|marc,
|
|I have simple suggestion. Can we modify the commit script to check that if
|anyone changes the 'special files' it sends an additional message to the
|list with the subject 'DANGER' or something simmilar. Then in the code we
|put the comment (in addition to what is already there) mentions that every
|one will get a DANGER message and will be very upset if the change was not
|approved by marc and bill.
well, Bill brought me back to earth.
Maybe locking is a bit extreme. I now hope this will be self organizing.
There are more people familiar with the core stuff that have spent time in
it and know the issues. The first one is of course Scott that knows this
part by heart and in fact debugged my commits some time ago.
What I am trying to say is that I appreciate your concern and trying to find
a solution in return I trust that we can work around these critical pieces
with the old trust system with the understanding that you really should not
commit on your own in these pieces (as opposed to pretty much every other
part of the code where you are encourage to commit).
most parts won't get "review" out of the box but only under usage, I am
thinking mostly of Vinay Menon here that wouldn't commit without the code he
wrote being first fully reviewed :) that was noble but unrealistic, people
do review when they use it. The 3 interceptor classes are an exception
where you really need approval, and in fact should not commit yourself.
that's all
|Just an idea.
thanks don't sweat it, over and out,
marcf
|
|-dain
|
|----- Original Message -----
|From: "marc fleury" <[EMAIL PROTECTED]>
|To: <[EMAIL PROTECTED]>
|Sent: Tuesday, August 07, 2001 10:51 AM
|Subject: RE: [JBoss-dev] synchronizeEntitiesWithinTransaction, null pk &
|remove
|
|
|> |I really disagree on locking the code in cvs. I'd prefer a solid
|> |test suite
|> |that tests this stuff backward and forward so that if somebody checks in
|> |bogus code, we will know who it is. We should encourage people to look
|at
|> |this code. The more eyes the better this code will be and get.
|> |Locking the
|> |code in CVS just discourages people from taking a look at it.
|>
|> I agree and disagree with your disagreeing.
|>
|> disagree: It seems to me that tagging the code as critical in
|fact attract
|> every developer like flies to a lightbulb. Not good
|> agree: I WANT more people to be familiar with it, because as you
|well know
|I
|> was on the verge of losing it when I realized that after 8 weeks
|of work I
|> went back to square one with logical steps in between.
|>
|> The probability space on that code is a big circle and it is
|dangerous for
|> one's sanity. I don't want people changing it before graduating
|to it and
|> the only way we can ensure that is by locking the interceptor code since
|it
|> seems the warnings are not respected? Otherwise it is I and
|more recently
|> you that have to go down and fix it... nightmare all over again.
|>
|> You seem to be more at ease with it than I am and I can't wait to find
|time
|> to review your N-Lock revision of mine.
|>
|> |BTW, I'll fix this stuff as soon as possible(I know the fix), I keep on
|> |getting pulled for stuff at the "real" job.
|>
|> great, thanks
|>
|> marcf
|> |
|> |Bill
|> |
|> |> -----Original Message-----
|> |> From: [EMAIL PROTECTED]
|> |> [mailto:[EMAIL PROTECTED]]On
|Behalf Of marc
|> |> fleury
|> |> Sent: Tuesday, August 07, 2001 11:12 AM
|> |> To: [EMAIL PROTECTED]
|> |> Subject: RE: [JBoss-dev] synchronizeEntitiesWithinTransaction, null pk
|&
|> |> remove
|> |>
|> |>
|> |> |bill,
|> |> |
|> |> |between you and I,
|> |>
|> |> It was between Bill and I and then I send the message to everyone.
|> |>
|> |> I am not even going to apologize for this since in fact it is better
|this
|> |> way.
|> |>
|> |> I will repeat that NO-ONE should touch the interceptor sync stuff (the
|3
|> |> core interceptor classes) without a deep review from top
|> |developers first.
|> |> I really recommend locking the code in cvs if that feature
|> |exists. It was
|> |> the Locking-waiting buzy wait bug, it was a ton of missed
|locks it is a
|> |> nightmare, even for me (who spends most time in it). don't touch
|> |> it please.
|> |>
|> |> So it is better that it is in the open
|> |>
|> |> marcf
|> |>
|> |> |this is typically the type of problem I want to avoid
|> |> |with the core container interceptors. I know (pretty sure) I
|> |didn't miss
|> |> |something as basic as that and obviously it wasn't you. If it is
|> |> |me then it
|> |> |only makes the following point even more true.
