Preserving CVS history

classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

Preserving CVS history

Cesar Strauss-2
Hi Keith,

On the ChangeLog of the legacy branch of mingw-org-wsl, you wrote:

> * w32api: New directory; use w32api-3.17-2 release as baseline for
> import; once again, original CVS history is NOT preserved.

> * mingwrt: New directory; use mingwrt-3.20-2 release as baseline
> for import; original CVS history is NOT preserved.

The authoritative CVS repository for mingwrt and w32api seems to be:

https://cygwin.com/viewvc/src/winsup/mingw/
https://cygwin.com/viewvc/src/winsup/w32api/

What do you think of importing the full CVS history as two separated git
repositories? It could be for historical purposes or even for starting
again from a clean state.

Regards,

Cesar

------------------------------------------------------------------------------
_______________________________________________
MinGW-dvlpr mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mingw-dvlpr
Reply | Threaded
Open this post in threaded view
|

Re: Preserving CVS history

Keith Marshall-3
Hi Cesar,

On 23/10/14 22:06, Cesar Strauss wrote:

> Hi Keith,
>
> On the ChangeLog of the legacy branch of mingw-org-wsl, you wrote:
>
>> * w32api: New directory; use w32api-3.17-2 release as baseline for
>> import; once again, original CVS history is NOT preserved.
>
>> * mingwrt: New directory; use mingwrt-3.20-2 release as baseline
>> for import; original CVS history is NOT preserved.
>
> The authoritative CVS repository for mingwrt and w32api seems to be:
>
> https://cygwin.com/viewvc/src/winsup/mingw/
> https://cygwin.com/viewvc/src/winsup/w32api/

Those are web view references; I don't think they represent actual
(current) CVS repositories.

> What do you think of importing the full CVS history as two separated git
> repositories? It could be for historical purposes or even for starting
> again from a clean state.

If it's possible, it may be worthwhile.  However, if you look at the log
messages on the HEAD commits, (as seen through the commit graphs for any
of the files), they suggest that the actual CVS directories have been
deleted.  I don't do git, so I don't know if it can import CVS history
from a web view representation of a defunct repository; perhaps you know
better?

--
Regards,
Keith.

------------------------------------------------------------------------------
_______________________________________________
MinGW-dvlpr mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mingw-dvlpr
Reply | Threaded
Open this post in threaded view
|

Re: Preserving CVS history

Cesar Strauss-2
On 23/10/2014 21:35, Keith Marshall wrote:
> On 23/10/14 22:06, Cesar Strauss wrote:
>> The authoritative CVS repository for mingwrt and w32api seems to be:
>>
>> https://cygwin.com/viewvc/src/winsup/mingw/
>> https://cygwin.com/viewvc/src/winsup/w32api/
>
> Those are web view references; I don't think they represent actual
> (current) CVS repositories.

It seems the underlying repository is accessible by rsync:

rsync -avz cygwin.com::src-cvs/src/winsup/w32api srcdir
rsync -avz cygwin.com::src-cvs/src/winsup/mingw srcdir

>> What do you think of importing the full CVS history as two separated git
>> repositories? It could be for historical purposes or even for starting
>> again from a clean state.
>
> If it's possible, it may be worthwhile.  However, if you look at the log
> messages on the HEAD commits, (as seen through the commit graphs for any
> of the files), they suggest that the actual CVS directories have been
> deleted.  I don't do git, so I don't know if it can import CVS history
> from a web view representation of a defunct repository; perhaps you know
> better?
>

I'm using this tool to convert directly from the raw CVS repositories to
Git:

http://cvs2svn.tigris.org/cvs2git.html

I'll just remove the last commit, which deleted the files.

Regards,

Cesar


------------------------------------------------------------------------------
_______________________________________________
MinGW-dvlpr mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mingw-dvlpr
Reply | Threaded
Open this post in threaded view
|

Re: Preserving CVS history

Keith Marshall-3
On 29/10/14 20:42, Cesar Strauss wrote:

> On 23/10/2014 21:35, Keith Marshall wrote:
>> On 23/10/14 22:06, Cesar Strauss wrote:
>>> The authoritative CVS repository for mingwrt and w32api seems to be:
>>>
>>> https://cygwin.com/viewvc/src/winsup/mingw/
>>> https://cygwin.com/viewvc/src/winsup/w32api/
>>
>> Those are web view references; I don't think they represent actual
>> (current) CVS repositories.
>
> It seems the underlying repository is accessible by rsync:
>
> rsync -avz cygwin.com::src-cvs/src/winsup/w32api srcdir
> rsync -avz cygwin.com::src-cvs/src/winsup/mingw srcdir

Good to know, thanks.

> I'm using this tool to convert directly from the raw CVS repositories to
> Git:
>
> http://cvs2svn.tigris.org/cvs2git.html
>
> I'll just remove the last commit, which deleted the files.

Seems good to me ... as a separate git repo; (rewriting history in the
existing WSL repo would not be a good idea).

--
Regards,
Keith.

