[gentoo-commits] repo/gentoo:master commit in: dev-python/nltk/files/, dev-python/nltk/

["Michał Górny" <mgorny@gentoo.org>]

commit:     2140875c4dafa046e8a658bde1ea0c2defc34faf
Author:     Michał Górny <mgorny <AT> gentoo <DOT> org>
AuthorDate: Fri Nov 26 20:27:42 2021 +0000
Commit:     Michał Górny <mgorny <AT> gentoo <DOT> org>
CommitDate: Fri Nov 26 20:30:17 2021 +0000
URL:        https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=2140875c

dev-python/nltk: Fix test failure

Closes: https://bugs.gentoo.org/819921
Signed-off-by: Michał Górny <mgorny <AT> gentoo.org>

 dev-python/nltk/files/nltk-3.6.5-test.patch | 28 ++++++++++++++++++++++++++++
 dev-python/nltk/nltk-3.6.5.ebuild           |  7 ++++---
 2 files changed, 32 insertions(+), 3 deletions(-)

diff --git a/dev-python/nltk/files/nltk-3.6.5-test.patch b/dev-python/nltk/files/nltk-3.6.5-test.patch
new file mode 100644
index 000000000000..ea4d9512fa74
--- /dev/null
+++ b/dev-python/nltk/files/nltk-3.6.5-test.patch
@@ -0,0 +1,28 @@
+From 9502cb3b5e43a787a16bc2f63ec34c69f9b151c6 Mon Sep 17 00:00:00 2001
+From: Tom Aarsen <Cubiegamedev@gmail.com>
+Date: Tue, 19 Oct 2021 16:49:36 +0200
+Subject: [PATCH] Skip doctest for printing out stopword languages
+
+People with outdated nltk_data will fail these tests, and this test is very vulnerable to updates in nltk_data
+---
+ nltk/test/corpus.doctest | 4 ++--
+ 1 file changed, 2 insertions(+), 2 deletions(-)
+
+diff --git a/nltk/test/corpus.doctest b/nltk/test/corpus.doctest
+index 536ef452f..47a6fea2f 100644
+--- a/nltk/test/corpus.doctest
++++ b/nltk/test/corpus.doctest
+@@ -385,8 +385,8 @@ examples illustrate the use of the wordlist corpora:
+     >>> words.words('en')
+     ['A', 'a', 'aa', 'aal', 'aalii', 'aam', 'Aani', 'aardvark', 'aardwolf', ...]
+ 
+-    >>> stopwords.fileids()
+-    ['arabic', 'azerbaijani', 'danish', 'dutch', 'english', 'finnish', 'french', ...]
++    >>> stopwords.fileids() # doctest: +SKIP
++    ['arabic', 'azerbaijani', 'bengali', 'danish', 'dutch', 'english', 'finnish', 'french', ...]
+     >>> sorted(stopwords.words('portuguese'))
+     ['a', 'ao', 'aos', 'aquela', 'aquelas', 'aquele', 'aqueles', ...]
+     >>> names.fileids()
+-- 
+2.34.1
+

diff --git a/dev-python/nltk/nltk-3.6.5.ebuild b/dev-python/nltk/nltk-3.6.5.ebuild
index 11dd1a3283d2..b706aa229c71 100644
--- a/dev-python/nltk/nltk-3.6.5.ebuild
+++ b/dev-python/nltk/nltk-3.6.5.ebuild
@@ -39,14 +39,15 @@ PDEPEND="dev-python/nltk-data"
 distutils_enable_tests pytest
 
 src_prepare() {
+	local PATCHES=(
+		"${FILESDIR}"/${P}-test.patch
+	)
+
 	# requires unpackaged pycrfsuite
 	sed -i -e '/>>>/s@$@ # doctest: +SKIP@' nltk/tag/crf.py || die
 	# replace fetching from network with duplicate file URL
 	sed -e 's@https://raw.githubusercontent.com/nltk/nltk/develop/nltk/test/toy.cfg@nltk:grammars/sample_grammars/toy.cfg@' \
 		-i nltk/test/data.doctest || die
-	# requires X and hangs in Xvfb
-	sed -e 's:test_plot:_&:' \
-		-i nltk/test/unit/test_cfd_mutation.py || die
 
 	distutils-r1_src_prepare
 }