[AD] Patch frenzy!

["Vincent Penquerc'h" <vincent@xxxxxxxxxx>]

First patch is (yet another patch) for my mouse press/release events.
If pressing mouse on a widget and release on no widget, it always ended
up in the scary 'goto getout' statement, until any of the widgets got
another mouse event. I hope this is the last one :)

The two others are old patches

Second one is a patch I proposed quite a while ago, but was rejected by
Shawn. I resubmit it, along with the mail Shawn sent at the time. Since
this was just before Shawn moved to Brighton, he never got to reply to
my answer. I attach this email for you to know what it was about.
IMHO, this code is right though, and I would like some of you to check
it (it's only -2 +3 lines) for correctness. It works for me anyway.
If it is found to work, I'd like it to be included, as it just removed
a hardcoded constant into a lookup, which is more future friendly :)

Third and last one is to be able to tell dat to selectively prevent some
properties from being stripped from a datafile, it's basically a more
precise version of the strip levels. If you want to strip properties but
keep the COOL propery, you do dat -s1 -s-=COOL

These three patches are from Allegro 3.9.30 (yes, I know I should
upgrade now :))

-- 
Lyrian

From: Shawn Hargreaves [shawn@xxxxxxxxxx]
Sent: Sunday, April 09, 2000 4:54 PM
To: Vincent Penquerc'h
Cc: conductors@xxxxxxxxxx
Subject: Re: [AD] Grabber patch

Vincent Penquerc'h <vincent@xxxxxxxxxx> writes:
> This little one replaced a hardcoded constant by a queried one.

I think this is wrong: the hard coded constant 3 is for the selection 
index of the DLG_PACKLIST object, which only ever contains three items 
(no compression, object compression, or global compression). You replaced 
that constant with obj_count, which is the number of objects from the 
loaded datafile, but that has nothing to do with DLG_PACKLIST...

> BTW, while I'm thinking of it:
> this patch (and the three previous ones as well) are made against
> a modified version of wip 30, so the lines numbers might be well off,
> and the patches might not apply automatically if something was changed
> in its way since wip 30...

No problem, as long as you use unified (-u) or context (-c) diff formats, 
patch is very good at figuring out where to apply the changes. Even when 
it gets confused, it will usually apply most of the patch, and then leave 
the bits it can't figure out in a reject (.rej) file, from where it is 
very easy for me to sort out those little conflicts by hand.


-- 
Shawn Hargreaves - shawn@xxxxxxxxxx - http://www.talula.demon.co.uk/
"A binary is barely software: it's more like hardware on a floppy disk."

Attachment: grabber.pack_getter.diff
Description: Binary data

Attachment: mouse_rel.diff
Description: Binary data

Attachment: strip.diff
Description: Binary data