------------------------------------------------------------------------------
_______________________________________________
MinGW-dvlpr mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mingw-dvlpr
Reply | Threaded
Open this post in threaded view
|

Re: Preserving CVS history

Cesar Strauss-2
On 10/29/2014 07:18 PM, Keith Marshall wrote:

> On 29/10/14 20:42, Cesar Strauss wrote:
>> On 23/10/2014 21:35, Keith Marshall wrote:
>>> On 23/10/14 22:06, Cesar Strauss wrote:
>> It seems the underlying repository is accessible by rsync:
>>
>> rsync -avz cygwin.com::src-cvs/src/winsup/w32api srcdir
>> rsync -avz cygwin.com::src-cvs/src/winsup/mingw srcdir
>
> Good to know, thanks.
>
>> I'm using this tool to convert directly from the raw CVS repositories to
>> Git:
>>
>> http://cvs2svn.tigris.org/cvs2git.html
>>
>> I'll just remove the last commit, which deleted the files.
>
> Seems good to me ... as a separate git repo; (rewriting history in the
> existing WSL repo would not be a good idea).
>

I converted the above subset of the Cygwin repository as two Git
repositories: mingwrt and w32api.

I used cvs-fast-export, roughly following this guide:

DVCS migration HOWTO
http://www.catb.org/esr/dvcs-migration-guide.html

The conversion is available temporarily at my SF user project at:

1) Web view:
https://sourceforge.net/u/cstrauss/mingwrt/ci/master/tree/
https://sourceforge.net/u/cstrauss/w32api/ci/master/tree/

2) Read-only git clone:
git clone git://git.code.sf.net/u/cstrauss/mingwrt
git clone git://git.code.sf.net/u/cstrauss/w32api

If you agree, I'll publish them both on the MinGW project.

It seems you resumed development of mingwrt under the legacy branch of
the WSL repository. Would it make sense to use the mingwrt repository
instead?

I noticed that the CVS repository contains newer commits that were never
released in the 3.x branch. For instance, the ChangeLog of mingwrt-3.20
ends in 08/23/2011 but the CVS history goes up to 08/08/2012. Likewise,
the ChangeLog of w32api-3.18 ends in 03/25/2011, but the CVS history
goes up to 08/10/2012. I suppose these additional commits can be
reverted in a subsequent commit, or erased altogether from the master
branch, if you find it convenient.

Regards,

Cesar


------------------------------------------------------------------------------
_______________________________________________
MinGW-dvlpr mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mingw-dvlpr
Reply | Threaded
Open this post in threaded view
|

Re: Preserving CVS history

Keith Marshall-3
On 09/11/14 21:54, Cesar Strauss wrote:
> On 10/29/2014 07:18 PM, Keith Marshall wrote:
>> On 29/10/14 20:42, Cesar Strauss wrote:
>>> It seems the underlying repository is accessible by rsync:
>>>
>>> rsync -avz cygwin.com::src-cvs/src/winsup/w32api srcdir
>>> rsync -avz cygwin.com::src-cvs/src/winsup/mingw srcdir
>
> I converted the above subset of the Cygwin repository as two Git
> repositories: mingwrt and w32api.

Thanks for doing this.

> I used cvs-fast-export, roughly following this guide:
>
> DVCS migration HOWTO
> http://www.catb.org/esr/dvcs-migration-guide.html

I know Eric, (yeah, ESR himself), did the conversion for GNU troff from
CVS to git, using the tools he describes there, and that all went quite
smoothly.  If I get a chance, over the next few days, and if I can find
my old CVS sandboxes, I'll clone these...

> The conversion is available temporarily at my SF user project at:
>
> 1) Web view:
> https://sourceforge.net/u/cstrauss/mingwrt/ci/master/tree/
> https://sourceforge.net/u/cstrauss/w32api/ci/master/tree/
>
> 2) Read-only git clone:
> git clone git://git.code.sf.net/u/cstrauss/mingwrt
> git clone git://git.code.sf.net/u/cstrauss/w32api

...and check them for consistency; can't promise that I can do better
than checking, at the appropriate release points, against historically
released tarballs, though; nor can I promise a quick turn-around.

> If you agree, I'll publish them both on the MinGW project.

If I don't get back to you, within two weeks at most, please do.

> It seems you resumed development of mingwrt under the legacy branch of
> the WSL repository. Would it make sense to use the mingwrt repository
> instead?

Yes, I think it would.  Even if we do eventually merge it with w32api,
to create a composite similar to WSL, we should maintain the two as
separate sub-repositories of a common overall project; WSL, as it stands
now, appears to be a disaster area best vacated.

> I noticed that the CVS repository contains newer commits that were never
> released in the 3.x branch. For instance, the ChangeLog of mingwrt-3.20
> ends in 08/23/2011 but the CVS history goes up to 08/08/2012.

So, there is some work which never made it into a mingwrt-3.x release;
some of it may even be my own.  We probably want to keep that, or at
least consider merging it with our own work, from after we abandoned the
Cygwin CVS.

> Likewise, the ChangeLog of w32api-3.18 ends in 03/25/2011, but the
> CVS history goes up to 08/10/2012.

I thought our last formal release was w32api-3.17; did Cygwin release
w32api-3.18, before they started to adopt mingw-w64?

