[M3devel] Checkins

[Jay]

hm, cvs release, didn't know about that.
I thought I had been much slower actually.
I added a file first, to better show history, then one well tested, on one platform, thought to be coherent checkin, then updated my Mac and found the missing file, fixed it, etc. There was also a missing file for NT386GNU that nobody would notice. I also set up my testing to build the more-Posix path on NT386GNU to be sure it built and ran. I am not much into the habit of testing on more than one machine/platform before submitting, though sometimes, and actually NT386GNU makes that easier -- two platforms on one machine. I am more in the habit of testing other machines right after commit.
I haven't found a convenient way to move changes..cvs diff into a file, copy that, and patch? Or maybe share out the tree over the network and build in one tree directly? I should probably try that..
 
And I can't checkin from the top as I have other definitely unrelated unfinished un-affecting changes, true that sets a trap for myself.
 
Without the commit mail, what do you run to see recent commits?
cvs filelog likes to default to showing the full history.
I'd like to say, just show the last 10 commits, or the last 20 commits..
I'm operating relatively blindly wrt cvs skills I'm afraid.
 
And you manually check the tinderbox pages? They don't send mail?
 
I do run cvs diff but I have a bunch of junk..gotta clean it up..
 
? bin? caltech-parser/parserlib/.parsertop? caltech-parser/parserlib/html/SeekRd.html? caltech-parser/parserlib/html/calc.html? caltech-parser/parserlib/html/index.html? caltech-parser/parserlib/html/kext.html? caltech-parser/parserlib/html/klex.html? caltech-parser/parserlib/html/ktok.html? caltech-parser/parserlib/html/kyacc.html? caltech-parser/parserlib/html/lr.html? caltech-parser/parserlib/html/m3build.html? caltech-parser/parserlib/html/macroizations? m3-db/smalldb/src/mine? m3-db/smalldb/src/orig? m3-libs/arithmetic/NT386? m3-libs/arithmetic/NT386GNU? m3-libs/commandrw/NT386? m3-libs/commandrw/SOLgnu? m3-libs/deepcopy/NT386? m3-libs/libm3/src/os/NT386GNU? m3-libs/libm3/src/os/POSIX/ProcessCygwin.m3 - Jay


CC: m3devel at elegosoft.com; jkrell at elego.deFrom: hosking at cs.purdue.eduTo: jayk123 at hotmail.comSubject: Re: [M3devel] CheckinsDate: Mon, 24 Mar 2008 12:43:31 -0400
I didn't take the time to look to closely at what caused the failure, but I did notice a smear of checkins spread over time.  Better to roll it all into one checkin.  I usually do a cvs release to see if there are files (marked by ?) that I forgot to add, and add them, before doing the big checkin as one top-level command.

On Mar 24, 2008, at 12:37 PM, Jay wrote:

Tony, The problem was I forgot to add a file, right?That's something the tools are setup to make you fail at as I understand. Right? I have to remember to cvs add whenever I create a file. Building works either way.After commiting, I updated another machine, noticed the missing file, added it. I thought other than the forgotten file, the changes were coherent, at least for non-NT386GNU.  - Jay


From: hosking at cs.purdue.eduTo: m3devel at elegosoft.comDate: Mon, 24 Mar 2008 10:36:56 -0400CC: jkrell at elego.deSubject: [M3devel] Checkins
It seems that the reason things failed is because a checkin of multiple related changes was not done as a single atomic commit.  This meant that partial checkout of inconsistent code occurred while bits and pieces of a coherent set of changes were being singly checked in.  Please try to bulk checkin related changes so as to avoid these sorts of inconsistencies.  cvs does a great job of bundling an entire set of changes as one commit.  It also has the nice side-effect of not bombarding folks with zillions of commit e-mails -- instead the bundled commit comes in one message.

If we use the tools properly, then everyone can play nicely together.


Antony Hosking | Associate Professor | Computer Science | Purdue University
305 N. University Street | West Lafayette | IN 47907 | USA
Office +1 765 494 6001 | Mobile +1 765 427 5484
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://m3lists.elegosoft.com/pipermail/m3devel/attachments/20080324/de9fde52/attachment-0002.html>