From: Kenneth J.Davis (jeremyd@computer.org)
Date: Thu May 23 2002 - 07:51:58 EDT
Hello,
I've got some other stuff to do now (sleep, work, ... :-) but lest y'all
think I haven't been doing the stuff I said I was going to work on...
http://abiword.pchasm.org/patches/bug1750andstuff.zip
This zip is actually a diff of all the differences in my tree [abi/] versus cvs,
but it is basically more of my work towards a correct [all platforms] fix for
bug 1750 (ZoomType in AbiWord.Profile only takes 100, Width and Page) so I'm
working on getting the platform specific ap_<platform>Frame's to derive
from ap_Frame (which derives from xap_<platform>Frame ala the ap/xap App method)
and factoring some of the mostly cut-n-paste code into there. While bug 1750 is
marked a duplicate of bug 1447, they aren't quite the same, bug 1750 involves
using the values stored in the preference, whereas bug 1447 I believe requires
the fix for 1750 plus the zoom value needs to be saved to the preference file
[which at least on Windows doesn't appear to occur]. I have only compiled the
code on Windows (I'm going to reinstall QNX tonight and try to fix/test there)
but unmodified AbiWord doesn't compile on my Slackware box so I would appreciate
if someone could test on a Unix System (note though it may not compile successfully
there just yet). This is based on my previous 1750 Win32 only patch, so I think it
handles all cases correctly, however the Windows build [with or without this patch]
currently crashes on zoom set to fit to Width or fit to Page.
Also in this patch is a refactoring of the auto dictionary
hash downloader so most of the logic is XP (once again only tested on Windows),
but in theory with only a little bit of added logic and assuming libcurl compiles,
then any platform could support at least autodownloading of the .tgz hashes.
[still needs a little work (always uses /tmp issue) but otherwise works great.]
And last but not least, a trivial fix for bug 3354 (encoding file for american.hash
not installed) which requires this tiny 1 line patch and of course the encoding file
actually committed to abidistfiles module.
Feel free to comment on any obvious mistakes, alternate ways to do anything done,
or other constructive feedback. I have no intention of committing any of this
until I'm sure it compiles on at least one Unix System and preferably at least
one other person thinks it looks ok. :-)
If someone could kindly point me either to what the rule is about binary files
or answer this tiny question I'll commit the update for the peer libpng to 1.2.3;
Should the test png image files (binary) be committed and the other nonessential
directories? That is, should I commit all files in the 1.2.3 archive, with the
png files as binary (whatever the proper cvs options are)? or just update/add/remove
the relevant source files?
(If you can't access it or get my hosting services page, the longer url should work
http://www.perditionschasm.org/abiword/patches/bug1750andstuff.zip )
Thank you,
Jeremy Davis
jeremyd@computer.org
This archive was generated by hypermail 2.1.4 : Thu May 23 2002 - 07:58:23 EDT