian|Gaelic|Galic
ian|Georgian|German|Greek|Greenlandic|Guarani|Gujarati|Hausa|Hebrew|Hindi|Hungar
ian|Icelandic|Indonesian|Interlingua|Interlingue|Inupiak|Irish|Italian|Japanese|
Javanese|Kannada|Kashmiri|Kazakh|Kinyarwanda|Kirghiz|Kirundi|Korean|Kurdish|Laot
hian|Latin|Latvian|Lingala|Lithuanian|Macedonian|Malagasy|Malay|Malayalam|Maltes
e|Maori|Marathi|Moldavian|Mongolian|Nauru|Nepali|Norwegian|Occitan|Oriya|Oromo|O
ther|Pashto|Persian|Polish|Portuguese|Punjabi|Quechua|Rhaeto-Romance|Romanian|Ru
ssian|Samoan|Sangro|Sanskrit|Serbian|Serbo-Croatian|Sesotho|Setswana|Shona|Sindh
i|Singhalese|Siswati|Slovak|Slovenian|Somali|Spanish|Sudanese|Swahili|Swedish|Ta
galog|Tajik|Tamil|Tatar|Telugu|Thai|Tibetan|Tigrinya|Tonga|Tsonga|Turkish|Turkme
n|Twi|Ukrainian|Urdu|Uzbek|Vietnamese|Volapuk|Welsh|Wolof|Xhosa|Yiddish|Yoruba|Z
ulu|,)+\Z'), | 59 (['XML_DIZ_INFO', 'Program_Info', 'Program_Release_Month'], r'^(0[1-9]|1[0-2])
\Z'), |
62 ('/XML_DIZ_INFO/Program_Info/File_Info/File_Size_Bytes', r'^[0-9]{3,16}\Z'), | 60 (['XML_DIZ_INFO', 'Program_Info', 'Program_Release_Day'], r'^(0[1-9]|[12][0-9]
|3[01])\Z'), |
63 ('/XML_DIZ_INFO/Program_Info/File_Info/File_Size_K', r'^[0-9.]{1,12}\Z'), | 61 (['XML_DIZ_INFO', 'Program_Info', 'Program_Release_Year'], r'^(19|20|21)[0-9]{
2}\Z'), |
64 ('/XML_DIZ_INFO/Program_Info/File_Info/File_Size_MB', r'^[0-9.]{1,8}\Z'), | 62 (['XML_DIZ_INFO', 'Program_Info', 'Program_Cost_Dollars'], r'^([0-9]+(\.[0-9]{
2})?)?\Z'), |
65 ('/XML_DIZ_INFO/Program_Info/Expire_Info/Has_Expire_Info', r'^[YyNn]\Z'), | 63 (['XML_DIZ_INFO', 'Program_Info', 'Program_Cost_Other_Code'], r'^(AED|AFN|ALL|
AMD|ANG|AOA|ARS|AUD|AWG|AZM|BAM|BBD|BDT|BGN|BHD|BIF|BMD|BND|BOB|BRL|BSD|BTN|BWP|
BYR|BZD|CAD|CDF|CHF|CLP|CNY|COP|COU|CRC|CSD|CZK|CUP|CVE|CYP|DJF|DKK|DOP|DZD|EEK|
EGP|ERN|ETB|EUR|FJD|FKP|GBP|GEL|GHC|GIP|GMD|GNF|GTQ|GYD|HKD|HNL|HRK|HTG|HUF|IDR|
ILS|INR|IQD|IRR|ISK|JMD|JOD|JPY|KES|KGS|KHR|KMF|KPW|KRW|KWD|KYD|KZT|LAK|LBP|LKR|
LRD|LSL|LTL|LVL|LYD|MAD|MDL|MGA|MKD|MMK|MNT|MOP|MRO|MTL|MUR|MVR|MWK|MXN|MYR|MZN|
NAD|NGN|NIO|NOK|NPR|NZD|OMR|PAB|PEN|PGK|PHP|PKR|PLN|PYG|QAR|RON|RUB|RWF|SAR|SBD|
SCR|SDD|SEK|SGD|SHP|SIT|SKK|SLL|SOS|SRD|STD|SYP|SZL|THB|TJS|TMM|TND|TOP|TRY|TTD|
TWD|TZS|UAH|UGX|USD|UYU|UZS|VEB|VND|VUV|WST|XAF|XCD|XOF|XPF|YER|ZAR|ZMK|ZWD)?\Z'
), |
66 ('/XML_DIZ_INFO/Program_Info/Expire_Info/Expire_Count', r'^[0-9]{0,15}\Z'), | 64 (['XML_DIZ_INFO', 'Program_Info', 'Program_Cost_Other'], r'^([0-9]+(\.[0-9]{2}
)?)?\Z'), |
67 ('/XML_DIZ_INFO/Program_Info/Expire_Info/Expire_Based_On', r'^(Days|Uses|Eithe
r\/Or)?\Z'), | 65 (['XML_DIZ_INFO', 'Program_Info', 'Program_Type'], r'^(Shareware|Freeware|Adwa
re|Demo|Commercial|Data Only)\Z'), |
68 ('/XML_DIZ_INFO/Program_Info/Expire_Info/Expire_Other_Info', r'^[^<\x09]{0,100
}\Z'), | 66 (['XML_DIZ_INFO', 'Program_Info', 'Program_Release_Status'], r'^(Major Update|
Minor Update|New Release|Beta|Alpha|Media Only)\Z'), |
69 ('/XML_DIZ_INFO/Program_Info/Expire_Info/Expire_Month', r'^(0[1-9]|1[0-2])?\Z'
