On Mon, Apr 06, 2009 at 10:50:45AM +0100, Eric Kow wrote:
> Haddockize literate comments in Darcs.Patch.Properties.
> -------------------------------------------------------
>> Trent W. Buck <[email protected]>**20090406090153
>>  Ignore-this: 49cb9c7e1e8627da17eada5d5a42281e
>>  I converted some LaTeX math to the corresponding Unicode characters,
>>  but it seems that Haddock 2.4 is back in the 70s: it converts high-bit
>>  bytes to control characters and similar gibberish.  Hopefully cases
>>  like this will encourage the Haddock codebase to get up to speed on
>>  the whole internationalization scene.
>
> So, I shall defer to you on all matters documentation.
>
> But before I apply this patch, are you sure this is a good idea?

No.  But I would like to hit the haddock people over the head with it,
grumble, etc.  (Note that this issue doesn't apply to later patches in
this bundle.)

> One concern I have is: how are people expected to edit this
> documentation easily?

With Emacs, and the TeX input method!  Yaaaaaaay!

See, you just type ^-^1 and get ⁻¹, or \leftrightarrow to get ↔.
What could be easier?

> Maybe we could bug the haddock folks to fix this for the hackathon?

That is indeed Phase #1 in this Cunning Plan.

> Is there another way to go about this?  Should we resign ourselves
> to ASCII art?

No way.  Particular since haddock is currently only able to produce
HTML (and annoying HTML at that, for a tty user like me) – and there
exist perfectly good definitions mechanisms like &prime; and <sup />
for the basic math markup called for in this instance.

Using ASCII in the source file itself would be more portable, but
haddock doesn't let you specify raw HTML output in your source
document.  IMO this is a Good Thing, because it would make it much
harder to add arbitrary other output formats in the future
(cf. markdown vs. rest).

> It seems like patch theory discussions have to go over email anyway,
> so darcs patch theory folks might as well devise a good ASCII
> notation and stick with it...

Shrug.  That's their call.

>> changepref test
>> set -x && ghc --make Setup; ./Setup configure --user -ftype-witnesses; 
>> ./Setup build; ./Setup test tests network
>> set -ex && ghc --make Setup; ./Setup configure --user -ftype-witnesses; 
>> ./Setup build; ./Setup haddock; ./Setup test tests network

Gah, that part wasn't supposed to go in!  Incidentally, test setpref
merging has completely borked --test in many of my repos :-/

> Resolve issue1413: remove duplicate documentation.
> --------------------------------------------------

I couldn't see a comment on this one; did you apply it?
I guess I'll find out tomorrow when I try to pull :-)
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users

Reply via email to