[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: My review of genx.h

• To: Rich Salz <rsalz@d...>
• Subject: Re: My review of genx.h
• From: Tim Bray <tbray@t...>
• Date: Thu, 22 Jan 2004 11:01:22 -0800
• Cc: xml-dev@l...
• In-reply-to: <40101323.8020907@d...>
• References: <A03C8438-4C8C-11D8-905C-000A95A51C9E@t...> <20040122035409.GD18459@m...> <20040122052954.GA2995@p...> <984DC92C-4C9F-11D8-905C-000A95A51C9E@t...> <20040122070336.GB8151@p...> <5EF3EAAD-4CFB-11D8-905C-000A95A51C9E@t...> <20040122170256.GD9212@s...> <40101323.8020907@d...>

Addressing issues not addressed elsewhere

On Jan 22, 2004, at 10:14 AM, Rich Salz wrote:

> Follow the standard open-modify-close model.  Support only one 
> document per genxWriter object.  That gets rid of genxNew; it would 
> also get rid of genxClose if it existed.  (See, I told you the current 
> model is too complicated. :)  Look at unix open syscall; flags specify 
> calling sequence.

Here's my thinking.  Genx is going to need a table for ascertaining 
which codepoints are legal Unicode characters, name characters, and so 
on.  I had the notion that rather than precompiling this I would cook 
it all up in code in a little one-time initializer routine.  Thus genx 
would have no static data, which would make it easier to compile into, 
for example Gnumacs and some embedded systems.  Thus you'd rather have 
a writer object that you generate multiple documents from.  On the 
other hand, if you don't mind having a precompiled table, then you're 
right, this eliminates one whole set of calls.

Am I going too far here?

> As someone else said, use typed opaque pointers:
>    typedef struct genxWriter_struct* genxWriter;

D'oh, that should have gone in the list of no-brainers.  Of course.

> Flags include:
>     gxoFILE -- next argument is FILE* for output
>     gxoNOPROLOG -- don't generate XML prolog
>     gxoSTANDALONE -- output "standalone='yes" in prolog
>     gxoABORT -- call abort() on any error (for debugging mode)
>     gxoFIXAMP -- turn & and < into &amp; and &lt;
>                  (perhaps gxoNOFIXMAP, I don't care)

I'm not too enthused about these, with the exception of the ABORT 
option, which could be a godsend at debugging time.

> genxScanUTF8 is a bad name -- looks to much like sscanf() et al.  
> Again, I'd merge that and the Check functions into the single Valid 
> predicate.

Yep, should be check not scan.

> How important is single-character output, really?  Making them use a 
> temporary buffer is not very inconvenient.  And if you believe must 
> single-char output will be in some kind of loop, anyway...

I'd use it all the time for odd characters by U+XXXX number

Cheers, Tim Bray  http://www.tbray.org/ongoing/

• References:
  • genx - abstract output
    • From: Tim Bray <tbray@t...>
  • Re: genx - abstract output
    • From: John Cowan <cowan@m...>
  • Re: genx - abstract output
    • From: Adam Turoff <ziggy@p...>
  • Re: genx - abstract output
    • From: Tim Bray <tbray@t...>
  • Re: genx - abstract output
    • From: Adam Turoff <ziggy@p...>
  • Re: genx - abstract output
    • From: Tim Bray <tbray@t...>
  • Re: genx - abstract output
    • From: jcowan@r...
  • My review of genx.h
    • From: Rich Salz <rsalz@d...>

• Prev by Date: RE: XLink and mixed vocabulary design
• Next by Date: Unicode input in sam/X11
• Previous by thread: My review of genx.h
• Next by thread: Re: genx - abstract output
• Index(es):
  • Date
  • Thread