Re: [Quantlib-dev] Some off-topic help on git

[Francois B. <ig...@gm...>]

Thanks for the advice, everyone.

Luigi/Piter, I didn't know about http://quantlib.org/install/vc10.shtml -
but what I did was similar. Instead of modifying the
Microsoft.cpp.<Platform>.users.props file, I created a separate boost.props
file and add that via property manager (a tip that I got off
StackOverflow). At the time it seems like a cleaner approach that modifying
Microsoft.cpp.<Platform>.users.props , but it still requires a .vcxproj
file change.

Dave, luckily I foresaw this, so I did ask it before starting to code. And
my code change was small, so I could experiment a bit.

In the end, git rebase --interactive master did what I wanted it to do. I
could remove the project file change commit. But I see that actually
reverts the change, which kind of makes sense. I would've like to keep the
commit, but just not push it, but I think the only clean way to do that is
with multiple branches, as Michael suggested. I want to shy away from
having multiple branches.

Thanks again.

Francois

Francois Botha


On 22 May 2014 09:37, Luigi Ballabio <lui...@gm...> wrote:

> Hi Francois,
>     all suggestions were good.
>
> The one I would pick up in this particular case is a combination of
> Piter's and Michael's: run "git rebase -i master", remove the commit
> you made on the projects (so that now you have a clean branch) and
> then use the Microsoft.Cpp.Win32.user property page to set the Boost
> directories (this doesn't modify the projects; instructions are at
> <http://quantlib.org/install/vc10.shtml>).
>
> Another, simpler way (which also works in more complex cases) would be
> to make the pull request and to tell me in a comment that I should
> skip one or more commits :)  I can cherry-pick the other ones myself
> (I don't merge the requests directly on GitHub anyway; I pull them on
> my machine and try to compile them first). It has the disadvantage
> that the pull request on GitHub will look as if I didn't accept it,
> but I can live with this if you can.
>
> In this case, though, I'd use the property page and revert the project
> changes. Otherwise, you'll have the same problem over and over again.
>
> Luigi
>
>
>
> On Wed, May 21, 2014 at 2:24 PM, Francois Botha <ig...@gm...> wrote:
> > Hi,
> >
> > Excuse me for going a bit off-topic. I'd like to contribute some changes,
> > but being new to git, I'm unsure how to progress.
> >
> > I've forked the quantlib repo, and cloned it. I've created a new branch
> to
> > work on. Everything is still clean. Now, I first have to modify my
> .vcxproj
> > files to point to my own boost installation. Obviously these are changes
> > that I don't want to include in the pull request later. After adding
> boost
> > and checking that it compiles correctly, I do a commit.
> >
> > Now I'm ready to start with the actual work. Let's assume I've finished
> that
> > and do another commit.
> >
> > How do I proceed now? I want to create a pull request, but I don't know
> how
> > to exclude the original, irrelevant commits. Is this easily possible?
> Should
> > I use git rebase or have used multiple branches? I also posed this
> question
> > on
> http://stackoverflow.com/questions/23781928/git-ignore-first-few-changes
> >
> > Once I figure this out, I want to apply a similar technique to pull in
> the
> > outstanding commits from Peter's changes (to let the QuantlibAddin build
> > successfully) and then do some work to expose more functions for the
> AddIn.
> > Yet again, I'll have to create a pull request, but I want to exclude
> Peter's
> > changes, because those are already in a pull request.
> >
> > Sorry for the newbie questions, but I hope you'll appreciate the
> > contributions that I can make.
> >
> > regards
> > Francois Botha
> >
> >
> ------------------------------------------------------------------------------
> > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> > Instantly run your Selenium tests across 300+ browser/OS combos.
> > Get unparalleled scalability from the best Selenium testing platform
> > available
> > Simple to use. Nothing to install. Get started now for free."
> > http://p.sf.net/sfu/SauceLabs
> > _______________________________________________
> > QuantLib-dev mailing list
> > Qua...@li...
> > https://lists.sourceforge.net/lists/listinfo/quantlib-dev
> >
>
>
>
> --
> <https://implementingquantlib.blogspot.com>
> <https://twitter.com/lballabio>
>