Perhaps I should also point out that I'm a library writer, so my
binaries go into someone else's binaries. I get bug reports from my
users, and from their users. It's a challenge to make sure that my
users are keeping up-to-date with patches, so the number of bug
reports are lower, so I can tell a known issue from a regression, and
so I don't get pulled into every end-user conversation.
On 2/28/07, jason marshall <firstname.lastname@example.org> wrote:
> On 2/28/07, Bryant Eastham <email@example.com> wrote:
> > Jason-
> > I believe that you are confusing the requirements of version control
> > with the issues of distribution tracking.
> Entirely possible, but I think the fact that the FAQ bears a question
> very close to what I'm asking indicates that's not entirely true.
> Unfortunately the FAQ answer reads more like an apology than a
> rationale for why nobody would want to do that.
> > Obviously the problem doesn't exist for people that have access the SVN
> > repository - they can just ask and they will know exactly what tag they
> > are working with.
> Actually, the problem ONLY exists for people that take the code from
> the repository. I have a little website where I publish the artifacts
> from a Continuous Integration tool. It tags and bags all of the
> builds (including the ones determined to be important), and everyone
> who uses that site always knows exactly which build of the code
> they're running.
> It's the people who want to grab the code from the repository (or
> can't access the website, for various sad reasons) that can't get the
> build artifacts as originally generated. They have to re-generate
> them through the build tool (in this case, Ant). But I have no way of
> injecting the build number into the build scripts, short of a manual
> process. I use the tools I use because they take the human factor out
> of things that people are bad at (like bookkeeping), and this throws
> the human (me) right back into the fray.
> > It seems that your issue is getting some level of information to people
> > that obtain your source through some other method. If so, I would try to
> > solve the problem at the point of creation of the distribution. What
> > process do you use at that point? A Makefile with "make dist-src"? Some
> > manual process? It seems that is the appropriate place to generate such
> > other reports.
> > For example, I can do a "make dist-clean; make dist-src" in our tree,
> > and it will do exactly what I need. Part of that process is creating a
> > tagged README, License, and version files. Those files are always static
> > in SVN, dynamic in the distribution.
> > The system is completely automatic, and it works for any src tree
> > anywhere (trunk, tag, branch) which has benefits internally as well as
> > externally. As it turns out this is the same solution used to get the
> > information into our binaries.
> > Just my .02. If you are interested I can forward the Makefile lines, but
> > it is all pretty easy.
> > --
> > Bryant Eastham firstname.lastname@example.org
> > Chief Architect
> > Panasonic Electric Works Laboratory of America, Inc. , Salt Lake City
> > Lab
> > 4525 South Wasatch Blvd., Suite 100, Salt Lake City, Utah 84124
> > Phone : 801.993.7124 Fax: 801.993.7260
> > MEW Intranet: https://pewla.mew.com/index.php/User:Beastham
> > -----Original Message-----
> > From: jason marshall [mailto:email@example.com]
> > Sent: Wednesday, February 28, 2007 4:57 PM
> > To: firstname.lastname@example.org
> > Subject: Putting the tag name in the tagged code
> > I have a requirement for my code repository that I don't think is
> > exotic, or honestly all that uncommon. I need to get accurate version
> > information into any binary built from an SVN tag directory, for
> > purposes of version tracking and customer support.
> > I can certainly accomplish this with pre-built binaries. Many
> > Continuous Integration tools have facilities for doing this (and
> > generating the tags to begin with), but that only helps when I'm doing
> > a binary distribution. Is there a way to accomplish the same thing
> > (automatically!) with a source distribution?
> > The general consensus seems to be that it's both unwise and fairly
> > brittle to have an on-commit modification (such as code beautifying)
> > to the source tree. However, with a tag we're generally NOT expecting
> > anyone to update it, and at least in our case it's being created
> > directly in the repository, so there is no repos/working directory
> > inconsistency to mitigate.
> > So my question is, how best should I accomplish this? A pre-commit
> > hook? Groveling around in the svn info data to extract the tag
> > directory or revision number information?
> > Any suggestions would be appreciated. Pointers to existing solutions
> > to this problem would be superb.
> > Thanks,
> > Jason
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: email@example.com
> > For additional commands, e-mail: firstname.lastname@example.org
> - Jason
To unsubscribe, e-mail: email@example.com
For additional commands, e-mail: firstname.lastname@example.org
Received on Wed Feb 28 17:58:54 2007