...of all the text files distributed with the game: the various
translations of the MANUAL, the changelog, the license information in
COPYING, the instructions in INSTALL, and the README. Changes
sequences of lines to DOS-style single paragraphs, and attempts to
preserve displayed text and tables. The resulting files should be
ready to be shipped with the Windows package, and for most people
these files will be shown in NotePad when double-clicked.
Run inside the top level source directory to produce a pseudo-diff
highlighting possible unused images. Many image filenames (eg. for
terrains) are actually constructed by the code, so these are not
actually unused.
This is more general than the in-game code in Damage Calculations
since it handles berserk. It computes the joint conditional
probability function of the hitpoints each unit will have at the end
of the skirmish. The algorithm is currently slow: O(r*(n+N)*(h*H))
operations (n and N are number of swings per round, h and H are
hitpoints, and r is 30 if berserk is in effect otherwise 1). Hence,
probably not yet suitable for AI use.
(though I am not sure perl really will give any back to the OS) before doing
the copy of the cached file, as that step doesn't need the undef'd data and
can take a while to run.
...it will sometimes pick up the same language twice for a given
campaign. (Currently Invasion of Eliador has this issue with sv.) This
isn't a problem on the campaign server side. However it looks bad on
the web page. So I now have the campaign list suppressing duplicate
langauges.
I also fixed a bug where the languages weren't sorted the intended way.
(except for when webtgz.pl successfully returns a tgz file).
I also changed webtgz.pl to only request campaign info for the campaign of
interest. In the future (after getting campaign summary info loaded for old
campaigns on the campaign server) this request won't be needed at all and will
be removed.
...attributes in the campaign WML as well as the config WML, so that
when campaigns are downloaded the overview information is sure to
match the campaign. The two web interface perl scripts have been
added to utils as weblist.pl and webtgz.pl.
...to select the single textdomain to extract from the files. It is
not really complete (not knowing about the WML scope of #textdomain,
and assuming textual scope), but sufficient for now. Changed default
string-extraction rules to use --domain. Added #textdomain in
"library" wml files from standard campaigns, in places where it would
be missing should those files be #included from different campaigns,
but mostly so their strings get extracted.
