[ts-gen] setting up mysql and editing data.c and once.c in shim/src directory

[R P Herrold]

On Sat, 10 Jan 2009, Justin wrote:

>
> I've run into a bit of a problem with understanding how I'm to add the
> lines below in
> the two files mentions.


The example cited was related to doing edits to set up 
userid's and such for using the database, with other than 
'stock' values, and written using a February 2007 version of 
the sources as an example.  In the last two years, we have 
come VERY FAR past that need, and I need to revisit that pate; 
I'll take a look when in the office Monday.

> I did some looking around at some examples of code from C++ 
> and Ruby and it seems to me to be C++ code. What I need to 
> know is how to properly introduce the lines below,

The good news is that you no longer need to do those source 
code edits, and can simply drop the variable / value pair into 
the .shimrc There is a sample in any recent tarball; the 
release from Friday looks like this:


[herrold at centos-5 shim-090109]$ cat .shimrc
DbmsName        mysql
DbmsHost        localhost
TableSet        testing
UserName        shim
Password        0
FeedName        tws
FeedHost        localhost
FeedPort        7496
HqPeriod        11
ClientId         8
Upstream        paper
LogFiles        log
PollTime        2
[herrold at centos-5 shim-090109]$

The first four entries are simple to edit.  My working copy 
(which connects on host: xeon.first.lan to a database: 
rph_testing  under userid: rph_shim  with NO password), looks 
like this:


[herrold at centos-5 ~]$ cat .shimrc
DbmsHost        xeon.first.lan
UserName        rph_shim
TableSet        rph_testing
    ...

Note that these are not order sensitive.


The .shimrc file was added, after Bill and I discovered that I 
was 'stepping on' his working dataset, with a failed source 
code patch for like the third time.  (additionally I moved my 
MySQL server and dataset to another computer -- it can be 
really difficult to debug a side effect which neither person 
is looking for)

>  but honestly I don't see how to do this without spending 
> hours maybe days learning basic C++ code, that is to say if 
> I'm right about it being C++ code.

outlined above, and no source edits needed.

Part of the benefit of coming to the project now is that we 
are three years in, and so as others had to do this, we 
responded to their confusion with the .shimrc file mechanism. 
We have one fairly minor set of additions to that file, and to 
the initiialization process left on our roadmap before we will 
move the project from a 'beta' to a 'production stable' status 
[there are some other milestones in the code as well, which we 
need as well].  The changes will be largely invisible to an 
external user, and any recent writeup of the .shimrc in teh 
mailing list archive will be substantailly complete.

Welcome, by the way.

For the wider audience, we are largely complete with the 
development cutover to the Debian 'lenny' (and once security 
updates are underway well in 'testing', back to 'testing'). 
A final member computer to run Debian 'lenny' in the distcc 
buildfarm which Bill uses to speed incremental testing and 
development arrived yesterday, and I will slot it in Monday.

Additionally Friday's release builds and runs cleanly on the 
following platforms in production:

Debian 'testing' 32 bit (i386)
Debian 'lenny' 64 bit   (amd64)
OS/X 10.4 PPC           (32 bit)
CentOS 5 32 bit         (i686)
CentOS 5 64 bit         (x86_64 multilib)
RHEL 4 64 bit           (s390 multilib)

I will speak with the IBM 'chiphopper' team, to get access to 
Power 5 (64 bit) RHEL turned on for verification of function; 
I anticipate no surprises or problems there at all.

The remaining possible platform (Windows 2000 and later with 
Cygwin 1.7) has a build problem with the MySQL libraries at 
the moment, and so we do not have function there at the 
moment.

-- Russ herrold