), | 67 (['XML_DIZ_INFO', 'Program_Info', 'Program_Install_Support'], r'^(Install and
Uninstall|Install Only|No Install Support|Uninstall Only)\Z'), |
70 ('/XML_DIZ_INFO/Program_Info/Expire_Info/Expire_Day', r'^(0[1-9]|[12][0-9]|3[0
1])?\Z'), | 68 (['XML_DIZ_INFO', 'Program_Info', 'Program_OS_Support'], r'^((Android|BlackBer
ry|Handheld\/Mobile Other|iPhone|iPad|iPod|iTouch|Java|Linux|Linux Console|Linux
Gnome|Linux GPL|Linux Open Source|Mac OS X|Mac Other|MS-DOS|Netware|OpenVMS|Pal
m|Pocket PC|Symbian|Unix|Win2000|Win7 x32|Win7 x64|Win98|WinMobile|WinOther|WinS
erver|WinVista|WinVista x64|WinXP|Windows 8|Windows Phone 7|Windows Phone 8|Wind
ows RT|Other|Not Applicable)[, ]*)+\Z'), |
71 ('/XML_DIZ_INFO/Program_Info/Expire_Info/Expire_Year', r'^((19|20|21)[0-9]{2})
?\Z'), | 69 (['XML_DIZ_INFO', 'Program_Info', 'Program_Language'], r'^(Abkhazian|Afar|Afri
kaans|Albanian|Amharic|Arabic|Armenian|Assamese|Aymara|Azerbaijani|Bashkir|Basqu
e|Bengali|Bhutani|Bihari|Bislama|Breton|Bulgarian|Burmese|Byelorussian|Cambodian
|Catalan|Chinese|ChineseSimplified|ChineseTraditional|Corsican|Croatian|Czech|Da
nish|Dutch|English|Esperanto|Estonian|Faeroese|Fiji|Finnish|French|Frisian|Gaeli
c|Galician|Georgian|German|Greek|Greenlandic|Guarani|Gujarati|Hausa|Hebrew|Hindi
|Hungarian|Icelandic|Indonesian|Interlingua|Interlingue|Inupiak|Irish|Italian|Ja
panese|Javanese|Kannada|Kashmiri|Kazakh|Kinyarwanda|Kirghiz|Kirundi|Korean|Kurdi
sh|Laothian|Latin|Latvian|Lingala|Lithuanian|Macedonian|Malagasy|Malay|Malayalam
|Maltese|Maori|Marathi|Moldavian|Mongolian|Nauru|Nepali|Norwegian|Occitan|Oriya|
Oromo|Other|Pashto|Persian|Polish|Portuguese|Punjabi|Quechua|Rhaeto-Romance|Roma
nian|Russian|Samoan|Sangro|Sanskrit|Serbian|Serbo-Croatian|Sesotho|Setswana|Shon
a|Sindhi|Singhalese|Siswati|Slovak|Slovenian|Somali|Spanish|Sudanese|Swahili|Swe
dish|Tagalog|Tajik|Tamil|Tatar|Telugu|Thai|Tibetan|Tigrinya|Tonga|Tsonga|Turkish
|Turkmen|Twi|Ukrainian|Urdu|Uzbek|Vietnamese|Volapuk|Welsh|Wolof|Xhosa|Yiddish|Y
oruba|Zulu|,)+\Z'), |
72 ('/XML_DIZ_INFO/Program_Info/Program_Change_Info', r'^[^<\x09]{0,300}\Z'), | 70 (['XML_DIZ_INFO', 'Program_Info', 'File_Info', 'File_Size_Bytes'], r'^[0-9]{3,
16}\Z'), |
73 ('/XML_DIZ_INFO/Program_Info/Program_Category_Class', r'^(Audio & Multimedia::
Audio Encoders\/Decoders|Audio & Multimedia::Audio File Players|Audio & Multimed
ia::Audio File Recorders|Audio & Multimedia::CD Burners|Audio & Multimedia::CD P
layers|Audio & Multimedia::Multimedia Creation Tools|Audio & Multimedia::Music C
omposers|Audio & Multimedia::Other|Audio & Multimedia::Presentation Tools|Audio
& Multimedia::Rippers & Converters|Audio & Multimedia::Speech|Audio & Multimedia
::Video Tools|Business::Accounting & Finance|Business::Calculators & Converters|
Business::Databases & Tools|Business::Helpdesk & Remote PC|Business::Inventory &
Barcoding|Business::Investment Tools|Business::Math & Scientific Tools|Business
::Office Suites & Tools|Business::Other|Business::PIMS & Calendars|Business::Pro
