Re: Is Tidy still being maintained?

Hi, thanks for your reply



> Speaking  for myself, if somebody made me aware of a security problem, or
> would like me  to review a patch, I would be happy to do so, beyond that,
> for quite some  time now nobody has been interested enough in the project
> to steer discussion  on the development list or otherwise engage with the
> project  actively.

I don't think I've seen any security problem.
And I'm not sure I want to get involved in fixing the C code, but I might submit 
some patches where I can see a simple fix.

> >I brought JTidy pretty much in sync with Tidy, to the point  where it
> >actually performs better, and I have to emulate bugs to get  similar results.
> >I have a bunch of bugs to report, but the ones I  reported last year are
> >still sitting there untouched.
> 
> Well, to  things like http://tidy.sf.net/issue/2917718 I think I've
> always reponded to  saying that there are very many errors Tidy does
> recover from silently, it's  never meant to be a fully featured HTML
> Validator, or something along that  line.

Well, most of the bugs I found (2917718 included) are regressions. They appear 
in the tests that come with Tidy, and they were introduced at some point, they 
didn't exist before.
Some of them are minor, but others cause broken output.

>(To me, usability of the
> Sourceforge bug tracking system has always  been terrible, and it has
> becomes worse recently, so I mostly stay away from  the tracker.)

I find it not too bad; of course, there's room for improvement.
Is there a better place to report bugs or submit patches for Tidy?
And if I submit patches (even for minor problems) will they get reviewed, and 
committed if acceptable?

> I would say we do still have infrastructure and people in  place to
> support maintenance, but nobody doing much groundwork like  working
> through the bug tracker, submitting patches, starting discussions  on
> the development list, and so on.

So basically nobody is fixing bugs anymore?

> There are basically four  components to Tidy, one is a HTML parser that
> can recover from errors in a  manner authors might expect (as sometimes
> opposed to what browsers do), a  component that checks for some errors
> and reports them, one where attempts  are made to fix problems, and one
> that "pretty prints" documents. The parser  is likely going to be re-
> placed by "HTML5" parsers, the validation component  probably has to be
> rewritten to accomodate "HTML5", I am not sure about the  "fixup" part,
> and the pretty printing should really be done by a separe  library that
> allows for more freedom than Tidy has done so far, pretty  printing is
> useful independently of all the other things (whereas parsing and  error
> reporting and fixups are not so easily separated).

You're talking about "replacing" and "rewriting" things for HTML5. Why not just 
add HTML5 support to what is already there?
If the new HTML5 code doesn't support previous HTML versions, then Tidy will 
lose at least 90% of its usefulness.
If it does, then it means reimplementing a huge amount of functionality instead 
of reusing it.
Anyway, HTML5 would be major new functionality, but the project is still 
suffering from lots of bugs.

> With that in  mind, the main thing that I would expect to happen in terms
> of Tidy  development is someone submitting a patch to support "HTML5"
> elements and  attributes in some rough manner, so you can continue to use
> Tidy where you  are used to it, and at the same time use "HTML5" features
> and I would not  expect much more to come of the project at the moment.

That sounds more like adding rather than replacing. That would be a good thing.
But I would rather expect bugfixes before "delving into new territory".

Regards,
Adrian

Received on Friday, 26 November 2010 07:02:26 UTC