Re: [Flex-help] working with two bufffers?

[Oliver B. <li...@ol...>]

Hello,

after another change I could get the unfolding work properly. But I like
your suggestion to separate the unfolding from the scanner though.

I changed my regular expressions to

from

> folded_line     ^[^\r\n]*{mycrlf}{wsp}

to

folded_line     ^[^\r\n]*{mycrlf}/{wsp}

and did some other minor changes because yyleng is different due to the
use of {wsp} as lookahead pattern.
I am not sure why the first expression did only match every second
result. I assume that this has to do with the length of actual lines
that are scanned but this is just a hunch.

Best regards,

Oliver Block


Marcel Laverdet schrieb:
> I would recommend writing your stream unfolder outside of Flex because
> it's an extremely simple transformation. Then run Flex to scan for tokens
> in your unfolded stream. A manual unfolder is going to be more effective
> since the transformed text will always be smaller than the original text so
> you can reuse memory and avoid having to alloc twice. I would just alloc
> the size of your file plus 2 bytes, run the manual unfolding (just a simple
> for loop), terminate the unfolded text with 2 NULL bytes, then use
> yy_scan_buffer on that text.
>