ject Management|Business::Vertical Market Apps|Communications::Chat & Instant Me
ssaging|Communications::Dial Up & Connection Tools|Communications::E-Mail Client
s|Communications::E-Mail List Management|Communications::Fax Tools|Communication
s::Newsgroup Clients|Communications::Other Comms Tools|Communications::Other E-M
ail Tools|Communications::Pager Tools|Communications::Telephony|Communications::
Web\/Video Cams|Desktop::Clocks & Alarms|Desktop::Cursors & Fonts|Desktop::Icons
|Desktop::Other|Desktop::Screen Savers: Art|Desktop::Screen Savers: Cartoons|Des
ktop::Screen Savers: Nature|Desktop::Screen Savers: Other|Desktop::Screen Savers
: People|Desktop::Screen Savers: Science|Desktop::Screen Savers: Seasonal|Deskto
p::Screen Savers: Vehicles|Desktop::Themes & Wallpaper|Development::Active X|Dev
elopment::Basic, VB, VB DotNet|Development::C \/ C\+\+ \/ C\#|Development::Compi
lers & Interpreters|Development::Components & Libraries|Development::Debugging|D
evelopment::Delphi|Development ::Help Tools|Development::Install & Setup|Develop
ment::Management & Distribution|Development::Other|Development::Source Editors|E
ducation::Computer|Education::Dictionaries|Education::G eography|Education::Kids
|Education::Languages|Education::Mathema tics|Education::Other|Education::Refere
nce Tools|Education::Science|Education::Teaching & Training Tools|Games & Entert
ainment::Action|Games & Entertainment::Adventure & Roleplay|Games & Entertainmen
t::Arcade|Games & Entertainment::Board|Games & Entertainment::Card|Games & Enter
tainment::Casino & Gambling|Games & Entertainment::Kids|Games & Entertainment::O
nline Gaming|Games & Entertainment::Other|Games & Entertainment::Puzzle & Word G
ames|Games & Entertainment::Simulation|Games & Entertainment::Sports|Games & Ent
ertainment::Strategy & War Games|Games & Entertainment::Tools & Editors|Graphic
Apps::Animation Tools|Graphic Apps::CAD|Graphic Apps::Converters & Optimizers|Gr
aphic Apps::Editors|Graphic Apps::Font Tools|Graphic Apps::Gallery & Cataloging
Tools|Graphic Apps::Icon Tools|Graphic Apps::Other|Graphic Apps::Screen Capture|
Graphic Apps::Viewers|Home & Hobby::Astrology\/Biorhythms\/Mystic|Home & Hobby::
Astronomy|Home & Hobby::Cataloging|Home & Hobby::Food & Drink|Home & Hobby::Gene
alogy|Home & Hobby::Health & Nutrition|Home & Hobby::Other|Home & Hobby::Persona
l Finance|Home & Hobby::Personal Interest|Home & Hobby::Recreation|Home & Hobby:
:Religion|Network & Internet::Ad Blockers|Network & Internet::Browser Tools|Netw
ork & Internet::Browsers|Network & Internet::Download Managers|Network & Interne
t::File Sharing\/Peer to Peer|Network & Internet::FTP Clients|Network & Internet
::Network Monitoring|Network & Internet::Other|Network & Internet::Remote Comput
ing|Network & Internet::Search\/Lookup Tools|Network & Internet::Terminal & Teln
et Clients|Network & Internet::Timers & Time Synch|Network & Internet::Trace & P
ing Tools|Security & Privacy::Access Control|Security & Privacy::Anti-Spam & Ant
