Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.
On Sat, 29 Jan 2005 15:34:22 -0700, Glen Starrett <grstarrett at cox.net> wrote: >Bo Berglund wrote: >>>>So for those of you wanting to preserve your code I recommend against >>>>installing build 1863 on anything but a test server. >>> >>>That's why it's a testing/development built.... >> >> Yes, indeed! >> I was foolhardy to install it on my private server though... > >*ahem* I think I have you beat here... I had installed it on my prod >server and witnessed a problem against a text file with binary >characters in a current project. I've reverted to 2.0.58d, but it still >has the taginfo problem where it was outputting the pointer to the name >instead of the name (I don't recall the exact problem, but it was >something like that). > >Tony -- Any chance of back-porting the taginfo fix to 2.0.58d until we >get 2.0.62.* stabilized? I supposed I could revert back to 2.0.51d, but >I don't know if I remember the changes backward to that release so I'm >hesitant. > >BTW: Tony has already addressed the problem I mention in 1864, but I >haven't verified it yet (and I learned my lesson on upgrading my prod >server that fast). If one made a branch off 58d and then merged in the changes in the taginfo handling into that and built it, wouldn't that fix 58d? Only thing is I can't anymore build cvsnt since it moved to DOTNET... And we don't really know where Tony fixed the pointer problem, right? Earlier on it was inside the tag.c (renamed to tag.cpp) file, but nowadays I guess the handling is in the trigger_something file.. It was done in build 1817 if my memory serves me right, though. /Bo (Bo Berglund, developer in Sweden)