You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by "C. Michael Pilato" <cm...@collab.net> on 2011/05/06 23:01:30 UTC
New)(?) Subversion client user-agent string ugliness.
Is there a convention in HTTP user-agent strings to use whitespace to
delimit various bits of information carried in that string? I asked because
our client strings now look like so in trunk co:
"SVN/1.7.0-dev (under development) neon/0.28.2"
I *think* the " (under development)" bit is new(ish) and the result of
Julian's relatively recent tweaks to our version strings. I'm not so fond
of its appearance in this location, but wanted to hear other's opinion on
the matter.
So ... what say you?
--
C. Michael Pilato <cm...@collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
Re: New)(?) Subversion client user-agent string ugliness.
Posted by Julian Foad <ju...@wandisco.com>.
On Mon, 2011-05-09 at 14:41 +0300, Daniel Shahaf wrote:
> Julian Foad wrote on Mon, May 09, 2011 at 12:17:43 +0100:
> > On Fri, 2011-05-06 at 21:47 -0400, C. Michael Pilato wrote:
> > > On 05/06/2011 07:08 PM, Greg Stein wrote:
> > > > On Fri, May 6, 2011 at 17:06, Hyrum K Wright <hy...@hyrumwright.org> wrote:
> > > >> On Fri, May 6, 2011 at 4:01 PM, C. Michael Pilato <cm...@collab.net> wrote:
> > > >>> Is there a convention in HTTP user-agent strings to use whitespace to
> > > >>> delimit various bits of information carried in that string? I asked because
> > > >>> our client strings now look like so in trunk co:
> > > >>>
> > > >>> "SVN/1.7.0-dev (under development) neon/0.28.2"
> > > >>>
> > > >>> I *think* the " (under development)" bit is new(ish) and the result of
> > > >>> Julian's relatively recent tweaks to our version strings. I'm not so fond
> > > >>> of its appearance in this location, but wanted to hear other's opinion on
> > > >>> the matter.
> > > >>>
> > > >>> So ... what say you?
> > > >>
> > > >> Seems kinda redundant when used along with the '-dev' postfix in the
> > > >> version number, says I.
> > > >
> > > > Agreed. We should switch the User-Agent string from using SVN_VERSION
> > > > to SVN_VER_NUMBER.
> > >
> > > Okey dokey. r1100424. Thanks, guys.
> >
> > Good catch. Note that this consideration may affect the 1.6.x backport
> > proposal for r1084575, r1084581.
> >
> > ------------------------------------------------------------------------
> > r1084575 | julianfoad | 2011-03-23 13:15:11 +0000 (Wed, 23 Mar 2011) | 6 lines
> >
> > Change the descriptive version identifier text from '1.x.y (dev build)' to
> > '1.x.y (under development)' to help clarify that it's not merely a build
> > with debug symbols included or something like that.
> >
> > * subversion/include/svn_version.h
> > (SVN_VER_TAG): Change.
> > ------------------------------------------------------------------------
> > r1084581 | julianfoad | 2011-03-23 13:31:26 +0000 (Wed, 23 Mar 2011) | 11 lines
> >
> > Change the descriptive version identifier text from '1.x.y (under
> > development)' to '1.x.y-dev (under development)' to help clarify that it's
> > not merely a build with debug symbols included or something like that. This
> > is the second part of a change started in r1084575.
> >
> > * subversion/include/svn_version.h
> > (SVN_VERSION): Include the tagged version number before the parenthesis.
> >
> > * subversion/tests/cmdline/getopt_tests.py
> > (rep_lines_res): Adjust comment and regex for an optional "-dev" part.
> > ------------------------------------------------------------------------
> >
> > To avoid changing any programmatic uses of the version string, including
> > in the user-agent string, I recommend not backporting r1084581.
> >
> > danielsh, pburba: you both voted +1 for backporting both those revs.
> > Want to reconsider?
