Re: lynx-dev bloating binaries

[Philip Webb]

990225 Henry Nelson, echoed by Matt Ackeret, wrote: 
> I can see no use for the `send a C)omment' function.

certainly i'ld say that's dispensable.

> there needs to be a way to turn off CTRACE and TRACE.
> In _release_ code sets, the toggle should not be compiled in as default.

this is one of a number of features which could be so treated.
 
> There are dozens of examples like the discovery of the WWW rules code.
> how many of you are going to use that code?

very few users are going to find out about it:
again, it can be kept, but subject to a configure switch.

> I could see lumping these tiny snippets of code into one NO_FRILLS,
> switched from off as default during development
> to on by default in (pre-)release code.

this gets hairier, as different people will want different things.
 
someone> Do you mean to remove most chartrans tables for CJK users?

i'm not clear how far chartrans code is switched off by default:
the vast majority of real-life users don't need it;
then there's the  .po  stuff, which is even more esoteric.
a configure switch would allow those who want these things to compile them,
while maintaining Lynx's reputation for non-bloatedness.
 
TD> The code only grows.

but the binary doesn't have to, except for those who want the features.

> is 100% of the cookies code <omitted from> the binary
> when configured for no-cookies?

another item some people can dispense with.

PW> my 2-8-1rel.1 binary is 1,45 M , 2-8-2dev.17 is 1,53 M (both stripped).
LP> that is  5 %  only - not so much difference, IMHO.
> it is significant: after 2 releases it could be 10%, after 3  15% etc.

actually,  1,05  1,10  1,16  1,22  ... : it really does add up.
 
TD> it is much smaller on Linux/BSD

IRIX could well be inefficient, but CHASS seems to have plenty of hardware.
I (capitalised for once) am not affected by the gradual bloat (yet),
but Henry's position may be like that of frogs Worldwide (no offense),
whose startling decline is a warning of something impending for humans:
one task for lynx-dev is to set about if-def'ing the code
to allow configure/compile-time exclusion of unwanted features.

[ and can everyone try to change Subject when the subject changes?
i corrected this one several messages ago,
but another thread has been going on under `dev.17: clue?' ... :
it's important for browsing the Archive (serious frown) ]

-- 
========================,,============================================
SUPPORT     ___________//___,  Philip Webb : address@hidden
ELECTRIC   /] [] [] [] [] []|  Centre for Urban & Community Studies
TRANSIT    `-O----------O---'  University of Toronto