[Test] Changes newsteaser attribute to newsteaser class #1097

Closed
mweimann wants to merge 1 commits from feature/282-newstag-to-class into test
Member

@max.mehl this PR goes to test to see if the changes are working. It looks dangerous since so many files are changed.

@max.mehl this PR goes to test to see if the changes are working. It looks dangerous since so many files are changed.
mweimann changed title from Changes newsteaser attribute to newsteaser class to [Test] Changes newsteaser attribute to newsteaser class 2019-10-05 14:13:35 +00:00
Owner

Thanks. I will merge master into test first to resolve all these conflicts

Thanks. I will merge master into test first to resolve all these conflicts
Owner

A few questions:

  1. Can I rebase on test and merge?
  2. Did you check whether these files were actually up-to-date translations? We should be careful not to touch outdated translations.
A few questions: 1. Can I rebase on test and merge? 2. Did you check whether these files were actually up-to-date translations? We should be careful not to touch outdated translations.
Author
Member

Can I rebase on test and merge?

yes

Did you check whether these files were actually up-to-date translations? We should be careful not to touch outdated translations.

I didn't and now the system would say it's up to date because the change date is the same - right? We need the logical "version" system.. I have no clue how to check this in a reasonable way for all of the files.

> Can I rebase on test and merge? yes > Did you check whether these files were actually up-to-date translations? We should be careful not to touch outdated translations. I didn't and now the system would say it's up to date because the change date is the same - right? We need the logical "version" system.. I have no clue how to check this in a reasonable way for all of the files.
Owner

I didn’t and now the system would say it’s up to date because the change date is the same - right? We need the logical “version” system.. I have no clue how to check this in a reasonable way for all of the files.

Yes, exactly, that's how the systems works. I agree that a version system, cleverly integrated in different workflows and well-communicated, may solve such issues. However, we would have to think of a way to handle old files, and those who already are outdated.

For a method to only change files which are up-to-date, see #1092

> I didn’t and now the system would say it’s up to date because the change date is the same - right? We need the logical “version” system.. I have no clue how to check this in a reasonable way for all of the files. Yes, exactly, that's how the systems works. I agree that a version system, cleverly integrated in different workflows and well-communicated, may solve such issues. However, we would have to think of a way to handle old files, and those who already are outdated. For a method to only change files which are up-to-date, see #1092
Owner

This becomes relevant as soon as we've adapted the version tag as discussed on web@.

This becomes relevant as soon as we've adapted the version tag as discussed on web@.
Member

As we have not yet applied this (and have to resolve all the conflicts anyway), let me throw in the idea of removing the "newsteaser" attribute/class completely and define that the first paragraph of the news entry is always the newsteaser - which is, AFAICT, a hard rule of how the attribute is used currently.

As we have not yet applied this (and have to resolve all the conflicts anyway), let me throw in the idea of removing the "newsteaser" attribute/class completely and define that the first paragraph of the news entry is always the newsteaser - which is, AFAICT, a hard rule of how the attribute is used currently.
Owner

Makes sense to me. However, we have to check some XSL files that make use of this class. For instance, the logic behind the excerpts for the share buttons and some meta tags uses this.

While at it, should we also review whether we still need the id introduction which is used for newer articles? The difference is that some introductions span multiple paragraphs, so this might be trickier.

Makes sense to me. However, we have to check some XSL files that make use of this class. For instance, the logic behind the excerpts for the share buttons and some meta tags uses this. While at it, should we also review whether we still need the id `introduction` which is used for newer articles? The difference is that some introductions span multiple paragraphs, so this might be trickier.
Member

I think the explicit definition of which paragraphs form the introduction is still required. Whether it should be an id= or a class= might be matter for discussion, but I guess that would be a different issue.

I certainly agree that if we remove the newsteaser completely, all of its usage must be replaced by selecting the first paragraph of the news item.

I think the explicit definition of which paragraphs form the introduction is still required. Whether it should be an id= or a class= might be matter for discussion, but I guess that would be a different issue. I certainly agree that if we remove the `newsteaser` completely, all of its usage *must* be replaced by selecting the first paragraph of the news item.
Owner

OK, let's start with that and deal with the article ideas later :)

OK, let's start with that and deal with the article ideas later :)
Member

My assumption that the first paragraph is always the newsteaser is wrong. See for example https://fsfe.org/news/2009/news-20090120-02.en.html

So the complete removal of the newsteaser marker is out of discussion.

Now I am somewhat torn between class="newsteaser" and id="newsteaser". Since the purpose of the marker is to exactly identify a single element for several purposes, and the visual presentation is only one of these purposes, I'd actually lean more towards id="newsteaser". What do others think?

My assumption that the first paragraph is always the newsteaser is wrong. See for example https://fsfe.org/news/2009/news-20090120-02.en.html So the complete removal of the newsteaser marker is out of discussion. Now I am somewhat torn between `class="newsteaser"` and `id="newsteaser"`. Since the purpose of the marker is to exactly identify a single element for several purposes, and the visual presentation is only one of these purposes, I'd actually lean more towards `id="newsteaser"`. What do others think?
Owner

I think a newsteaser is always the first paragraph, and would consider anything else (like the article you mentioned) a bug.

In the linked article, this is rather being used as a way to make the text bold rather than a teaser to the full news item.

I think a newsteaser is always the first paragraph, and would consider anything else (like the article you mentioned) a bug. In the linked article, this is rather being used as a way to make the text bold rather than a teaser to the full news item.
Member

Would you think that this is such a hard rule that we should enforce it technically? What would we do with old news items (still visible in the news archive) where this rule isn't followed?

We also have news items like https://fsfe.org/news/2008/news-20081215-01.en.html where forcing the first paragraph to be the newsteaser would not work very well at all.

I have meanwhile come to the contrary result: that technically we should not make any assumptions neither about the position of the newsteaser within the article, nor about the number of paragraphs it consists of. With a class="newsteaser" we leave all options open, but on the other hand it does not stop us from following our own rules and preferences if we want.

@max.mehl and everybody else, please let me know what you think.

Would you think that this is such a hard rule that we should enforce it technically? What would we do with old news items (still visible in the news archive) where this rule isn't followed? We also have news items like https://fsfe.org/news/2008/news-20081215-01.en.html where forcing the first paragraph to be the newsteaser would not work very well at all. I have meanwhile come to the contrary result: that *technically* we should not make any assumptions neither about the position of the newsteaser within the article, nor about the number of paragraphs it consists of. With a class="newsteaser" we leave all options open, but on the other hand it does not stop us from following our own rules and preferences if we want. @max.mehl and everybody else, please let me know what you think.
reinhard closed this pull request 2020-05-01 09:06:46 +00:00
max.mehl deleted branch feature/282-newstag-to-class 2020-05-04 11:41:11 +00:00
Some checks failed
continuous-integration/drone/push Build is passing
continuous-integration/drone/pr Build is failing

Pull request closed

Sign in to join this conversation.
No description provided.