Re: [PATCH] Add option --path to allow to run tests with real systems. You must install git before running this test.
- From
Johannes Schindelin <johannes.schindelin@gmx.de>
- Date
- Dec 6, 2007, 15:50 UTC
- Message-ID
- <Pine.LNX.4.64.0712061549491.27959@racer.site>
- In-Reply-To
- <fcaeb9bf0712060732r70d07135lf6a01ae2410a0ada@mail.gmail.com>
Hi,
On Thu, 6 Dec 2007, Nguyen Thai Ngoc Duy wrote:
Show 27 quoted lines
> On Dec 6, 2007 10:23 PM, Johannes Schindelin <Johannes.Schindelin@gmx.de> wrote: > > Hi, > > > > On Thu, 6 Dec 2007, Nguyễn Thái Ngọc Duy wrote: > > > > > Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> > > > --- > > > On Thu, Dec 06, 2007 at 10:15:05AM +0700, Nguyen Thai Ngoc Duy wrote: > > > > On Dec 6, 2007 2:32 AM, Junio C Hamano <gitster@pobox.com> wrote: > > > > > This is wrong, isn't it? $path may point at the freshly built but not > > > > > installed git executable, but it reports --exec-path the location that > > > > > git-foo and friends are to be _eventually_ installed, not the location > > > > > they are sitting after built, being tested, waiting to be installed. > > > > > > > > > Yes, forgot to mention you must do "make install" first :) > > > > > > > > > > Perhaps a check to remind people to 'make install' like this? > > > > Well, the whole point of "make test" is to be sure that it works before > > installing it. > > > > How about something completely different? > > I still think testing another round after installing is good. But > that's not my aim. So Ack on your patch (at least the idea because I > have not used it yet).
No, I think that testing after installing is way too late.
BTW there was no sign-off on my patch (as well as no commit message), because it does not work on symbolic-link-challenged systems. So it was purely meant as inspiration.
Ciao, Dscho