i-Spy Tools|Security & Privacy::Anti-Virus Tools|Security & Privacy::Covert Surv
eillance|Security & Privacy::Encryption Tools|Security & Privacy::Other|Security
& Privacy::Password Managers|Servers::Firewall & Proxy Servers|Servers::FTP Ser
vers|Servers::Mail Servers|Servers::News Servers|Servers::Other Server Applicati
ons|Servers::Telnet Servers|Servers::Web Servers|System Utilities::Automation To
ols|System Utilities::Backup & Restore|System Utilities::Benchmarking|System Uti
lities::Clipboard Tools|System Utilities::File & Disk Management|System Utilitie
s::File Compression|System Utilities::Launchers & Task Managers|System Utilities
::Other|System Utilities::Printer|System Utilities::Registry Tools|System Utilit
ies::Shell Tools|System Utilities::System Maintenance|System Utilities::Text\/Do
cument Editors|Web Development::ASP & PHP|Web Development::E-Commerce|Web Develo
pment::Flash Tools|Web Development::HTML Tools|Web Development::Java & JavaScrip
t|Web Development::Log Analysers|Web Development::Other|Web Development::Site Ad
ministration|Web Development::Wizards & Components|Web Development::XML\/CSS Too
ls)\Z'), | 71 (['XML_DIZ_INFO', 'Program_Info', 'File_Info', 'File_Size_K'], r'^[0-9.]{1,12}
\Z'), |
74 ('/XML_DIZ_INFO/Program_Info/Program_Specific_Category', r'^(Audio|Business|De
velopmentTools|Education|Games|Graphics|Home\/Hobby|Internet|Miscellaneous|Scree
n Savers|Utilities)?\Z'), | 72 (['XML_DIZ_INFO', 'Program_Info', 'File_Info', 'File_Size_MB'], r'^[0-9.]{1,8}
\Z'), |
75 ('/XML_DIZ_INFO/Program_Info/Program_System_Requirements', r'^[^<\x09]{0,100}\
Z'), | 73 (['XML_DIZ_INFO', 'Program_Info', 'Expire_Info', 'Has_Expire_Info'], r'^[YyNn]
\Z'), |
76 ('/XML_DIZ_INFO/Program_Descriptions/English/Keywords', r'^[^<\x09]{0,250}\Z')
, | 74 (['XML_DIZ_INFO', 'Program_Info', 'Expire_Info', 'Expire_Count'], r'^[0-9]{0,1
5}\Z'), |
77 ('/XML_DIZ_INFO/Program_Descriptions/English/Char_Desc_45', r'^[^<\x09\x0a\x0d
]{0,45}\Z'), | 75 (['XML_DIZ_INFO', 'Program_Info', 'Expire_Info', 'Expire_Based_On'], r'^(Days|
Uses|Either\/Or)?\Z'), |
78 ('/XML_DIZ_INFO/Program_Descriptions/English/Char_Desc_80', r'^[^<\x09\x0a\x0d
]{0,80}\Z'), | 76 (['XML_DIZ_INFO', 'Program_Info', 'Expire_Info', 'Expire_Other_Info'], r'^[^<\
x09]{0,100}\Z'), |
79 ('/XML_DIZ_INFO/Program_Descriptions/English/Char_Desc_250', r'^[^<\x09\x0a\x0
d]{0,250}\Z'), | 77 (['XML_DIZ_INFO', 'Program_Info', 'Expire_Info', 'Expire_Month'], r'^(0[1-9]|1
[0-2])?\Z'), |
80 ('/XML_DIZ_INFO/Program_Descriptions/English/Char_Desc_450', r'^[^<\x09\x0a\x0
d]{0,450}\Z'), | 78 (['XML_DIZ_INFO', 'Program_Info', 'Expire_Info', 'Expire_Day'], r'^(0[1-9]|[12
][0-9]|3[01])?\Z'), |
81 ('/XML_DIZ_INFO/Program_Descriptions/English/Char_Desc_2000', r'^[^<]{0,2000}\
Z'), | 79 (['XML_DIZ_INFO', 'Program_Info', 'Expire_Info', 'Expire_Year'], r'^((19|20|21
)[0-9]{2})?\Z'), |
82 ('/XML_DIZ_INFO/Web_Info/Application_URLs/Application_Info_URL', r'^http:\/\/.