> I suppose these additional commits can be reverted in a subsequent
> commit, or erased altogether from the master branch, if you find it
> convenient.

Better to review, and merge as appropriate; it would be a shame to throw
away potentially useful work.

--
Regards,
Keith.

------------------------------------------------------------------------------
_______________________________________________
MinGW-dvlpr mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mingw-dvlpr
Reply | Threaded
Open this post in threaded view
|

Re: Preserving CVS history

Cesar Strauss
On 10/11/2014 01:01, Keith Marshall wrote:

> On 09/11/14 21:54, Cesar Strauss wrote:
> If I get a chance, over the next few days, and if I can find
> my old CVS sandboxes, I'll clone these...
>
>> git clone git://git.code.sf.net/u/cstrauss/mingwrt
>> git clone git://git.code.sf.net/u/cstrauss/w32api
>
> ...and check them for consistency; can't promise that I can do better
> than checking, at the appropriate release points, against historically
> released tarballs, though; nor can I promise a quick turn-around.

Thanks. Don't worry, take your time.

>
>> If you agree, I'll publish them both on the MinGW project.
>
> If I don't get back to you, within two weeks at most, please do.
>

I'll wait.

>> Likewise, the ChangeLog of w32api-3.18 ends in 03/25/2011, but the
>> CVS history goes up to 08/10/2012.
>
> I thought our last formal release was w32api-3.17; did Cygwin release
> w32api-3.18, before they started to adopt mingw-w64?
>

Sorry about the typo, I meant w32api-3.17.

>> I suppose these additional commits can be reverted in a subsequent
>> commit, or erased altogether from the master branch, if you find it
>> convenient.
>
> Better to review, and merge as appropriate; it would be a shame to throw
> away potentially useful work.
>

Indeed.

Regards,

Cesar

------------------------------------------------------------------------------
Comprehensive Server Monitoring with Site24x7.
Monitor 10 servers for $9/Month.
Get alerted through email, SMS, voice calls or mobile push notifications.
Take corrective actions from your mobile device.
http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk
_______________________________________________
MinGW-dvlpr mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mingw-dvlpr
Reply | Threaded
Open this post in threaded view
|

Re: Preserving CVS history

Keith Marshall-3
On 16/11/14 16:55, Cesar Strauss wrote:

> On 10/11/2014 01:01, Keith Marshall wrote:
>> On 09/11/14 21:54, Cesar Strauss wrote:
>> If I get a chance, over the next few days, and if I can find
>> my old CVS sandboxes, I'll clone these...
>>
>>> git clone git://git.code.sf.net/u/cstrauss/mingwrt
>>> git clone git://git.code.sf.net/u/cstrauss/w32api
>>
>> ...and check them for consistency; can't promise that I can do better
>> than checking, at the appropriate release points, against historically
>> released tarballs, though; nor can I promise a quick turn-around.
>
> Thanks. Don't worry, take your time.
I've now done an 'hg clone' on both, from which I've generated the
attached log files, and rolled the working copies back to reflect the
state at the mingwrt-3.20 and w32api-3.17 release points, respectively:

  $ hg -R mingwrt-hg co -r 562
  $ hg -R w32api-hg co -r 1118

Some comments, arising from subsequent comparison with WSL-legacy:

1) Both packages exhibit a significant number of files with extraneous
   (trailing) white-space; I removed this from the copies of all such
   files, when I imported them into WSL-legacy.

2) mingwrt has several files with malformed RCS keywords, (mea culpa);
   I also corrected these, when importing to WSL-legacy.

3) The readme.txt file in mingwrt is littered with instances of the
   (non-ASCII) closing single quote character; in each instance, the
   ASCII apostrophe *should* have been used, and I corrected this in
   the copy which I imported into WSL-legacy.

4) An oddity in mingwrt: changeset 543 introduced changes in function
   __dyn_tls_init(), which were then incorrectly reverted by changeset
   558, and subsequently corrected again by changeset 566, (apparently
   *after* the mingwrt-3.20 release).  The repository thus shows a bad
   tlssup.c at point of release, yet the released source tarball, (on
   which I based my import to WSL-legacy), incorporates the subsequent
   correction; however, it is lacking the changes introduced by the
   intervening changesets, 563..565.

5) The samples directory of mingwrt is present in the repository, but
   omitted from the source tarball; (a packaging error, perhaps?)  Thus,
   this directory was not imported into WSL-legacy.

6) The repository provides a mingwex/math/pow.c file; this seems to be
   redundant, since MinGW uses the MSVCRT.DLL implementation of pow().

7) The repository provides two files: mingwex/gdtoa/README.mingw, and
   mingwex/gdtoa/gdtoa_fltrnds.h, which are not present in the mingwrt
   source tarball, and are consequently omitted from WSL-legacy.

8) The w32api repository includes a CheckConflicts.sh script which is
   not included in the source tarball; thus, this too is missing from
   WSL-legacy.

--
Regards,
Keith.

------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=157005751&iu=/4140/ostg.clktrk
_______________________________________________
MinGW-dvlpr mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mingw-dvlpr

mingwrt-hg.log (256K) Download Attachment
w32api-hg.log (608K) Download Attachment