On 11/08/2014 19:35, Dirk Hohndel wrote:
Quite a few patches were sent over the past few weeks that for various
reasons didn't make it into 4.2. I am going through my inbox and am
struggling to make sure that I get the right ones and don't miss anything.
And I am thinking that maybe it would be better if you guys resend what
you think should actually go into master now.
Things that are on my bucket list:
- Willem's next set of CCR patches
- Linus' magic trip patches
- Robert's tissue loading like the Petrel patches
- Tomaz' continued work on cmake and unit tests
- a solution to the heading event issue #586 - I don't want to just stop
showing the heading events like was done in one of the patches... there
needs to be a better solution... likely a separate graph that shows the
heading like a wind flag between the profile and the partial pressure
graphs maybe?
- I saw a few random other patches (and even applied a couple that were
trivial, obvious, and still applied without any pain).
So in summary - send the patches that you've been working on but that
didn't make it into master becuase of the 4.2 release process.
I have a question about the submission of patches for CCR functionality.
To make my point I take the liberty of attaching a diagram that I use in
working with the code. The next set of patches concern the tank
pressures. Tank pressure calculation is done in profile.c. The
high-level function that controls tank pressure calculations is the one
in pink in the diagram: create_plot_info_new. The way in which functions
are called from here is indicated in the diagram. The functions in
yellow work as a tight unit in setting up the tank pressures. The chief
activity is estimating tank pressures for points on the profile for
which no tank pressure values exist. This may be because of drop-out in
the pressure transmitter or because of the insertion of bookmarks in the
profile corresponding to events during the dive. In order to work with
either the oxygen cylinder or the diluent cylinder, I had to add an
extra argument to the calling procedure for these four functions.
Now, in submitting the patches, I thought of working bottom-up, starting
at the lowest level. But this will not work as the calling function will
call the lower level function with too few arguments. Likewise, working
top-down will yield similar problems. So the only sensible solution
appears to be to submit a single patch that includes all 4 the functions
in yellow as well as the one in pink. But this contains maybe about 300
lines of code and may seem too large a single change in one single shot.
So I need advice from the experts in this forum. What is the most
sensible way to go about this?
Kind regards,
willem
_______________________________________________
subsurface mailing list
subsurface@hohndel.org
http://lists.hohndel.org/cgi-bin/mailman/listinfo/subsurface