|> |> |
|> |> |I put "you must be this tall to ride the dragon" messages in the
|> |> |code but it
|> |> |seems to have no effect in fact it seems to attract more people
|> |> wanting to
|> |> |take the challenge. You and I know the nightmare that solid synch is
|(I
|> |> |almost went crazy with it) so it *enfuriates* me when someone comes
|and
|> |> |changes it without paying attention. Can you track down the
|> |> change (still
|> |> |compiling painfully in RH) and let me know and will send a "friendly
|> |> |reminder privately". :( It could very well be me because I
|> |> honestly don't
|> |> |remember.
|> |> |
|> |> |thanks for being on top of this, I am just tired of going back
|> |> and back to
|> |> |it. It is a waste of our time.
|> |> |
|> |> |we should do it once and for all and then **lock** the 3 classes
|> |> |code in CVS
|> |> |:)
|> |> |
|> |> |marcf
|> |> |
|> |> ||-----Original Message-----
|> |> ||From: [EMAIL PROTECTED]
|> |> ||[mailto:[EMAIL PROTECTED]]On Behalf Of
|Bill
|> |> ||Burke
|> |> ||Sent: Tuesday, August 07, 2001 10:34 AM
|> |> ||To: [EMAIL PROTECTED]
|> |> ||Subject: RE: [JBoss-dev] synchronizeEntitiesWithinTransaction,
|> |null pk &
|> |> ||remove
|> |> ||
|> |> ||
|> |> ||
|> |> ||
|> |> ||> -----Original Message-----
|> |> ||> From: [EMAIL PROTECTED]
|> |> ||> [mailto:[EMAIL PROTECTED]]On
|> |> Behalf Of Jason
|> |> ||> Dillon
|> |> ||> Sent: Monday, August 06, 2001 9:19 PM
|> |> ||> To: [EMAIL PROTECTED]
|> |> ||> Subject: [JBoss-dev] synchronizeEntitiesWithinTransaction, null pk
|&
|> |> ||> remove
|> |> ||>
|> |> ||>
|> |> ||> I am running into a problem with entity bean removal where the
|> |> ||pk is null,
|> |> ||> which causes most everything passed (and including)
|> |> ||> Application.synchronizeEntitiesWithinTransaction() to fail
|> |> (called from
|> |> ||> EntityContainer.remove()).
|> |> ||>
|> |> ||
|> |> ||The person who moved synchronizeEntitiesWithinTransaction forgot
|> |> |to perform
|> |> ||that same logic that is
|> |EntitySynchronizationInterceptor.storeEntity. I
|> |> ||should have caught this when I reviewed the code. Sorry, I'll
|> |> fix it now.
|> |> ||
|> |> ||BTW, pk get's set to null when a remove() is called.
|> |> ||
|> |> ||Bill
|> |> ||
|> |> ||
|> |> ||
|> |> ||_______________________________________________
|> |> ||Jboss-development mailing list
|> |> ||[EMAIL PROTECTED]
|> |> ||http://lists.sourceforge.net/lists/listinfo/jboss-development
|> |> |
|> |> |
|> |> |_______________________________________________
|> |> |Jboss-development mailing list
|> |> |[EMAIL PROTECTED]
|> |> |http://lists.sourceforge.net/lists/listinfo/jboss-development
|> |>
|> |>
|> |> _______________________________________________
|> |> Jboss-development mailing list
|> |> [EMAIL PROTECTED]
|> |> http://lists.sourceforge.net/lists/listinfo/jboss-development
|> |>
|> |
|> |
|> |
|> |_______________________________________________
|> |Jboss-development mailing list
|> |[EMAIL PROTECTED]
|> |http://lists.sourceforge.net/lists/listinfo/jboss-development
|>
|>
|> _______________________________________________
|> Jboss-development mailing list
|> [EMAIL PROTECTED]
|> http://lists.sourceforge.net/lists/listinfo/jboss-development
|>
|
|
|_______________________________________________
|Jboss-development mailing list
|[EMAIL PROTECTED]
|http://lists.sourceforge.net/lists/listinfo/jboss-development
_______________________________________________
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