{2,120}\Z'), | 80 (['XML_DIZ_INFO', 'Program_Info', 'Program_Change_Info'], r'^[^<\x09]{0,300}\Z
'), |
83 ('/XML_DIZ_INFO/Web_Info/Application_URLs/Application_Order_URL', r'^((http|ht
tps):\/\/.{2,120})?\Z'), | 81 (['XML_DIZ_INFO', 'Program_Info', 'Program_Category_Class'], r'^(Audio & M
ultimedia::Audio Encoders\/Decoders|Audio & Multimedia::Audio File Players|A
udio & Multimedia::Audio File Recorders|Audio & Multimedia::CD Burners|A
udio & Multimedia::CD Players|Audio & Multimedia::Multimedia Creation To
ols|Audio & Multimedia::Music Composers|Audio & Multimedia::Other|Audio
& Multimedia::Presentation Tools|Audio & Multimedia::Rippers & Conve
rters|Audio & Multimedia::Speech|Audio & Multimedia::Video Tools|Busines
s::Accounting & Finance|Business::Calculators & Converters|Business::Dat
abases & Tools|Business::Helpdesk & Remote PC|Business::Inventory &
Barcoding|Business::Investment Tools|Business::Math & Scientific Tools|Busin
ess::Office Suites & Tools|Business::Other|Business::PIMS & Calendars|Bu
siness::Project Management|Business::Vertical Market Apps|Communications::Chat &
amp; Instant Messaging|Communications::Dial Up & Connection Tools|Communicat
ions::E-Mail Clients|Communications::E-Mail List Management|Communications::Fax
Tools|Communications::Newsgroup Clients|Communications::Other Comms Tools|Commun
ications::Other E-Mail Tools|Communications::Pager Tools|Communications::Telepho
ny|Communications::Web\/Video Cams|Desktop::Clocks & Alarms|Desktop::Cursors
& Fonts|Desktop::Icons|Desktop::Other|Desktop::Screen Savers: Art|Desktop::
Screen Savers: Cartoons|Desktop::Screen Savers: Nature|Desktop::Screen Savers: O
ther|Desktop::Screen Savers: People|Desktop::Screen Savers: Science|Desktop::Scr
een Savers: Seasonal|Desktop::Screen Savers: Vehicles|Desktop::Themes & Wall
paper|Development::Active X|Development::Basic, VB, VB DotNet|Development::C \/
C\+\+ \/ C\#|Development::Compilers & Interpreters|Development::Components &
amp; Libraries|Development::Debugging|Development::Delphi|Development::Help Tool
s|Development::Install & Setup|Development::Management & Distribution|De
velopment::Other|Development::Source Editors|Education::Computer|Education::Dict
ionaries|Education::Geography|Education::Kids|Education::Languages|Education::Ma
thematics|Education::Other|Education::Reference Tools|Education::Science|Educati
on::Teaching & Training Tools|Games & Entertainment::Action|Games &
Entertainment::Adventure & Roleplay|Games & Entertainment::Arcade|Games
& Entertainment::Board|Games & Entertainment::Card|Games & Entertain
ment::Casino & Gambling|Games & Entertainment::Kids|Games & Entertai
nment::Online Gaming|Games & Entertainment::Other|Games & Entertainment:
:Puzzle & Word Games|Games & Entertainment::Simulation|Games & Enter
tainment::Sports|Games & Entertainment::Strategy & War Games|Games &
Entertainment::Tools & Editors|Graphic Apps::Animation Tools|Graphic Apps::
CAD|Graphic Apps::Converters & Optimizers|Graphic Apps::Editors|Graphic Apps
::Font Tools|Graphic Apps::Gallery & Cataloging Tools|Graphic Apps::Icon Too
ls|Graphic Apps::Other|Graphic Apps::Screen Capture|Graphic Apps::Viewers|Home &
amp; Hobby::Astrology\/Biorhythms\/Mystic|Home & Hobby::Astronomy|Home &
Hobby::Cataloging|Home & Hobby::Food & Drink|Home & Hobby::Genealog
y|Home & Hobby::Health & Nutrition|Home & Hobby::Other|Home & Ho
bby::Personal Finance|Home & Hobby::Personal Interest|Home & Hobby::Recr
eation|Home & Hobby::Religion|Network & Internet::Ad Blockers|Network &a
mp; Internet::Browser Tools|Network & Internet::Browsers|Network & Inter
