[glob2-devel] .hgignore added to the repostory

[Kai Antweiler]

Hi *,

I have added an .hgignore file to our repository, so that "hg status"
becomes cleaner.
But this will also have an effect on "hg add".
The file:

syntax: glob

*.Po
*.map
*.gif
*.png
*.txt
*.ogg
Makefile
Makefile.*
config.*
configure
*.ttf
aclocal.m4
output.?
requests
traces.?
default.apspec
libgag.a
src/glob2
stamp-h1


Every file which matches will be ignored.
I have created this file a few month ago, when I didn't have much time.
I'm sure much of it is silly and can be improved.

I also added a .hgtags file, so that now our old cvs tags (like "alpha20") are
in mercurial.

Thereby I recognized that my old repository is not consistent
with the one we are using right now.  Each repository has some revisions that
the other one doesn't have.  Maybe they are just sorted differently, but I
think some (<1%) are missing.  Also my old repository seems to have
slightly more revisions in it.

tagname:             remote:   old:

tip                    2564 : 2560
alpha21                2387 : 2400
alpha20a               2384 : 2397
alpha20                2383 : 2396
alpha19                2235 : 2250
alpha18a               2180 : 2195
alpha18                2179 : 2194
alpha17a               2058 : 2073
alpha17                2057 : 2072
alpha16                2037 : 2052
alpha15                1949 : 1964
alpha14                1925 : 1937
alpha13                1919 : 1931
alpha12                1900 : 1912
alpha11                1873 : 1885
alpha10                1822 : 1834
alpha9                 1730 : 1741
alpha8                 1688 : 1699
alpha7                 1625 : 1636
alpha6                 1538 : 1547
alpha5                 1401 : 1410
alpha4                 1362 : 1371
alpha3                 1254 : 1258
alpha2                 1251 : 1255
alpha1                 1238 : 1242
stable-0-2              817 :  821
semi-stable-0-1         517 :  522
globalbug                31 :   30


"tip" is not representative. "tip" old is revision 2547 in the new repository.


Ok, I have greped the summary of both logs, concatenated and sorted this,
used uniq -c on it, clean recent revisions up manually

     3 summary:     16 bits GL fix
     3 summary:     Added an assert to prevent too big packet to be sent.
     3 summary:     Added some forbidden pathfinding dump.
     3 summary:     Fixed FillRect in GL
     5 summary:     Fixed manual
     1 summary:     imported sources
     3 summary:     Now players with differents checksums are dropped
     3 summary:     Patched YOG timeouts
     5 summary:     Removed a bug in case a host don-t join his own game!
    17 summary:     Removed many dependencies.
     3 summary:     Removed Nicowar from the Menus in preperation of
a Nicowar-less release
     3 summary:     Updated network checkSums system to have a list
of recents checks sums.

Looks to me like:
One conversion tools implemented some commits twice (maybe those revisions
from which other branches forked).

Looking into my old log:
changeset:   500:524a282b56f3
user:        nct <nct>
date:        Mon Jan 06 09:43:30 2003 +0000
summary:     Fixed manual

...

changeset:   486:c76a10fb0767
user:        nct <nct>
date:        Mon Dec 30 17:35:35 2002 +0000
summary:     Fixed manual

changeset:   485:065f3220b12f
user:        nct <nct>
date:        Mon Dec 30 17:34:54 2002 +0000
summary:     Fixed manual

OK, I have compared the diffs:
Stephs tool has fused some revisions.  The second commit above undid some
changes from the first one.

So this should be OK.
--
Kai Antweiler