[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnash-commit] gnash ChangeLog gui/NullGui.cpp

From: Martin Guy
Subject: Re: [Gnash-commit] gnash ChangeLog gui/NullGui.cpp
Date: Fri, 6 Jul 2007 15:40:33 +0100

2007/7/6, Tomas Groth <address@hidden>:
I don't understand why it's so important not to depend on boost_date_time

Keeping the code size down. Gnash is already huge. and pulls in more
libraries than mplayer :-/
boost_date_time is large and slow (that's why it's split off into a
separate package) and the local timezone/daylight saving time stuff
brings in complex tables and hairy algorithms for predicting what the
local governments of the world have decided over the last hundred

If you just need microsecond timing (which we don't need; millisecond
is enough for human-visible video) you don't want libraries that go
looking up the local timezone and daylight saving time through layers
of extra code and only end up calling the "how many milliseconds"
function that you would have used anyway.

why not keep the changes in NullGui, and work on from there to remove the use 
of tu_timer?

The only thing inherently wrong with tu_* is the egotistical use of
the author's initials in the identifiers.
In this specific case it seems to me to make more unnecessary bloat
for no practical gain whatsoever, plus a touch of "don't fix what
ain't broke".
As well as increasing code size even further, extra dependencies
complexify the build process for everyone (think Windows, think

However you can make that change throughout Gnash if you wish.
I just happened to notice a tiny and largely pointless miodification
made 3 days ago that has a negative impact on the software for no
practical gain.

But that's just my opinion of course. If you're itching to use
date_time and don't care about code size and complexity, go ahead. You
can reimplement Date.cpp to use it too if you're feeling strong :)


reply via email to

[Prev in Thread] Current Thread [Next in Thread]