>
> The difference is whether off-the-branch builds will report themselves
> as "1.7.0-dev (under development)" or "1.7.0 (under development)"...
> No, I'm happy with both of my +1 votes.
Gah. I need to remember that the changes I made to the version strings
-- showing "-dev" and/or "(under development)" -- don't affect release
builds at all, and dev-build users are by definition deveopers/testers,
and therefore there is little concern about making such changes. That's
why it was a safe change in the first place.
I'll remove my veto.
If the changes happen to upset the validity of user-agent strings or
whatever, so be it: it doesn't really matter.
Sorry for the noise.
- Julian
> > Another option is also backporting r1100424, of course, but that seems
> > OTT as the problem of people misunderstanding/misreporting the version
> > is neither new nor severe.
> >
> > - Julian
> >
> >
Re: New)(?) Subversion client user-agent string ugliness.
Posted by Daniel Shahaf <da...@elego.de>.
Julian Foad wrote on Mon, May 09, 2011 at 12:17:43 +0100:
> On Fri, 2011-05-06 at 21:47 -0400, C. Michael Pilato wrote:
> > On 05/06/2011 07:08 PM, Greg Stein wrote:
> > > On Fri, May 6, 2011 at 17:06, Hyrum K Wright <hy...@hyrumwright.org> wrote:
> > >> On Fri, May 6, 2011 at 4:01 PM, C. Michael Pilato <cm...@collab.net> wrote:
> > >>> Is there a convention in HTTP user-agent strings to use whitespace to
> > >>> delimit various bits of information carried in that string? I asked because
> > >>> our client strings now look like so in trunk co:
> > >>>
> > >>> "SVN/1.7.0-dev (under development) neon/0.28.2"
> > >>>
> > >>> I *think* the " (under development)" bit is new(ish) and the result of
> > >>> Julian's relatively recent tweaks to our version strings. I'm not so fond
> > >>> of its appearance in this location, but wanted to hear other's opinion on
> > >>> the matter.
> > >>>
> > >>> So ... what say you?
> > >>
> > >> Seems kinda redundant when used along with the '-dev' postfix in the
> > >> version number, says I.
> > >
> > > Agreed. We should switch the User-Agent string from using SVN_VERSION
> > > to SVN_VER_NUMBER.
> >
> > Okey dokey. r1100424. Thanks, guys.
>
> Good catch. Note that this consideration may affect the 1.6.x backport
> proposal for r1084575, r1084581.
>
> ------------------------------------------------------------------------
> r1084575 | julianfoad | 2011-03-23 13:15:11 +0000 (Wed, 23 Mar 2011) | 6 lines
>
> Change the descriptive version identifier text from '1.x.y (dev build)' to
> '1.x.y (under development)' to help clarify that it's not merely a build
> with debug symbols included or something like that.
>
> * subversion/include/svn_version.h
> (SVN_VER_TAG): Change.
> ------------------------------------------------------------------------
> r1084581 | julianfoad | 2011-03-23 13:31:26 +0000 (Wed, 23 Mar 2011) | 11 lines
>
> Change the descriptive version identifier text from '1.x.y (under
> development)' to '1.x.y-dev (under development)' to help clarify that it's
> not merely a build with debug symbols included or something like that. This
> is the second part of a change started in r1084575.
>
> * subversion/include/svn_version.h
> (SVN_VERSION): Include the tagged version number before the parenthesis.
>
> * subversion/tests/cmdline/getopt_tests.py
> (rep_lines_res): Adjust comment and regex for an optional "-dev" part.
> ------------------------------------------------------------------------
>
> To avoid changing any programmatic uses of the version string, including
> in the user-agent string, I recommend not backporting r1084581.
>
> danielsh, pburba: you both voted +1 for backporting both those revs.
> Want to reconsider?
>
The difference is whether off-the-branch builds will report themselves
as "1.7.0-dev (under development)" or "1.7.0 (under development)"...
