Re: hello again

it's been awhile myself so i don't really remember what i've done or not, and
where it might be documented.  but in general i don't think there's
documentation other than what you mentioned and the ChangeLogs -- but i usually
put a detailed summary at the top of the ChangeLog entries.  some of these
summaries, if very large, may have been moved to the internals manual; there
should [i think] be pointers from the ChangeLogs to the manual in these cases.

there is also a great deal of stuff sitting in workspaces on my computer, not
checked in.  one of the first things i'll be doing is figuring out exactly
what's there and documenting it.

as for large strings: i did implement something where i maintain a count of the
number of ascii chars at the beginning of a string, and use it to speed up
string ops.  i think this is checked in, but i'm not sure.  take a look at the
defn of the string object and see if there is a 21-bit field holding an "ascii
length" or something of the sort.  the wide-char stuff is mostly written, but
some stuff is missing like the code to automatically convert buffers as
necessary, and it's not tested.  i think the stuff already written is already
checked in.  take a look at the macros in text.h and see if they mention a
format.

ben

----- Original Message ----- 
From: "Hrvoje Niksic" <[email protected]>
To: "Ben Wing" <[email protected]>
Cc: "XEmacs Design" <[email protected]>
Sent: Tuesday, October 28, 2003 5:37 AM
Subject: Re: hello again


> Good to have you back, Ben.  I myself am curious about some things:
>
> There have been reports about instability and slowness of the 21.5
> code base.  I'm not sure if the people who complained about slowness
> compiled with or without error checking, but we might want to address
> these.
>
> I myself have been away for quite a while.  Is there documentation on
> what you have done, other than the mailing list archives and the
> CHANGES-ben-mule file?
>
> I'm interested in speeding up some string operations on largish
> strings, but that may be unnecessary if the wide char strings will be
> implemented.  What is the status of that?
>