[nem-en] Re: [svn] r7496: nemerle/trunk/ncc/parsing/Lexer.n

[Dmitry Ivankov]

On 2/26/07, Kamil Skalski <kamil.skalski at gmail.com> wrote:
>
> Could you explain the purpose of this patch? Maybe a testcase could
> demonstrate why it is needed?
> The difference I see is that now the newline char is not removed from
> input stream... which is a little bit confusing.

For example bug 945
If it does matter newline char can be explicitly removed at the end of
directive parsing.

Also, I don't like the (maybe little) performance hit of using option:
> extra allocation per every character of input.

If it is really a bottleneck, I can convert it to tuple (char * bool), or to
int. The last wont be slower than char.
Oh, by the way, tuple is struct, maybe option-like wrap for tuple('a,
bool) is a good idea?

I guess the old workings of code was better.
>
If replacing char with any other type is totally inacceptable, i'll revert
to old interface and maybe extend it in order to read chars not jumping over
newline.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /mailman/pipermail/devel-en/attachments/20070226/ea639546/attachment.html