net::Download Managers|Network & Internet::File Sharing\/Peer to Peer|Networ
k & Internet::FTP Clients|Network & Internet::Network Monitoring|Network
& Internet::Other|Network & Internet::Remote Computing|Network & In
ternet::Search\/Lookup Tools|Network & Internet::Terminal & Telnet Clien
ts|Network & Internet::Timers & Time Synch|Network & Internet::Trace
& Ping Tools|Security & Privacy::Access Control|Security & Privacy:
:Anti-Spam & Anti-Spy Tools|Security & Privacy::Anti-Virus Tools|Securit
y & Privacy::Covert Surveillance|Security & Privacy::Encryption Tools|Se
curity & Privacy::Other|Security & Privacy::Password Managers|Servers::F
irewall & Proxy Servers|Servers::FTP Servers|Servers::Mail Servers|Servers::
News Servers|Servers::Other Server Applications|Servers::Telnet Servers|Servers:
:Web Servers|System Utilities::Automation Tools|System Utilities::Backup & R
estore|System Utilities::Benchmarking|System Utilities::Clipboard Tools|System U
tilities::File & Disk Management|System Utilities::File Compression|System U
tilities::Launchers & Task Managers|System Utilities::Other|System Utilities
::Printer|System Utilities::Registry Tools|System Utilities::Shell Tools|System
Utilities::System Maintenance|System Utilities::Text\/Document Editors|Web Devel
opment::ASP & PHP|Web Development::E-Commerce|Web Development::Flash Tools|W
eb Development::HTML Tools|Web Development::Java & JavaScript|Web Developmen
t::Log Analysers|Web Development::Other|Web Development::Site Administration|Web
Development::Wizards & Components|Web Development::XML\/CSS Tools)\Z'), |
84 ('/XML_DIZ_INFO/Web_Info/Application_URLs/Application_Screenshot_URL', r'^http
:\/\/.{2,120}\.(gif|jpg|png)\Z'), | 82 (['XML_DIZ_INFO', 'Program_Info', 'Program_Specific_Category'], r'^(Audio|Busi
ness|Development Tools|Education|Games|Graphics|Home\/Hobby|Internet|Miscellaneo
us|Screen Savers|Utilities)?\Z'), |
85 ('/XML_DIZ_INFO/Web_Info/Application_URLs/Application_Icon_URL', r'^http:\/\/.
{2,120}\.(gif|jpg|png)\Z'), | 83 (['XML_DIZ_INFO', 'Program_Info', 'Program_System_Requirements'], r'^[^<\x09]{
0,100}\Z'), |
86 ('/XML_DIZ_INFO/Web_Info/Application_URLs/Application_XML_File_URL', r'^http:\
/\/.{2,120}\.(xml|cgi|php|asp)\Z'), | 84 (['XML_DIZ_INFO', 'Program_Info', 'FacebookProductPage'], r'^((http|https):\/\
/.{4,120}|)\Z'), |
87 ('/XML_DIZ_INFO/Web_Info/Download_URLs/Primary_Download_URL', r'^(http|ftp):\/
\/.{2,120}\Z'), | 85 (['XML_DIZ_INFO', 'Program_Info', 'GooglePlusProductPage'], r'^((http|https):\
/\/.{4,120}|)\Z'), |
88 ('/XML_DIZ_INFO/Web_Info/Download_URLs/Secondary_Download_URL', r'^((http|ftp)
:\/\/.{2,120})?\Z'), | 86 (['XML_DIZ_INFO', 'Program_Descriptions', 'English', 'Keywords'], r'^[^<\x09]{
0,250}\Z'), |
89 ('/XML_DIZ_INFO/Web_Info/Download_URLs/Additional_Download_URL_1', r'^((http|f
tp):\/\/.{2,120})?\Z'), | 87 (['XML_DIZ_INFO', 'Program_Descriptions', 'English', 'Char_Desc_45'], r'^[^<\x
09\x0a\x0d]{0,45}\Z'), |
90 ('/XML_DIZ_INFO/Web_Info/Download_URLs/Additional_Download_URL_2', r'^((http|f
tp):\/\/.{2,120})?\Z'), | 88 (['XML_DIZ_INFO', 'Program_Descriptions', 'English', 'Char_Desc_80'], r'^[^<\x
09\x0a\x0d]{0,80}\Z'), |
91 ('/XML_DIZ_INFO/Permissions/Distribution_Permissions', r'^[^<]{0,2000}\Z'), | 89 (['XML_DIZ_INFO', 'Program_Descriptions', 'English', 'Char_Desc_250'], r'^[^<\
x09\x0a\x0d]{0,250}\Z'), |
92 ('/XML_DIZ_INFO/Permissions/EULA', r'^[^<]{0,20000}\Z'), | 90 (['XML_DIZ_INFO', 'Program_Descriptions', 'English', 'Char_Desc_450'], r'^[^<\
x09\x0a\x0d]{0,450}\Z'), |