No, I'm happy with both of my +1 votes.
Daniel
> Another option is also backporting r1100424, of course, but that seems
> OTT as the problem of people misunderstanding/misreporting the version
> is neither new nor severe.
>
> - Julian
>
>
Re: New)(?) Subversion client user-agent string ugliness.
Posted by Julian Foad <ju...@wandisco.com>.
On Fri, 2011-05-06 at 21:47 -0400, C. Michael Pilato wrote:
> On 05/06/2011 07:08 PM, Greg Stein wrote:
> > On Fri, May 6, 2011 at 17:06, Hyrum K Wright <hy...@hyrumwright.org> wrote:
> >> On Fri, May 6, 2011 at 4:01 PM, C. Michael Pilato <cm...@collab.net> wrote:
> >>> Is there a convention in HTTP user-agent strings to use whitespace to
> >>> delimit various bits of information carried in that string? I asked because
> >>> our client strings now look like so in trunk co:
> >>>
> >>> "SVN/1.7.0-dev (under development) neon/0.28.2"
> >>>
> >>> I *think* the " (under development)" bit is new(ish) and the result of
> >>> Julian's relatively recent tweaks to our version strings. I'm not so fond
> >>> of its appearance in this location, but wanted to hear other's opinion on
> >>> the matter.
> >>>
> >>> So ... what say you?
> >>
> >> Seems kinda redundant when used along with the '-dev' postfix in the
> >> version number, says I.
> >
> > Agreed. We should switch the User-Agent string from using SVN_VERSION
> > to SVN_VER_NUMBER.
>
> Okey dokey. r1100424. Thanks, guys.
Good catch. Note that this consideration may affect the 1.6.x backport
proposal for r1084575, r1084581.
------------------------------------------------------------------------
r1084575 | julianfoad | 2011-03-23 13:15:11 +0000 (Wed, 23 Mar 2011) | 6 lines
Change the descriptive version identifier text from '1.x.y (dev build)' to
'1.x.y (under development)' to help clarify that it's not merely a build
with debug symbols included or something like that.
* subversion/include/svn_version.h
(SVN_VER_TAG): Change.
------------------------------------------------------------------------
r1084581 | julianfoad | 2011-03-23 13:31:26 +0000 (Wed, 23 Mar 2011) | 11 lines
Change the descriptive version identifier text from '1.x.y (under
development)' to '1.x.y-dev (under development)' to help clarify that it's
not merely a build with debug symbols included or something like that. This
is the second part of a change started in r1084575.
* subversion/include/svn_version.h
(SVN_VERSION): Include the tagged version number before the parenthesis.
* subversion/tests/cmdline/getopt_tests.py
(rep_lines_res): Adjust comment and regex for an optional "-dev" part.
------------------------------------------------------------------------
To avoid changing any programmatic uses of the version string, including
in the user-agent string, I recommend not backporting r1084581.
danielsh, pburba: you both voted +1 for backporting both those revs.
Want to reconsider?
Another option is also backporting r1100424, of course, but that seems
OTT as the problem of people misunderstanding/misreporting the version
is neither new nor severe.
- Julian
Re: New)(?) Subversion client user-agent string ugliness.
Posted by Julian Foad <ju...@wandisco.com>.
On Fri, 2011-05-06 at 21:47 -0400, C. Michael Pilato wrote:
> On 05/06/2011 07:08 PM, Greg Stein wrote:
> > On Fri, May 6, 2011 at 17:06, Hyrum K Wright <hy...@hyrumwright.org> wrote:
> >> On Fri, May 6, 2011 at 4:01 PM, C. Michael Pilato <cm...@collab.net> wrote:
> >>> Is there a convention in HTTP user-agent strings to use whitespace to
> >>> delimit various bits of information carried in that string? I asked because
> >>> our client strings now look like so in trunk co:
> >>>
> >>> "SVN/1.7.0-dev (under development) neon/0.28.2"
> >>>
> >>> I *think* the " (under development)" bit is new(ish) and the result of
> >>> Julian's relatively recent tweaks to our version strings. I'm not so fond
> >>> of its appearance in this location, but wanted to hear other's opinion on
> >>> the matter.
