Just a sample of the Echomail archive
[ << oldest | < older | list | newer > | newest >> ]
|  Message 59  |
|  Richard Webb to all  |
|  another convoluted brainteaser for the g  |
|  12 Jan 12 21:46:00  |
 
HI Folks,
following up a message from Richard Webb to all:
RW> One thing I've always wanted to implement was a bit of a
RW> history section, keyed by date. These go out on Sunday
RW> mornings, of course.
RW> This batch would run at midnight utc on Sunday of course adn prepare
RW> the history information, compiling into it a file
RW> that the weekly email generator code would find.
A couple of changes since that post ...
First, I came to the brilliant realization that when HOrst's count increments
or decrements a numeric environment
variable it's going to nuke the leading 0 if it exists.
So, we rename all the files for months under 10 such as
315.* 518.* etc. etc.
This means we set the %NEWS% environment variable a little
bit differently as well, by playing with HOrst's count then
moving it to where we want it, to lose any leading 0 that
would have been generated.
AS for what happens with the month turnover, we could lose
the prepend leading 0 to single digit months code of course, and then use
ifnumber to see if %NEWS% and %TURNOVER% are
equal. IF so, we reset %NEWS% to the date found in
%TURNOVER%
and call another batch. Because of this, we also create
semaphores for each day in the original batch.
An example, since we'll always have the proper date for
Sundays we don't need to test for turnover, but on mondays ...
:chkmon
count NEWS +1
ifnumber %NEWS% biggerequivalthan %TURNOVER%
if errorlevel 255 goto conmon
if errorlevel 1 goto turnover
goto conmon
:conmon
echo %NEWS% >> c:\hold\work\mon.sem
call mmsnhist.bat %NEWS%
if exist c:\hold\work\history.dat goto monhist
goto chktue
So, if turnover is reached, we branch to that label, reset
the env var %NEWS% and call rollhist.bat which checks for
the flag files, branches, and moves on.
Still open to suggestions if anybody sees an easier way to
roll this.
Regards,
Richard
---
* Origin: (1:116/901)
|
[ << oldest | < older | list | newer > | newest >> ]