| 91 (['XML_DIZ_INFO', 'Program_Descriptions', 'English', 'Char_Desc_2000'], r'^[^<
]{0,2000}\Z'), |
| 92 (['XML_DIZ_INFO', 'Web_Info', 'Application_URLs', 'Application_Info_URL'], r'^
(http|https):\/\/.{2,120}\Z'), |
| 93 (['XML_DIZ_INFO', 'Web_Info', 'Application_URLs', 'Application_Order_URL'], r'
^((http|https):\/\/.{2,120})?\Z'), |
| 94 (['XML_DIZ_INFO', 'Web_Info', 'Application_URLs', 'Application_Screenshot_URL'
], r'^(http|https):\/\/.{2,120}\.(gif|jpg|png)\Z'), |
| 95 (['XML_DIZ_INFO', 'Web_Info', 'Application_URLs', 'Application_Icon_URL'], r'^
(http|https):\/\/.{2,120}\.(gif|jpg|png)\Z'), |
| 96 (['XML_DIZ_INFO', 'Web_Info', 'Application_URLs', 'Application_XML_File_URL'],
r'^(http|https):\/\/.{2,120}\.(xml|cgi|php|asp)\Z'), |
| 97 (['XML_DIZ_INFO', 'Web_Info', 'Application_URLs', 'Video_Link_1_URL'], r'^((ht
tp|https):\/\/.{4,120}|)\Z'), |
| 98 (['XML_DIZ_INFO', 'Web_Info', 'Application_URLs', 'Video_Link_2_URL'], r'^((ht
tp|https):\/\/.{4,120}|)\Z'), |
| 99 (['XML_DIZ_INFO', 'Web_Info', 'Download_URLs', 'Primary_Download_URL'], r'^(ht
tp|https|ftp):\/\/.{2,120}\Z'), |
| 100 (['XML_DIZ_INFO', 'Web_Info', 'Download_URLs', 'Secondary_Download_URL'], r'^(
(http|https|ftp):\/\/.{2,120})?\Z'), |
| 101 (['XML_DIZ_INFO', 'Web_Info', 'Download_URLs', 'Additional_Download_URL_1'], r
'^((http|https|ftp):\/\/.{2,120})?\Z'), |
| 102 (['XML_DIZ_INFO', 'Web_Info', 'Download_URLs', 'Additional_Download_URL_2'], r
'^((http|https|ftp):\/\/.{2,120})?\Z'), |
| 103 (['XML_DIZ_INFO', 'Permissions', 'Distribution_Permissions'], r'^[^<]{0,2000}\
Z'), |
| 104 (['XML_DIZ_INFO', 'Permissions', 'EULA'], r'^[^<]{0,20000}\Z'), |
| 105 (['XML_DIZ_INFO', 'Press_Release', 'Keywords'], r'^[^<\x09\x0a\x0d]{0,250}\Z')
, |
| 106 (['XML_DIZ_INFO', 'Press_Release', 'Headline'], r'^([^<\x09\x0a\x0d]{20,100}|)
\Z'), |
| 107 (['XML_DIZ_INFO', 'Press_Release', 'Summary'], r'^([^<\x09\x0a\x0d]{20,250}|)\
Z'), |
| 108 (['XML_DIZ_INFO', 'Press_Release', 'Press_Release'], r'^[^<\x09\x0a\x0d]{0,300
0}\Z'), |
| 109 (['XML_DIZ_INFO', 'Press_Release', 'Press_Release_Plain'], r'^[^<\x09\x0a\x0d]
{0,3000}\Z'), |
| 110 (['XML_DIZ_INFO', 'Press_Release', 'Related_URL'], r'^((http|https):\/\/.{0,10
0}|)\Z'), |
| 111 (['XML_DIZ_INFO', 'NewsFeed', 'NewsFeed_Feed_URL'], r'^((http|https):\/\/.{0,1
00}|)\Z'), |
| 112 (['XML_DIZ_INFO', 'NewsFeed', 'NewsFeed_Type'], r'^(RSS 0\.90|RSS 0\.91|RSS 0\
.92|RSS 0\.93|RSS 0\.94|RSS 1\.0|RSS 2\.0|Atom 0\.3|Atom 1\.0|)\Z'), |
| 113 (['XML_DIZ_INFO', 'NewsFeed', 'NewsFeed_Title'], r'^[^<\x09\x0a\x0d]{0,60}\Z')
, |
| 114 (['XML_DIZ_INFO', 'NewsFeed', 'NewsFeed_Keywords'], r'^[^<\x09]{0,250}\Z'), |
| 115 (['XML_DIZ_INFO', 'NewsFeed', 'NewsFeed_Description_70'], r'^[^<\x09\x0a\x0d]{
0,70}\Z'), |
| 116 (['XML_DIZ_INFO', 'NewsFeed', 'NewsFeed_Description_250'], r'^[^<\x09\x0a\x0d]
{0,250}\Z'), |
| 117 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_ShareIt_Order_Page'], None), |
| 118 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_ShareIt_Vendor_ID'], None), |
| 119 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_ShareIt_Product_ID'], None), |
| 120 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_ShareIt_Maximum_Commission_Rate'],
None), |
| 121 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_PayPro_Order_Page'], None), |
| 122 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_PayPro_Vendor_ID'], None), |
| 123 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_PayPro_Product_ID'], None), |
| 124 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_PayPro_Maximum_Commission_Rate'],
None), |
