[TxMt] Re: GetBundles Request: GitHub Link / More info

Hans-Jörg Bibiko bibiko at eva.mpg.de
Wed Jul 23 18:00:23 UTC 2008


On 23.07.2008, at 19:27, Thomas Aylott wrote:
> Maybe we could add the link to the description in the info.plist. And
> for Github you could just tack on the homepage value for the repo.
> Seems Allan already officially added support for html in the
> description over a year ago:
>>>> Author: duff
>>>> Date:   Sun Mar 4 20:11:31 2007 +0000
>>>>
>>>>     Add the following keys to the bundle’s info.plist:
>>>>
>>>>     contactName:
>>>>        The full name of the contact for this bundle.
>>>>     contactEmailRot13:
>>>>        A ROT13 encoded email address for the bundle contact (we
>>>> obfuscate it to avoid spammers from picking up the files, seeing
>>>> how bundles will often be available via anonymous svn over http).
>>>>     description:
>>>>        A short description for this bundle. Do link to more info
>>>> about the language (or whatever) the bundle is about. I decided on
>>>> using HTML for this (but leave out the initial paragraph tag)
>>>> since it seemed like overkill to run a single line of text through
>>>> Markdown.pl, just to convert [foo](link) into <a  
>>>> href="link">foo</a>
>>>>
>>>>     Many of the current descriptions could use some improvements,
>>>> I just wanted to get the ball rolling :)
Well, if this is consistently done one could parse a href for getting  
more infos but I also saw bundle descriptions mentioned something like:
[fictive]
bla foo, the language deals with Python (www.python.org), ....

How to distinguish between such a link at the link to an HTML  
description?

Or did this mean either a plain text for describing the bundle or a  
link to a external document, but not both?
If so then it'd possible I guess.


> You could just parse a README file in the root of the bundle (if it
> exists) like for GitHub bundles.
>
>> For github, well, to parse the README file would be tricky caused by
>> the issue that these are coming in different formats.
>
> There are only a couple of options on GitHub: plain text, textile,
> markdown, etc...
> You could just simply support the first three with the existing
> textile and markdown parsers in the textmate support folder. Nobody
> seems to really use any but those anyway.
>
>> But it could be, maybe, possible to get the README stuff from
>> github's HTML page meaning e.g. the <div id="readme">...</div> part
>> of http://github.com/subtleGradient/subtlegradient-tmbundle/tree/
>> master
>>
>> Then the HTML rendering will be done by github.
>
> Well, whichever has the best performance. It might be faster to do
> that than to load the README file yourself and then parse it yourself.
I'll try to set up a basic thing.

>> BTW I uploaded a new GetBundles to the review. The entire error
>> handling is strongly improved. I also added a help nib plus button,
>> and a check box to make it possible to install a git bundle by
>> downloading it as zipball. This is useful for users whose computer is
>> behind a firewall which blocks the git port.
>
>
> Yep! I saw it. There are a few minor issues I was going to nit-pick
> about but figured I'd wait until you were done. I don't think nib
> changes merge very well, so should I post another zip if I change
> anything?
If you have something post it to me.

On the other hand I have to fix some "tiny" issues, e.g. how to  
cancel a runnig installation. Up to now I disallow it, because an  
installation process is a fragile thing. Even if you quit TextMate  
the installation runs further and finally it opens TM again, but this  
I have to fix.
Or an other issue I'm using Ruby's Timeout::timeout but this is a bit  
tricky. E.g. I cannot interrupt such a running timeout. And some  
other stuff.

--Hans


More information about the textmate mailing list