Release of new ExternalMedia #13
Comments
|
OK, but I guess the more elegant way would be to specify the start values in a way that is independent from the reference state. One could for example use a saturation state an enthalpy difference or something like that. |
Or just start with a T,p pair? That would be the most reliable. On Thu, Oct 1, 2015 at 6:51 AM, Jorrit Wronski notifications@github.com
|
I can see about the phases function. It's a bit of work to make sure that On Thu, Oct 1, 2015 at 8:07 AM, Ian Bell ian.h.bell@gmail.com wrote:
|
Test_Cell1D_ht is a CO2 TestBasePropertiesExplicit is a simple |
See my push in CoolProp/CoolProp@80851f4 On Thu, Oct 1, 2015 at 10:09 AM, JonWel notifications@github.com wrote:
|
Hi, |
BTW, would it be easier to handle the dependencies if we moved ExternalMedia to Git? |
Hi, The example codes haven’t been checked and corrected yet (there is sometimes some issues due to a change of enthalpy reference as I understood). I updated my checkbox at the beginning of this issue. For the dependencies, current |
Yes, git would simplify things. As for now, I think that we should go for some basic testing (I do not have access to Dymola anymore) and the release a new version with CoolProp 5. I am not sure how many people actually use the splined derivatives, but I guess that we should put a something about it in the release notes. |
At least one person seems in need for this feature (see issue 16). This function does not only compute the derivatives but is also used to compute the value of the smoothed density : https://github.com/CoolProp/ExternalMedia/blob/master/Projects/Sources/coolpropsolver.cpp#L241 (which may not be obvious by the name). Computing the two derivatives or the two derivatives + rho, in a single function call may be quicker as part of the code is identical. This may be an improvement to do for later if considered worth it (this may be interesting when using the tabular backed as state update is then quite fast). For the testing, I don't have time until mid February (examination time). At that moment, I'll will have to see with @squoilin what can be done in the frame of my thesis. |
Is a new release of ExternalMedia coming? Also, what is the "canonical"/upstream/definitive repository for ExternalMedia?
If it's 2., I can help with getting the changes from here back upstream, if necessary. |
The "canonical" repo is the one on the Modelica SVN server. The version with coolprop v5 is currently on a branch. Whatever is shown on GitHub is a mirror, so some functionality might be broken. |
To my understanding the main missing element is explained in #20 and #16 , and lies on the CoolProp's tabular backend side (ie once implemented in CoolProp it should work out of the box with this ExternalMedia repository) Currently the compilation has mainly be tested under Visual Studio. The compilation under Linux lack testing (#21) and I have no idea if the OMC script still work. |
sure thing, what do you need specifically? |
So, I just completed a Github import of the SVN repo, it can be found here. Interestingly, the last commit in the coolprop5 branch by @jowr says "Synchronising with 6f50f6d ". Since then there have been more commits in CoolProp/ExternalMedia, so the most recent state is actually in CoolProp/ExternalMedia, and there is no need to "move to github natively", as there is no new state we could transfer (the other branches have latest commits from 2013, so I don't think there is usable code in there, right @casella ?) The only thing I could imagine being useful is bugging Github support about connecting CoolProp/ExternalMedia with the upstream repo modelica/ExternalMedia - apparently CoolProp/ExternalMedia was not created with the fork button, so this connection is broken (no "forked from blabla text available), and it's much more difficult to send PRs from one to the other. |
Thank you for the suggestions. My idea was to have a sandbox repository here for us to play with the code. Whenever we get something working, I would sync this repo to the SVN server. This is because this repository breaks every now and then and the code on the SVN should at least be compilable... I have no problem with changing this, but I would like to stress that the repositories work as they should and I am happy to merge things from here to the SVN on demand. |
I see, so modelica/ExternalMedia would really serve as your upstream. Re: Github vs SVN I am extrapolating (from my own experience/feeling) that it would be much easier for other people to contribute or inspect the current state of development, if both repos were on Github. I guess it boils down to what @casella wants: If you want to switch the main upstream repo to Github, just say the word and I'll bug support about restoring the link between the two, so that PRs can be easily created. |
Fine with me. Please let me know if I should merge the current state into the SVN prior to the transition to GitHub - if @casella wants to move to Git as the primary VCS. |
I don't think we should copy the SVN repo to Github, to be honest. Right now, both the Coolprop and the Modelica one have one ancestry (i.e. same commit hashes, compare this and that), so it's easy to share/transfer code between. |
I have fixed a couple of bugs and if someone could help me testing, we might actually get close to a release. |
Hi Jorrit, |
@casella would you please give an update here of the new, current situation? |
@bilderbuchi, I wrote an email to several people including yourself yesterday evening. Did you get it? |
Yes I got the email, which prompted me - not everyone stumbling over this issue got it, probably. :-) Yeah, I was referring to the repo under the coolprop organisation, i.e. the one in whose issue tracker we are right now. Sure, it's smart to wait with deleting, it should just be clear what the canonical repo is w.r.t to source, issue tracker, ... :-) |
I guess it would be wise to delete this repo, but only when we demonstrate on the new canonical one that everything works fine (or in fact, better than it does now). |
Agreed, the only thing is that we made this repo work with CMake on Linux and Windows. Would you be interested in this functionality? I might be able to make a slimmed down commit to your repo that ships the CMake stuff over. |
FYI, this is the diff from the modelica repo master to the coolprop repo master, excluding whitespace changes. |
@jowr There was also the changes in Projects/Sources/coolpropsolver.cpp to make it use CoolProp 5 (then CoolProp 6) instead of CoolProp 4. But I agree with the idea that it's better to have two distinct entities with a more stable interface in between, rather than integrating CoolProp inside ExternalMedia as before, and having the link breaking frequently with CoolProp's updates. |
Francesco would like to release a new version of ExternalMedia soon. Do you @JonWel, @squoilin and @ibell think that we should push for including CoolProp5 in it?
I think the remaining tasks of #4 should be fixed before we include a new CoolProp. If we provide a binary, we cannot expect everyone to help us debugging and the software should work somewhat reliably. How well tested is the current code base? Do we feel comfortable releasing it?
The text was updated successfully, but these errors were encountered: