a
#.c @ sr d Z ddlmZ ddlZddlZddlmZmZ ddlm Z ddl
m Z m Z G dd d e e Z
G d
d d e ZdS ) zGTests to ensure that the html.parser tree builder generates good
trees. ) set_traceN)HTMLParserTreeBuilderXMLParsedAsHTMLWarning)BeautifulSoupHTMLParser )SoupTestHTMLTreeBuilderSmokeTestc @ sP e Zd ZeZdd Zdd Zdd Zdd Zd d
Z d d Z
d
d Z dd Z dS )TestHTMLParserTreeBuilderc C s d S N selfr r S/var/www/brookimports/venv/lib/python3.9/site-packages/bs4/tests/test_htmlparser.py test_namespaced_system_doctype s z8TestHTMLParserTreeBuilder.test_namespaced_system_doctypec C s d S r
r r r r r test_namespaced_public_doctype s z8TestHTMLParserTreeBuilder.test_namespaced_public_doctypec C s: | d}t|d}t|}t|jt|js6J dS )zfUnlike most tree builders, HTMLParserTreeBuilder and will
be restored after pickling.
z
foo N)souppickledumpsloads
isinstanceZbuildertype)r
treeZdumpedZloadedr r r test_builder_is_pickled s
z1TestHTMLParserTreeBuilder.test_builder_is_pickledc C s | dd | dd d S )Nz
z
z Z assert_soupr r r r )test_redundant_empty_element_closing_tags# s zCTestHTMLParserTreeBuilder.test_redundant_empty_element_closing_tagsc C s | dd d S )Nz
foo &# barzfoo &# barr r r r r test_empty_element' s z,TestHTMLParserTreeBuilder.test_empty_elementc C sz d}| |}d|jjks J d|jjks.J d|jdjksDJ | j |dd}d|jjjksdJ d|jjjksvJ d S )Nz=
text
r
sourcelineF)Zstore_line_numbers sourcepos)r pr r findname)r
markupr r r r test_tracking_line_numbers, s
z4TestHTMLParserTreeBuilder.test_tracking_line_numbersc s d }d|jd ks J dg|jd ks4J d|jd ksFJ fdd}|d d |tjd |tjd d
d }||g d d S )
Nz;url3hrefclsclassidc sJ j | d}||jd ks J dg|jd ks4J d|jd ksFJ d S )N)on_duplicate_attributer' r( r) r* )r a)r+ expectedr r$ r
r r assert_attributeH s zOTestHTMLParserTreeBuilder.test_on_duplicate_attribute..assert_attributeurl1c S s. t | | ts | | g| |< | | | d S r
)r listappend)attrskeyvaluer r r
accumulateX s zITestHTMLParserTreeBuilder.test_on_duplicate_attribute..accumulate)r0 url2r& )r r, r ZREPLACEZIGNORE)r
r r/ r6 r r. r test_on_duplicate_attribute: s
z5TestHTMLParserTreeBuilder.test_on_duplicate_attributec C sj dD ]`\}}}d| }| |j}| }d|d }||ksDJ |jdd}d| }||ksJ qd S )N))z⇄u ⇄s ⇄)z⊧u ⊧s ⊧)z𝔑u 𝔑s 𝔑)z≧̸u ≧̸s ≧̸)z¬ ¬s ¬)z⫬u ⫬s ⫬)z"" ")z∴ ∴ ∴)z ∴r< r= )z ∴r< r= )zfjfjs fj)z⊔u ⊔s ⊔)z⊔︀u ⊔︀s ⊔︀)z'' ')z|| |z
%s
s
%s
utf8html) formatter)r divencode) r
Z
input_elementZoutput_unicodeZoutput_elementr$ rF Zwithout_elementexpectZ with_elementr r r test_html5_attributes^ s z/TestHTMLParserTreeBuilder.test_html5_attributesN)
__name__
__module__ __qualname__r Zdefault_builderr r r r r r% r8 rI r r r r r s $r c @ s e Zd Zdd ZdS )TestHTMLParserSubclassc C sX t }tjdd}|d W d n1 s20 Y |\}dt|jksTJ dS )zlVerify that our HTMLParser subclass implements error() in a way
that doesn't cause a crash.
T)recordz don't crashN)r warningscatch_warningserrorstrmessage)r
parserZwarnswarningr r r
test_error s
(z!TestHTMLParserSubclass.test_errorN)rJ rK rL rV r r r r rM ~ s rM )__doc__pdbr r rO Z bs4.builderr r Zbs4.builder._htmlparserr r r r r rM r r r r s p