Multiple pages: Add homepages

[principis]

Added some more homepages, see: #2660
It seemed handy for all the git pages to link directly to the docs although it's not really a homepage. But it is the same website so... Let me know if I need to change anything.

[tldr-bot]

The build for this PR has failed with the following error(s):

Error: Parse error on line 32:
...arget_file}}.patch`> Homepage: <https:/
----------------------^
Expecting 'NEWLINE', 'TEXT', 'DASH', 'BACKTICK', got 'GREATER_THAN'

Please fix the error(s) and push again.

[blunket]

(Sorry for my previous comment -- I screwed up, lol)

[principis]

(Sorry for my previous comment -- I screwed up, lol)

NP :) my script messed it up.

[sbrl]

Thanks, @principis 😺

[waldyrious]

@sbrl what happened to your insistence on using the slashes at the end of URLs? ;)

In all seriousness, though, I think it might be better if we don't include them because people are more likely to write links without a trailing slash than with them. If we insist in keeping them, we'd have to constantly remind people of them -- case in point -- which is a bit of a drag for everyone.

Btw, @principis, I agree that using specific pages for each subcommand is more useful than linking all of them to the homepage of the main command. Perhaps we should use a different term than "homepage" for the main link, to avoid giving out the wrong idea. WDYT?

[principis]

@waldyrious I added slashes consistently with the websites as was said in #2660 I find them ugly so I would rather remove them but it's not that big of a deal.

Perhaps we should use a different term than "homepage" for the main link, to avoid giving out the wrong idea. WDYT?

Ow I forgot to mention that sorry 😆 It was my intention to mention it in my PR.
I couldn't think of a good term but I think it's a good idea.

Maybe something that means "more info"

[jedahan]

maybe we don't specify and let people put whatever they think is most appropriate? Most likely it will be
homepage: http://mycoolproject.org
but sometimes
documentation: http://project.readthedocs.org
and maybe even
community: http://project.gitter.im

[sbrl]

@waldyrious: Oops! Missed that :P I'm unsure that it's worth ensuring that all urls are formatted in exactly the same way though.

@jedahan: Sounds like a great idea!

[waldyrious]

I still vote for consistently not adding slashes :) they're redundant and quite prone to cause inconsistencies, as we've seen here and even in #2660.

As for making the URL free-form, I am not convinced it would be a good idea. I'm inclined to agree with @principis' suggestion of "More info:". Or we could simply call it "Link:" or "Web page:". Any of these could be used for all URLs, without the need to add different labels for different pages.

[sbrl]

Sounds good to me, @waldyrious

[vladimyr]

Prefer https if it is available

[vladimyr]

@sbrl Is this linter (https://github.com/tldr-pages/tldr-lint) still used for validating tldr pages? If so this should be added as rule there...

[agnivade]

Did you mean to comment on this PR ? What has tldr-lint got to do with adding a homepage link ?

[vladimyr]

In ideal world tldr-lint or any other linter used for checking tldr pages should report if http (home)page is linked and https version is available. It should report dead links also?

[agnivade]

Oh I see. I think that would put too much logic into the linter. I would want the heading text to be free-form without any strict format. The Homepage: format is just something we agreed to. I think it is not necessary to have linter checks for that.

[sbrl]

Some commands' homepages may be only available over http and not https though.

[vladimyr]

Quoting myself:

if http (home)page is linked and https version is available.

That is relatively easy to check.

[sbrl]

....in C/C++ with bison?

We need to do something about the maintainability of that linter at some point. I've studied bison at University since the last time I looked at it, and I think it might be overkill here.

I think a simple lexer + recursive descent parser would be more maintainable. IIRC, we discussed it elsewhere, but I can't remember where.

[vladimyr]

This is probably not a place to discuss it but parsing tldr pages with bison is like killing fly with orbit ion cannon 🙄
After all tldr page is markdown document and there are plenty of markdown parsers around that can be adapted/wrapped to check page format. Such parser could provide hooks for specific page segments like urls which can in turn be used for http/https and/or dead link checking.

[agnivade]

Couldn't agree more. Sadly, that is the state of affairs now. If someone wants to rewrite the linter completely, they are more than welcome.

[sbrl]

What's remaining for getting this and #2660 merged, @agnivade?

[agnivade]

We were waiting for the node client to have home pages support. Currently, the node client breaks trying to display the home page link.

@vladimyr - You had mentioned about subclassing the markdown parser to add the support. Are you planning to do that ?

[vladimyr]

Not atm 😞 Right now I don't have enough free time for doing it properly including extensive test suite but maybe some time in the future...

[agnivade]

Ok, how about we do @jedahan's solution so that this is unblocked for now, and you can do it properly later whenever you have time ?

[vladimyr]

Sure, totally agree 👍

[agnivade]

Done. Merging this.