There is a hardcoded array limit based on one-time measurements.

I think that would be possible, but would the gain in startup speed be worth the coding effort?

On 11-04-07 04:08 PM, Nate wrote:
Ah, interesting. Does Ragel actually approximate the bytecode size of
the separate functions and break them into multiple init functions if
near the 64kb limit? If it does, maybe an approximation could be done to
avoid init functions entirely. Or alternatively, how about a "write
init" option to not use functions at all?

-Nate


On Thu, Apr 7, 2011 at 1:34 PM, Adrian Thurston <[email protected]
<mailto:[email protected]>> wrote:

    If I recall correctly, the functions are there because the
    initializers become code in the calling function. Since java has a
    limit on the amount of (compiled) code per function, we always use
    separate functions so we can test for a reasonable approximation of
    the limit on a per-array basis. I think that was it.


    On 04/06/2011 01:20 AM, Nate wrote:

        Ragel currently outputs a method and static field for Java, eg:

        private static byte[] init__tableLayout_trans_targs_0 () {
        return new
        byte[] {...}; }
        private static final byte _tableLayout_trans_targs[] =
        init__tableLayout_trans_targs_0();

        We could save a few bytes on class file size by just assigning the
        array, without an method:

        private static final byte _tableLayout_trans_targs[] = new
        byte[] {...};

        The end result is the exactly the same, but less code needs to
        be output
        and the class files will be a little bit smaller.

        -Nate



        _______________________________________________
        ragel-users mailing list
        [email protected] <mailto:[email protected]>
        http://www.complang.org/mailman/listinfo/ragel-users
        --===============1102049


    --
    Adrian D. Thurston
    http://www.complang.org/thurston/

    _______________________________________________
    ragel-users mailing list
    [email protected] <mailto:[email protected]>
    http://www.complang.org/mailman/listinfo/ragel-users




_______________________________________________
ragel-users mailing list
[email protected]
http://www.complang.org/mailman/listinfo/ragel-users

--
Dr. Adrian D. Thurston
http://www.complang.org/thurston/

_______________________________________________
ragel-users mailing list
[email protected]
http://www.complang.org/mailman/listinfo/ragel-users

Reply via email to