Re: xmlCtxReadIO and BOM

Hi,

  I have wrappers around libxml calls, and already implemented the 
"hack" to skip the BOM. But I think there is an issue here, because when 
I use for instance xmlParserInputBufferCreateIO(), it skips the BOM when 
the encoding is provided. So for me there is an inconsistency here as 
xmlCtxtReadIO() should work the same way, or maybe I missed something.

Regards,

Frank


Le 02/02/2018 à 19:52, Eric S. Eberhard a écrit :
> Same advice I just gave to someone else.  Unless it is HUGE this 
> works.  Read it into a memory buffer (calloc, malloc, whatever).  
> Remove BOM.  Parse the memory buffer.
>
> If you do this often you can make the buffer address and it's size 
> static so that you don't release it (deliberate memory leak) and then 
> keep using it for minimal context switching (and more memory) ... if 
> you need it bigger, realloc.
>
> libxml2.a cannot do everything for everyone -- putting small wrappers 
> on things is good.  I generally use it with giant wrappers (meaning 
> the open, calloc, parse, etc are all one routine).  Then when changes 
> occur you can change your wrapper and generally life is good.  I would 
> not recommend coding directly with raw libxml2 calls -- they are lower 
> level but complex.
>
> BTW -- if the data is HUGE then write it to a /tmp file (removing the 
> BOM as you do it) and parse and delete the file ... modern machines 
> are so fast it won't notice.  I have systems sending and receiving 2-4 
> million XML docs per day.  Several have to deal with quirks -- 
> especially when dealing with "big box" places or shipping companies 
> (you cannot get Target or USPS to change for you).  One does not put 
> spaces between attribute ending quote and the start of the next 
> attribute.  It is wrong.  It won't parse.  So I filter it with my 
> wrapper.  And so forth.
>
> The specs are often interpreted differently by other organizations 
> that you cannot win against.  So work around them.
>
> Daniel is a great guy but ... if he had to make an exception and 
> change for everything I have (and I imagine thousands of others) he 
> need 100 clones :-)
>
> E
>
> On 2/2/2018 1:19 AM, Frank Gross wrote:
>> Hi,
>>
>> I came to an issue where I try to parse an XML document from an HTTP 
>> stream. I decode the charset from the HTTP header and then create a 
>> xmlCtxReadIO with that charset value as encoding parameter. The 
>> problem is that the XML document has three BOM characters, and it 
>> seems that xmlCtxReadIO considers the document as malformed in that 
>> case. (XML document with BOM value and when we call xmlCtxReadIO with 
>> an encoding value).  Notice that if I don't provide the encoding 
>> value to xmlCtxReadIO, the parsing works well as BOM is decoded. Is 
>> there a way to ignore the BOM when parsing with xmlCtxReadIO ?
>>
>> Regards,
>>
>> Frank
>>
>>
>
> -- 
> Eric S. Eberhard
> VICS
> 2933 W Middle Verde Road
> Camp Verde, AZ  86322
>
> 928-567-3727  work                      928-301-7537  cell
>
> http://www.vicsmba.com/index.html              (our work)
> http://www.vicsmba.com/ourpics/index.html      (fun pictures)

-- 
Frank GROSS
Software Engineer - Web Services
Four J's Development Tools - http://www.4js.com

_______________________________________________
xml mailing list, project page  http://xmlsoft.org/
[email protected]
https://mail.gnome.org/mailman/listinfo/xml