similarity index 86%
rename from meta-python/recipes-devtools/python/python3-dateparser_1.5.0.bb
rename to meta-python/recipes-devtools/python/python3-dateparser_1.5.1.bb
@@ -3,7 +3,7 @@ HOMEPAGE = "https://github.com/scrapinghub/dateparser"
LICENSE = "BSD-3-Clause"
LIC_FILES_CHKSUM = "file://LICENSE;md5=3d3ed25571191e7aa3f55d0a6efe0051"
-SRC_URI[sha256sum] = "08d62f2c00d3a1e806fc59d40c79e7554e31d84d3b8c6145820af051c92e7292"
+SRC_URI[sha256sum] = "1c3d4a568b4f4d615adfc2945b54596adffc14f955c9b435030b5236226a4366"
inherit pypi python_setuptools_build_meta
Upgrade to release 1.5.1: Security fixes: - Prevent quadratic regex matching on long whitespace runs in the Jalali calendar (GHSA-rq5p-ww4m-666c) - Prevent quadratic parsing in search_dates() on a long run of skipped words after a number (GHSA-c83m-fh72-x3jp) - Prevent unbounded recursion and quadratic parsing in search_dates() on many quantities written in words (GHSA-8hqv-h88c-f367) - Prevent quadratic time in search_dates() on a long text whose words and Unicode-normalized words differ in number (GHSA-wr9g-q4w6-fp42) - Prevent quadratic parsing in parse() on a long run of separators followed by skipped words (GHSA-5g92-f88x-gwj5) New features: - Support "less than a minute ago" in Russian, which was not parse Fixes: - search_dates() looks for the words of a quantity written in words among at most 16 words, so a longer quantity can give a different result - search_dates() yields at most eight dates from a run of more than eight quantities written in words, where it yielded all of them before - Keep a full stop that belongs to a date inside the date in search_dates(), so "Sept. 5, 2020" is no longer split at the full stop - Use the reference date's year when a custom format has no year, so PREFER_DAY_OF_MONTH="last" returns February 29 in a leap year instead of February 28, and skip invalid leap days instead of raising ValueError; with PREFER_DAY_OF_MONTH or PREFER_MONTH_OF_YEAR set to "current", the day or month of RELATIVE_BASE is used when it is set - Leave out a time span that does not fit in the years 1 to 9999, instead of raising ValueError in search_dates() with RETURN_TIME_SPAN, so the other dates of the text are kept - Treat digits that the regular expression module matches but int() rejects as not a number, instead of raising ValueError in parse() and search_dates() - Do not raise KeyError for a month or weekday name in date_formats that is spelled with a letter the regular expression module equates with a letter of the name when ignoring case, such as the long s for "s"; the date is read as that name - Stop the Vietnamese year simplification from eating the month name "năm", so a month name such as "Tháng năm" (May) followed by a year parses - Fix parsing time-only strings such as "3pm" when RELATIVE_BASE is timezone-aware, which returned None Cleanups and internal improvements: - Stop a test from depending on the current year - Move the pytest and coverage settings into pyproject.toml, which requires pytest 9 or later for the tests Signed-off-by: Leon Anavi <leon.anavi@konsulko.com> --- ...{python3-dateparser_1.5.0.bb => python3-dateparser_1.5.1.bb} | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) rename meta-python/recipes-devtools/python/{python3-dateparser_1.5.0.bb => python3-dateparser_1.5.1.bb} (86%)