Diary
This is a diary of my activities on this project. It is generally updated
every few days.
Note
While the Netmind service will tell me how many people have
signed up to receive e-mail when this page changes (thousands and thousands,
I am sure), it won't tell me the individual e-mail addresses. It will give
me statistical summary of demographics, should you choose to fill out the
mostly optional form which Netmind will present to you.
So, feel free to lurk in complete safety with the assurance
that your identity is unknown to me...
January 1 – January 7:
Goals:
Get familiar with D0 Release Process and L3 Build Activities
Activities:
-
Got separate briefings from Amber, Alan, and Paul
-
Began trying to understand WNT Release sequence
-
Set up office, acquired a PC from Dave Fagan
-
Began putting Linux onto the PC in a dual boot format so can get into ODS
resources if necessary
-
Met with Ruth and Amber
January 8 – January 14:
Goals:
Get familiar with D0 Release Process and L3 Build Activities
Activities:
-
Completed putting Linux on PC – some issues remain but expect to deal with
them eventually
-
Began trying to get PC (in it’s Windows NT incarnation) to have all the
necessary things on it so I could become a D0 WNT developer: cygwin, CERN
product, others as described on http://d0ntwg01.fnal.gov/nt/
-
Got NetMeeting installed on PC (for use with WNT build process).
-
Began trying to get Visual C++ 6.0 installed on PC. Got access to APPS
directory. Attempted install.Various problems. (Need product number and
user id) Still no success.
-
Paul gave me a walk-through of the build process on WNT. Will need another
but am getting the hang of it, I think. The goal is to get to the point
where I could do the WNT weekly/daily test release.
-
Met with Gordon.He will e-mail me some info about the current arrangements
of things. We will talk the next time he is out—if not before via e-mail.
-
Did Build 8 of t01.29.00 and nt01.29.00 under Paul's direction (i.e., I
typed at the keyboard on NT, Linux, and IRIX and Paul told me what to type).
January 15 – January 21:
Goals:
Get familiar with D0 Release Process and L3 Build Activities
Activities:
-
Got product number and user id for Visual C++. Made it further with install.
Needed access to MSDN_Lib_Disc1 and _Disc2 on APPS directory. Got that.
Completed Visual C++ install on D0NT114. Installed Service Pack 3. I appear
to be able now to develop C++ code on this PC. Will go back to WNT at D0
check list and see how to proceed.
-
Did the first build of the week for WNT under Alan's direction (i.e., I
typed at the keyboard on NT. Alan told me what to type--maybe a little
less of Alan telling me now, given my earlier experiences.)
-
Learned more about how D0 NT builds actually make it to the L3 trigger
processor PC's.
-
Pretty much did the NT builds from Wednesday on with a minium of handholding.
-
Got some more education in the things you have to do to build the executables
on the actually trigger processor PC's.
January 22 - January 29:
Goals:
Do an NT release (as opposed to a build). Get a C++ development environment
specialized to D0 working on my PC so I know how it is done. Continue NT
builds. Learn about the web builder internals from Gordon.
Activities:
-
(Dealt with some non-D0-related backup problems on fndaub)
-
Learned about how optimized builds in NT are being done.Was reminded that
I need to learn how to freeze a release.
-
Did an NT release (nt01.31.00) Build 01. Messed it up afterwards,
however, so Alan had to restart it for me leading to Build 02 being done.
(Thanks, Alan)
-
Did NT Build 03.
-
Learned from Gordon Watts how to make a new release available for use with
the web builder. Made nt01.29.00 and nt01.30.00 available. Asked Daniel
to try it. Learned how to kick the web builder when it hangs.
-
Organized my e-mail folders to deal with the ~130 messages a day that the
release process generates.
-
Did NT Build 04. Unfortunately, I mistakenly added a ZOOM package
into the build because it's version had changed and, for some reason, I
didn't realize it was a ZOOM package. So, Alan had to kill that build
and I had to back out the build. Fortunately, according to Paul,
it was a package that was only used internally by ZOOM so it was not necessary
to re-build the whole thing--only to remove the tmp folder within the particular
Zoom Package in the D0reldb area and then begin again but with Build 05,
having appropriately indicated that we were down reving the package in
question which I had up-rev'd. Hence the NT build begain about 1:30
or 2 pm which means it probably won't be done until midnight.
-
Saw a bit of the Linear Collider talks over lunch.
-
With NT Build 05 going, I turned to trying to get the document that instructs
one how to do release packages out of the D0reldb product area with an
idea to update it with what I am learning (NOT?) to do in building WNT
releases. Though there are easy ways to do this now, I ought to be
able (if I understand things) to get this document onto my PC in the local
/d0reldb (?) directories by checking it out of cvs.
-
Since I need this also to work for doing package development, I returned
to trying to get CYGWIN going on my PC (a 266 MHZ with 64 Mb). For
some reason the Cygwin device table was having difficulty mounting the
Microsoft Visual Studio C++ include directory. I fiddled around with
this for some time and got it to work but then the d0dist local directory
mount was hosed. Eventually by unmounting that and bringing Cygwin
up again, I managed to get it to all come up correctly with no mount errors.
-
Lesson seems to be to remember that the mount table has a memory and once
it gets the wrong memory one needs to umount it and bring it up again.
I now have Cygwin up and Visual C++ up. Setup d0cvs works but attempting
to actually access the cvs repository fails. I will talk to Alan
about this tomorrow to learn the secret of making rsh talk to the d0cvs
repository.
-
I have also been collecting various documents, lessons, etc. off the d0
web site that pertain to developing packages and such. Goal is to
eventually be able to develop some sort of a brain dead package on this
box in order to go through the development cycle that others are going
through in order to see what their life is like and possibly make it better.
January 29 - February 4:
Goals:
Now that I have had some schooling in the NT build process, the thinking
is that I should turn my attention to looking at the causes of the broken
messages on NT builds and interfacing between the build effort and the
developers who have packages that are built on NT (but who possibly don't
live on NT otherwise).
-
Get this PC into a C++ WNT Build Station for the Level 3 related packages
-
Update Gordon's web builder so it can do the next release
-
Review "broken" messages from NT builds that are being done through the
week. Determine the causes of the "broken"s. Review with Paul and
Alan and others and feedback the cause info to the developers.
-
Categorize the "broken" messages and be prepared to present the categorization
at the Subsystem Coordinators meeting on Friday.
Activities:
February 5 - February 12:
Goals:
-
Analyze Level 3 Messages
-
Get PC to be a WNT D0 Build Station
Activities:
-
I analyzed and posted on my web site the errors resulting from three builds
of the D0 WNT Level 3 packages
-
I have successfully removed the release (nt01.29.00) and installed the
new release (nt01.31.00).
-
I have successfully (I think) gone through G. Watt's steps for building
a local Thread Util package against the new release. I will document these
steps next week.