> >>>
> >>> So ... what say you?
> >>
> >> Seems kinda redundant when used along with the '-dev' postfix in the
> >> version number, says I.
> >
> > Agreed. We should switch the User-Agent string from using SVN_VERSION
> > to SVN_VER_NUMBER.
>
> Okey dokey. r1100424. Thanks, guys.
r1100424 switched the string for Neon and Serf. r1102218 now does the
same for RA-svn.
- Julian
Re: New)(?) Subversion client user-agent string ugliness.
Posted by "C. Michael Pilato" <cm...@collab.net>.
On 05/06/2011 07:08 PM, Greg Stein wrote:
> On Fri, May 6, 2011 at 17:06, Hyrum K Wright <hy...@hyrumwright.org> wrote:
>> On Fri, May 6, 2011 at 4:01 PM, C. Michael Pilato <cm...@collab.net> wrote:
>>> Is there a convention in HTTP user-agent strings to use whitespace to
>>> delimit various bits of information carried in that string? I asked because
>>> our client strings now look like so in trunk co:
>>>
>>> "SVN/1.7.0-dev (under development) neon/0.28.2"
>>>
>>> I *think* the " (under development)" bit is new(ish) and the result of
>>> Julian's relatively recent tweaks to our version strings. I'm not so fond
>>> of its appearance in this location, but wanted to hear other's opinion on
>>> the matter.
>>>
>>> So ... what say you?
>>
>> Seems kinda redundant when used along with the '-dev' postfix in the
>> version number, says I.
>
> Agreed. We should switch the User-Agent string from using SVN_VERSION
> to SVN_VER_NUMBER.
Okey dokey. r1100424. Thanks, guys.
--
C. Michael Pilato <cm...@collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
Re: New)(?) Subversion client user-agent string ugliness.
Posted by Greg Stein <gs...@gmail.com>.
On Fri, May 6, 2011 at 17:06, Hyrum K Wright <hy...@hyrumwright.org> wrote:
> On Fri, May 6, 2011 at 4:01 PM, C. Michael Pilato <cm...@collab.net> wrote:
>> Is there a convention in HTTP user-agent strings to use whitespace to
>> delimit various bits of information carried in that string? I asked because
>> our client strings now look like so in trunk co:
>>
>> "SVN/1.7.0-dev (under development) neon/0.28.2"
>>
>> I *think* the " (under development)" bit is new(ish) and the result of
>> Julian's relatively recent tweaks to our version strings. I'm not so fond
>> of its appearance in this location, but wanted to hear other's opinion on
>> the matter.
>>
>> So ... what say you?
>
> Seems kinda redundant when used along with the '-dev' postfix in the
> version number, says I.
Agreed. We should switch the User-Agent string from using SVN_VERSION
to SVN_VER_NUMBER.
Cheers,
-g
Re: New)(?) Subversion client user-agent string ugliness.
Posted by Hyrum K Wright <hy...@hyrumwright.org>.
On Fri, May 6, 2011 at 4:01 PM, C. Michael Pilato <cm...@collab.net> wrote:
> Is there a convention in HTTP user-agent strings to use whitespace to
> delimit various bits of information carried in that string? I asked because
> our client strings now look like so in trunk co:
>
> "SVN/1.7.0-dev (under development) neon/0.28.2"
>
> I *think* the " (under development)" bit is new(ish) and the result of
> Julian's relatively recent tweaks to our version strings. I'm not so fond
> of its appearance in this location, but wanted to hear other's opinion on
> the matter.
>
> So ... what say you?
Seems kinda redundant when used along with the '-dev' postfix in the
version number, says I.
-Hyrum