Consider renaming port to devel/py-jinja as jinja2 is now version 3.x and devel/py-jinja is devel/py-Jinja expired on 2017-03-03 Changes Version 3.1.2 Released 2022-04-28 Add parameters to Environment.overlay to match __init__. #1645 Handle race condition in FileSystemBytecodeCache. #1654 Version 3.1.1 Released 2022-03-25 The template filename on Windows uses the primary path separator. #1637 Version 3.1.0 Released 2022-03-24 Drop support for Python 3.6. #1534 Remove previously deprecated code. #1544 WithExtension and AutoEscapeExtension are built-in now. contextfilter and contextfunction are replaced by pass_context. evalcontextfilter and evalcontextfunction are replaced by pass_eval_context. environmentfilter and environmentfunction are replaced by pass_environment. Markup and escape should be imported from MarkupSafe. Compiled templates from very old Jinja versions may need to be recompiled. Legacy resolve mode for Context subclasses is no longer supported. Override resolve_or_missing instead of resolve. unicode_urlencode is renamed to url_quote. Add support for native types in macros. #1510 The {% trans %} tag can use pgettext and npgettext by passing a context string as the first token in the tag, like {% trans "title" %}. #1430 Update valid identifier characters from Python 3.6 to 3.7. #1571 Filters and tests decorated with @async_variant are pickleable. #1612 Add items filter. #1561 Subscriptions ([0], etc.) can be used after filters, tests, and calls when the environment is in async mode. #1573 The groupby filter is case-insensitive by default, matching other comparison filters. Added the case_sensitive parameter to control this. #1463 Windows drive-relative path segments in template names will not result in FileSystemLoader and PackageLoader loading from drive-relative paths. #1621 Version 3.0.3 Released 2021-11-09 Fix traceback rewriting internals for Python 3.10 and 3.11. #1535 Fix how the native environment treats leading and trailing spaces when parsing values on Python 3.10. #1537 Improve async performance by avoiding checks for common types. #1514 Revert change to hash(Node) behavior. Nodes are hashed by id again #1521 PackageLoader works when the package is a single module file. #1512 Version 3.0.2 Released 2021-10-04 Fix a loop scoping bug that caused assignments in nested loops to still be referenced outside of it. #1427 Make compile_templates deterministic for filter and import names. #1452, #1453 Revert an unintended change that caused Undefined to act like StrictUndefined for the in operator. #1448 Imported macros have access to the current template globals in async environments. #1494 PackageLoader will not include a current directory (.) path segment. This allows loading templates from the root of a zip import. #1467
Created attachment 236211 [details] git diff Update to 3.1.2, fixed COMMENT
Created attachment 236212 [details] portlint
Created attachment 236213 [details] poudriere testport
This also fixes using (complex) macros with Ansible as described in https://github.com/ansible/ansible/issues/77272 With 3.0.1 Jinja2 $ ansible localhost -m debug -a "msg='{% macro foo(v) %}{{ v }}{% endmacro %} {{ foo(1) }}'" [WARNING]: No inventory was parsed, only implicit localhost is available localhost | FAILED! => { "msg": "Unexpected templating type error occurred on ({% macro foo(v) %}{{ v }}{% endmacro %} {{ foo(1) }}): sequence item 0: expected str instance, int found" } $ pkg info -I py39-{ansible,Jinja2} py39-ansible-6.1.0 Radically simple IT automation py39-Jinja2-3.0.1 Fast and easy to use stand-alone template engine With 3.1.2 Jinja2 $ ansible localhost -m debug -a "msg='{% macro foo(v) %}{{ v }}{% endmacro %} {{ foo(1) }}'" [WARNING]: No inventory was parsed, only implicit localhost is available localhost | SUCCESS => { "msg": " 1" } $ pkg info -I py39-{ansible,Jinja2} py39-ansible-6.1.0 Radically simple IT automation py39-Jinja2-3.1.2 Fast and easy to use stand-alone template engine
If the two cents could be of any interest, a newer Jinja might be needed for devel/py-jupyterlab. Great to see if someone's already worked out a patch for it
Port updated. Thanks
A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/ports/commit/?id=23cc43859179b1bb1b6932ccee2559ed5f66a968 commit 23cc43859179b1bb1b6932ccee2559ed5f66a968 Author: Yuri Victorovich <yuri@FreeBSD.org> AuthorDate: 2022-11-08 15:03:00 +0000 Commit: Yuri Victorovich <yuri@FreeBSD.org> CommitDate: 2022-11-09 00:21:12 +0000 devel/py-Jinja2: Update 3.0.1 -> 3.1.2; Fix test target PR: 260440 264990 Approved by: nivit@ (maintainer's timoeut on 260440; 10.5 months) Submitted by: brd@, p5B2EA84B3@t-online.de devel/py-Jinja2/Makefile | 11 +++++------ devel/py-Jinja2/distinfo | 6 +++--- 2 files changed, 8 insertions(+), 9 deletions(-)
Copied from bug 260440 comment 4 ... @Yuri This sphinx update caused a regression. This port is maintained by python@, nivit@ just happened to be assigned to nivit@. The correct/better course of action would have been a reset of assignee to python@ (as it wasn't on our radar). It was an assignee timeout, not a maintainer timeout. Not withstanding, comment 1 detailed what was necessary to progress, which wasn't ack'd. The missing bulk / consumer build test would have picked up the regression. Further, this update was for 3.0.3 (which wouldn't have broken the build) The commit updating to 3.1 should have gone through a cycle on this issue, including updating the summary to match.