From mboxrd@z Thu Jan 1 00:00:00 1970 From: John Keeping Date: Fri, 21 Jun 2019 13:36:24 +0100 Subject: [Buildroot] [RFC PATCH] download/git: ban branch references In-Reply-To: References: <20190619151817.6331-1-john@metanate.com> Message-ID: <20190621131846.0f8f74e4@donbot> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net On Thu, 20 Jun 2019 11:27:02 -0600 Joel Carlson wrote: > On Wed, Jun 19, 2019 at 9:19 AM John Keeping wrote: > > > > As described in the manual, using a branch name as a version is not > > supported. However, nothing enforces this so it is easy to specify a > > branch name either accidentally or because new developers have not read > > through the manual. > > At a more general discussion level, I have to admit I like having the > ability to use a branch name even if it isn't supported. During > development I find it easier to use a branch name and just delete the > cached version to force it to re-download when I know I have new > updates. I can switch to a tag when I have a version I actually want > tagged and "final." Otherwise I'd have to keep tagging changes and > updating what tag to grab in the config or package file to pull in new > changes. > > Granted, it isn't a huge inconvenience to switch how I do things, and > I could see it being useful to "idiot-proof" for people who don't > realize the potential problems with using branch names. But I at > least wanted to make my opinion known that I like it the way it is. I'm curious what advantages you see to this method compared to _OVERRIDE_SRCDIR. I find pointing to an external version of the source to be more flexible and generally faster than pulling from Git every time. Regards, John