Increased codebase consistency is not necessarily an outcome of applying the expanded rule. I can imagine a codebase that has a lot of xcom retrievals where many of them do not request the default return. In this case, instead of a single pattern of a "long jinja string", there will be 3 different pattern: long jonja, short jinja, and non-jinja.
Btw, I should've mentioned that my opinion is aligned with Wei's, and that MC has already been implemented (PR pending in ruff) - we're simply trying to determine the community's views on this. I'm playing the devil's advocate a bit... On Fri, 24 Apr 2026, 18:23 Przemysław Mirowski, <[email protected]> wrote: > +1 to Wei comments. Also it will improve consistency of user's codebase > when we will detect these cases too. > ________________________________ > From: Wei Lee <[email protected]> > Sent: 24 April 2026 09:37 > To: [email protected] <[email protected]> > Subject: Re: [DISCUSS] Extending the AIR201 rule to mixed-content strings > > 1. Yes — the work is already done in the PR, so no additional effort is > needed to land this. > 2. I think this should fall within the scope of AIR201. > > Best, > Wei > > Dev-iL <[email protected]> 於 2026年4月23日週四 下午6:55寫道： > > > Hi everyone, > > > > This discussion revolves around the ruff rule that replaces > > `"{{ ti.xcom_pull(task_ids='task_1') }}"` -> `task_1.output` > > > > The rule previously approved by the community explicitly excluded the > > "mixed-content" (MC) case, e.g.: > > `"echo {{ ti.xcom_pull(task_ids='task_1') }}"` -> `"echo {{ task_1.output > > }}"` > > > > Ultimately, in the MC case we're replacing a string by another string. > This > > replacement has a minor benefit in readability, but does not provide the > > same IDE and refactoring benefits that the original does. Several > > questions: > > 1. Do we even consider the MC case as worthy of replacing, in light of > the > > minor benefits it brings? > > 2. If yes, should this be part of AIR201, or a new rule? Are there cases > > where users might only want to lint for the "pure" case and not MC? > > > > Will appreciate your views on this matter! > > > > Best, > > Dev-iL > > > > 1: https://lists.apache.org/thread/1oh64pgco8bcbsxyyyflm7ccg1x3frol > > 2: > https://docs.astral.sh/ruff/rules/airflow-xcom-pull-in-template-string/ > > 3: https://github.com/astral-sh/ruff/pull/24673 > > >
