As for the refinements mentioned in JEP454

[Maurizio Cimadamore]

See the "throws" section of the javadoc:

    |IllegalArgumentException
    <https://docs.oracle.com/en/java/javase/22/docs/api/java.base/java/lang/IllegalArgumentException.html>|
    - if |charset| is not a standard charset
    <https://docs.oracle.com/en/java/javase/22/docs/api/java.base/java/nio/charset/StandardCharsets.html>

We only support common fixed-size charsets, as being fixed-size allows 
us to determine string length and also how many terminator bytes to add.

Maurizio

On 11/06/2024 14:05, Jonathan Rosenne wrote:
>
> Hi,
>
> regarding (3), the link just says Charset, not StandardCharset. Please 
> clarify.
>
> Jonathan
>
> *From:*panama-dev <panama-dev-retn at openjdk.org> *On Behalf Of 
> *Maurizio Cimadamore
> *Sent:* Monday, June 10, 2024 12:46 PM
> *To:* Cheng Jin <jincheng at ca.ibm.com>; panama-dev at openjdk.org
> *Subject:* Re: As for the refinements mentioned in JEP454
>
> Hi, some answers below:
>
> (1) refers to the ability to query what primitive layouts are like, 
> using Linker::canonicalLayouts.
> See: 
> https://docs.oracle.com/en/java/javase/22/docs/api/java.base/java/lang/foreign/Linker.html#canonicalLayouts()
>
> (2) refers to the new methods in MemoryLayout, namely 
> MemoryLayout::scale/scaleHandle, together with the fact that now all 
> var handles generated by this API take a start offset (so they are 
> easier to combine).
> see: 
> https://docs.oracle.com/en/java/javase/22/docs/api/java.base/java/lang/foreign/Linker.html#canonicalLayouts()
>
> (3) refers to the fact that you can now read and write strings with 
> more charsets than just UTF8. You can now use any charset in 
> StandardCharsets (to which we have added UTF32, for good measure).
> see: 
> https://docs.oracle.com/en/java/javase/22/docs/api/java.base/java/lang/foreign/MemorySegment.html#getString(long,java.nio.charset.Charset)
>
> Maurizio
>
> On 07/06/2024 23:30, Cheng Jin wrote:
>
>     Hi there,
>
>     I’ve got a question about JEP454 at https://openjdk.org/jeps/454
>     in which some of refinements confused me as follows:
>
>      1. /Enabled clients to build C-language function descriptors
>         programmatically, avoiding platform-specific constants;/
>      2. /Improved support for variable-length arrays in native memory;/
>      3. /Added support for arbitrary charsets for native strings./
>
>     I didn’t find any description in the JEP to clarify how they work
>     and what they mean to users.
>
>     Could anybody explain these refinements in more details? Many thanks.
>
>     Best Regards
>
>     Cheng
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/panama-dev/attachments/20240611/08cca362/attachment-0001.htm>