| 125 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_Avangate_Order_Page'], None), |
| 126 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_Avangate_Vendor_ID'], None), |
| 127 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_Avangate_Product_ID'], None), |
| 128 (['XML_DIZ_INFO', 'Affiliates', 'Affiliates_Avangate_Maximum_Commission_Rate']
, None), |
| 129 (['XML_DIZ_INFO', 'ASP', 'ASP_Member'], r'^[YyNn]\Z'), |
| 130 (['XML_DIZ_INFO', 'ASP', 'ASP_Member_Number'], None), |
93 ] | 131 ] |
94 | 132 |
| 133 def validate_fields(fields, nodes, filename): |
| 134 expected_nodes = set() |
| 135 |
| 136 for node_name, fields in itertools.groupby(fields, lambda (path, regex): path[
0]): |
| 137 expected_nodes.add(node_name) |
| 138 |
| 139 regex = None |
| 140 nested_fields = [] |
| 141 for path, regex_ in fields: |
| 142 if path == [node_name]: |
| 143 regex = regex_ |
| 144 else: |
| 145 nested_fields.append((path[1:], regex_)) |
| 146 |
| 147 found = False |
| 148 for node in nodes: |
| 149 if node.nodeName == node_name: |
| 150 if found: |
| 151 warnings.warn('invalid PAD file (duplicate node)\n' |
| 152 'filename: %s\n' |
| 153 'node: %s' % (filename, node_name)) |
| 154 |
| 155 if regex: |
| 156 value = ''.join(child.toxml() for child in node.childNodes) |
| 157 if not re.match(regex, value): |
| 158 warnings.warn('invalid PAD file (invalid value)\n' |
| 159 'filename: %s\n' |
| 160 'node: %s\n' |
| 161 'value: %s\n' |
| 162 'regex: %s' % (filename, node_name, value, regex)) |
| 163 |
| 164 if nested_fields: |
| 165 validate_fields(nested_fields, node.childNodes, filename) |
| 166 |
| 167 found = True |
| 168 |
| 169 if not found: |
| 170 if regex and not re.match(regex, ''): |
| 171 warnings.warn('invalid PAD file (missing node)\n' |
| 172 'filename: %s\n' |
| 173 'node: %s' % (filename, node_name)) |
| 174 |
| 175 validate_fields(nested_fields, [], filename) |
| 176 |
| 177 for node in nodes: |
| 178 if node.nodeType == node.COMMENT_NODE: |
| 179 continue |
| 180 if node.nodeType == node.TEXT_NODE and node.nodeValue.strip() == '': |
| 181 continue |
| 182 |
| 183 if node.nodeName not in expected_nodes: |
| 184 warnings.warn('invalid PAD file (unexpected node)\n' |
| 185 'filename: %s\n' |
| 186 'node: %s' % (filename, node.nodeName)) |
| 187 |
95 def validate_pad(pad, filename): | 188 def validate_pad(pad, filename): |
96 if not libxml2: | 189 validate_fields(FIELDS, minidom.parseString(pad).childNodes, filename) |
97 return | 190 |
98 | 191 def print_fields(): |
99 doc = libxml2.parseDoc(pad) | 192 doc = minidom.parse(urllib2.urlopen('http://repository.appvisor.com/padspec/fi
les/padspec40.xml')) |
100 xpctx = doc.xpathNewContext() | 193 |
101 | 194 print '[' |
102 for xpath, regex in FIELDS: | 195 for field in doc.getElementsByTagName('Field'): |
103 nodes = xpctx.xpathEval(xpath) | 196 path, regex = [ |
104 | 197 ''.join(node.nodeValue for node in field.getElementsByTagName(name)[0].chi
ldNodes) |
105 if not nodes: | 198 for name in ('Path', 'RegEx') |
106 warnings.warn('invalid PAD file (missing node)\n' | 199 ] |
107 'filename: %s\n' | 200 print ' (%r, %s),' % (str(path).split('/'), "r'%s'" % regex if regex else '
None') |
108 'XPath: %s' % (filename, xpath)) | 201 print ']' |
109 else: | 202 |
110 value = nodes[0].getContent() | 203 if __name__ == '__main__': |
111 | 204 print_fields() |
112 if not re.match(regex, value): | |
113 warnings.warn('invalid PAD file (invalid value)\n' | |
114 'filename: %s\n' | |
115 'XPath: %s\n' | |
116 'value: %s\n' | |
117 'regex: %s' % (filename, xpath, value, regex)) | |
LEFT | RIGHT |