[cvsnt] Re: unexpected '\x0' reading revision number in RCS file

[Jesse]

I certainly hope there was no way to do it, as I'll feel foolish.
I had a bunch of files that had been committed by users using the -kv or -kb 
option that needed to use the -kB or -kBx option.  The resulting larger 
filesize was literally crashing the server when people attempted to obtain 
older versions of these resources (the ,v files reached 800+ mb ).  So I had 
to get every old version, copy it over to a different directory, and 
recommit it.

In some cases, I used a program I wrote to chop up the ,v files into 
managable files to get the old versions without crashing ther server.

Then, I wanted to keep the revision history, so I wrote a program that went 
through all the old and new ,v files, and copied the relevent information 
(author, date, state, commitid and log message) from the old ,v to the new 
,v.

I was able to retrieve old versions, and diffed them, etc.

It's possible that the code I used to copy the header info added the null 
char at the end of the file.

And it certainly was risky, but I did try my best.

Apparently I failed...

Jesse
"Tony Hoyle" <tony.hoyle at march-hare.com> wrote in message 
news:e5nm70$330$2 at paris.nodomain.org...
> Glen Starrett wrote:
>> Bo Berglund wrote:
>>>> I recently try manually modifying some CVS files, both to gain some 
>>>> knowledge and to make some batch changes.  Afterwards, I tried 
>>>> retriving histories, old versions, etc, and was successful.
>>>
>>> Are you saying that you are directly modifying he ,v files in tghe
>>> repository?
>>> I certainly hope not....
>>
>> That seems to be what he was indicating.  Certainly VERY risky to the 
>> repository!
>>
> Definitely... all bets are off once that happens.
>
> I wonder what the changes were - there's probably a builtin way to do it.
>
> Tony