svn-cur

[Stephen Montgomery-Smith]

On 12/27/2011 12:27 PM, Julian H. Stacey wrote:

>> From:		Stephen Montgomery-Smith<stephen at missouri.edu>
>> When I modify or take over someone elses code, I try to follow their
>> style, even if this leads to slight inconsistencies.  The original
>> writer of the script mkCTM must have intended for the first delta to be
>> called xxx.0001.  There is a flag in the script called "CTMfirst."
>> Setting this to "1" causes the script to generate xxx.0001 by comparing
>> it to an empty directory.
>>
>> This makes it very easy for me to start a new ctm series of deltas.  I
>> set this flag to "1," run the mkCTM script, then set this flag back to
>> "0," and then sit back and hope the whole thing works.
>>
>> It probably wouldn't be terribly hard for me to make slight
>> modifications to mkCTM to rename it to .0001xEmpty.  But I prefer to
>> stick to the original design as much as possible.
>
> OK, forget it then, Thanks.
>    ( Sorry, I had not been intending to suggest modifying a script, I
>    thought I was just remembering files names from the past to be
>    conformant with, that I thought/ presumed had started all zeros Empty .gz.
>    I was presumably wrong on that if its always been the same script
>    in use (&  archives I've tried to check on for interest have been
>    stripped to save space) )

When I first took over CTM, I didn't know about this CTMfirst flag.  I 
discovered it by accident when reading through the script.

Also it is quite likely that prior maintainers of CTM had forgotten or 
not known about this flag.  So some of the CTM lists may well have 
started with xEmpty file.  This might include some created by me.  But I 
also don't remember whether this happened or not.

The early days when I took over CTM, I didn't know anything about cvs 
and stuff like that.  The first ctm list I created was, I think, src-7. 
  It was a lot of trial and error.  Now I know a lot more about the 
process, and I can create a new ctm list very efficiently and quickly.

One day, I will convert the src-* lists to be generated from svn rather 
than cvs, and svn is in some ways easier to work from.  But I am 
delaying that process until some kind of svnsup is created.  cvsup is 
very stable with respect to internet time outs or computer crashes.  It 
seems to pick up from where it left off, and is to a large extent 
self-repairing.  cvs seems to be similarly stable.  I have no idea 
whether svnsync or svn are anything like as reliable.

> Chris's assurance covers what was most of interest :-)
> 	"Yes, when it gets to src-9.0100 there'll be an Empty file here"

Yes.  That is the main point.  And I am trusting that the script will 
automatically make this happen without any further intervention on my part.