Re: [pdf-devel] Text module test data encodings

[Aleksander Morgado]

Not really, the purpose of getting the internal_hex in most tests is
to really know the internals of the pdf_text_t, so it should be
retrieved as UTF-32HE, not as UTF-8.

On Tue, Dec 23, 2008 at 10:33 PM,  <address@hidden> wrote:
>
> Hi Aleks.
>
>    > I have finished task #79, the tests compile without errors.
>    > But, I have 40 failing test cases due to calling the text API with
>   the wrong
>    > UTF encoding type for each test data.
>    >
>    > Aleks could you help me with this ? I attach here the patch from my
>   branch so
>    > you can see what test cases are failing.
>    >
>
>   Most of the changes you did are between the INTERACTIVE_DEBUG
>   preprocessor conditions, which mean that you want to know the exact data
>   stored in the pdf_text_t. Before the opaque pointer change, the
>   internals of the pdf_text_t where accessible directly, but now you need
>   to use the pdf_text_t API to access the internals. Now, UTF-8 is not the
>   encoding used internally by the pdf_text_t: use UTF-32HE instead (Host
>   encoded UTF-32, which means UTF-32LE in LE machines and UTF-32BE in BE
>   machines, without any extra option (no country/lang or 32-bit NUL suffix).
>
> Do you agree with Gerel's patch, then? Should we install it?
>
>
>
>