Re: [lmi] [PATCH] Allow installing wxWidgets from GitHub

[Vadim Zeitlin]

On Wed, 05 Aug 2015 23:36:23 +0000 Greg Chicares <address@hidden> wrote:

GC> On 2015-08-05 16:43, Vadim Zeitlin wrote:
GC> > On Wed, 05 Aug 2015 15:09:13 +0000 Greg Chicares <address@hidden> wrote:
GC> [...]
GC> 
GC> > GC> I don't get it. Why do we want '--output-document'? Isn't it better
GC> > GC> without that?
GC> > 
GC> >  No, when you use wx_commit_sha the output of wget would be a local file
GC> > named $(wx_commit_sha).zip which is really not what we want.
GC> 
GC> Oh. The test I showed downloaded a static file, which is permanently
GC> stored on the server and naturally has a timestamp. When I repeat the
GC> test using wx_commit_sha, I get a file that's created on the fly,
GC> which doesn't have the desired name unless '--output-document' is used.
GC> And it has a current timestamp.

 I admit I hadn't tested this (I should have had), but, retrospectively,
this is not really surprising.

GC> (I imagine a meaningful historical timestamp could be determined by
GC> looking up the date of the commit, but GitHub doesn't do that on the
GC> server and I'm not going to do it myself on the client side.)

 Just to further confirm your choice, I don't even think this can be done
in any reasonable way: svn can use the remote repository directly, but git
only works with the local clone, so to get any information about the commit
we'd need to clone the repository, at which point we wouldn't have to
download the ZIP from GitHub. Anecdotally, downloading the entire
repository with all its history doesn't actually take much longer than just
getting the ZIP file because even if the total amount of data transferred
is roughly twice as big with git, it seems to use more efficient
compression than ZIP.

GC> Okay, I think I'm ready to write and commit this.

 Thank you!
VZ