Обсуждение: RE: cvs-commit-digest V1 #856
> -----Original Message----- > From: owner-cvs-commit-digest@hub.org > [mailto:owner-cvs-commit-digest@hub.org] > Sent: Sunday, January 30, 2000 1:01 PM > To: committers-digest@hub.org > Subject: cvs-commit-digest V1 #856 > > ------------------------------ > > Date: Sat, 29 Jan 2000 11:58:52 -0500 (EST) > From: Peter Eisentraut - PostgreSQL <petere> > Subject: [COMMITTERS] pgsql/src/interfaces/libpq (fe-misc.c > fe-print.c libpq-fe.h) > > Date: Saturday, January 29, 2000 @ 11:58:51 > Author: petere > > Update of /usr/local/cvsroot/pgsql/src/interfaces/libpq > from hub.org:/home/tmp/cvs-serv53967/src/interfaces/libpq > > Modified Files: > fe-misc.c fe-print.c libpq-fe.h > > - ----------------------------- Log Message > ----------------------------- > > A few minor psql enhancements > Initdb help correction > Changed end/abort to commit/rollback and changed related notices > Commented out way old printing functions in libpq > Fixed a typo in alter table / alter column > pqbool is removed from libpq-fe.h. Couldn't compile interfaces/perl5 now. In addition,this seems to change external interface of PQprint(). Is it OK ? Regards. Hiroshi Inoue Inoue@tpf.co.jp
"Hiroshi Inoue" <Inoue@tpf.co.jp> writes:
>> From: Peter Eisentraut - PostgreSQL <petere>
>> Commented out way old printing functions in libpq
> pqbool is removed from libpq-fe.h.
> Couldn't compile interfaces/perl5 now.
> In addition,this seems to change external interface of PQprint().
> Is it OK ?
Not IMHO. It looks like Peter has removed typedef pqbool (potentially
breaking application sources, not just perl5) and changed what were
pqbool == char fields into int fields (thereby breaking binaries that
depend on shared libraries of libpq). Not to mention the advertised
change of removing documented API entry points.
Peter, you need to have a little more respect for stability of
library APIs. Gratuitous breaking of backwards compatibility
is not the done thing around here. It's especially not done
without any discussion.
regards, tom lane
At 10:52 PM 2/1/00 -0500, Tom Lane wrote: >Peter, you need to have a little more respect for stability of >library APIs. Gratuitous breaking of backwards compatibility >is not the done thing around here. It's especially not done >without any discussion. I thought we went over this a week ago...was I dreaming? PG is intended to be a PROFESSIONAL product. You don't arbitrarily break things for the hell of it. PG has CUSTOMERS. Not in the formal "we bought it" sense, but in the moral and professional engineering sense. You don't screw your customers without good reason, and when you do you at least provide them cushions and soft mattresses and advance notice. Especially advance notice. And if you do screw them, you do so after you explore alternatives and come to realize that there is no other course open to you. And you offer them a condom (i.e. an upgrade path). Because they depend on you. Is professionalism so hard to understand? - Don Baccus, Portland OR <dhogaza@pacifier.com> Nature photos, on-line guides, Pacific Northwest Rare Bird Alert Serviceand other goodies at http://donb.photo.net.
On Tue, 1 Feb 2000, Don Baccus wrote:
> At 10:52 PM 2/1/00 -0500, Tom Lane wrote:
>
> >Peter, you need to have a little more respect for stability of
> >library APIs. Gratuitous breaking of backwards compatibility
> >is not the done thing around here. It's especially not done
> >without any discussion.
>
> I thought we went over this a week ago...was I dreaming?
>
> PG is intended to be a PROFESSIONAL product. You don't arbitrarily
> break things for the hell of it.
>
> PG has CUSTOMERS. Not in the formal "we bought it" sense, but in
> the moral and professional engineering sense.
>
> You don't screw your customers without good reason, and when you
> do you at least provide them cushions and soft mattresses and
> advance notice. Especially advance notice. And if you do screw
> them, you do so after you explore alternatives and come to realize
> that there is no other course open to you. And you offer them
> a condom (i.e. an upgrade path).
>
> Because they depend on you.
>
> Is professionalism so hard to understand?
Don ... I try to stay out of stuff like this but ... TONE IT DOWN!
Peter is making mistakes, granted, but he is making them in a *NONE
PRODUCTION RELEASE* code tree ... if he messes with a -STABLE release in
this way, fine, your responses are justified, but, right now, I don't
think they are ...
Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
On Tue, Feb 01, 2000 at 09:34:17PM -0800, Don Baccus wrote: > At 10:52 PM 2/1/00 -0500, Tom Lane wrote: > <deleted to save having to read that again> Ah Don, that's a little harsh, isn't it? Did Peter's actions in checking in code lose you one minutes work, either time? I seem to recall that your not tracking the CVS (which you shouldn't). So your basically bitching about _theoretical_ problems? Tom, as a core developer, is directly affected, and has earned the right to chew out Peter. You, on the other hand, are a kibitzer here, as am I. Your comments in technical discussions lead me to believe that you are a professional developer, but you haven't stepped up to the plate for postgresql, yet, and submitted code. Let the current core developers deal with this: we all know your position! Ross -- Ross J. Reedstrom, Ph.D., <reedstrm@rice.edu> NSBRI Research Scientist/Programmer Computer and Information Technology Institute Rice University, 6100 S. Main St., Houston, TX 77005
> > Is professionalism so hard to understand? > > Don ... I try to stay out of stuff like this but ... TONE IT DOWN! > > Peter is making mistakes, granted, but he is making them in a *NONE > PRODUCTION RELEASE* code tree ... if he messes with a -STABLE release in > this way, fine, your responses are justified, but, right now, I don't > think they are ... I also told Peter that 7.0 was a good time to remove routines that were no longer needed. Yes, it is a migration problem, but why drag around functions forever that are useless. Maybe he pulled one too many? -- Bruce Momjian | http://www.op.net/~candle pgman@candle.pha.pa.us | (610) 853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania19026
> On Tue, Feb 01, 2000 at 09:34:17PM -0800, Don Baccus wrote: > > At 10:52 PM 2/1/00 -0500, Tom Lane wrote: > > > <deleted to save having to read that again> > > Ah Don, that's a little harsh, isn't it? Did Peter's actions in checking in > code lose you one minutes work, either time? I seem to recall that your not > tracking the CVS (which you shouldn't). So your basically bitching about > _theoretical_ problems? Tom, as a core developer, is directly affected, > and has earned the right to chew out Peter. You, on the other hand, are > a kibitzer here, as am I. Your comments in technical discussions lead > me to believe that you are a professional developer, but you haven't > stepped up to the plate for postgresql, yet, and submitted code. Let the > current core developers deal with this: we all know your position! This is excellent advise. -- Bruce Momjian | http://www.op.net/~candle pgman@candle.pha.pa.us | (610) 853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania19026
Bruce Momjian <pgman@candle.pha.pa.us> writes:
> I also told Peter that 7.0 was a good time to remove routines that were
> no longer needed. Yes, it is a migration problem, but why drag around
> functions forever that are useless. Maybe he pulled one too many?
We have in fact talked about removing some of the older-generation
print functions (though I was envisioning a slow process of labeling
them deprecated for a few releases...). It was the quite unnecessary
modification of the exported PQprintOpt struct that got my Irish up.
I've fought way too many hard-to-debug crashes caused by that sort
of change.
regards, tom lane
On Wed, 2 Feb 2000, Hiroshi Inoue wrote: > pqbool is removed from libpq-fe.h. > Couldn't compile interfaces/perl5 now. > > In addition,this seems to change external interface of PQprint(). > Is it OK ? Darn, seems like I'm doing everything wrong these days. I gotta take some time off to get my wits together. This is not anyone's fault out there, maybe I just wasn't ready quite yet. I don't want to be the problem person. I'll be back. (psql quoting bug will be fixed.) -- Peter Eisentraut Sernanders vaeg 10:115 peter_e@gmx.net 75262 Uppsala http://yi.org/peter-e/ Sweden
At 12:21 AM 2/2/00 -0600, Ross J. Reedstrom wrote: >On Tue, Feb 01, 2000 at 09:34:17PM -0800, Don Baccus wrote: >> At 10:52 PM 2/1/00 -0500, Tom Lane wrote: >> ><deleted to save having to read that again> > >Ah Don, that's a little harsh, isn't it? Yes, it is, and I apologize. - Don Baccus, Portland OR <dhogaza@pacifier.com> Nature photos, on-line guides, Pacific Northwest Rare Bird Alert Serviceand other goodies at http://donb.photo.net.
"Ross J. Reedstrom" wrote: > Ah Don, that's a little harsh, isn't it? Yes, it was a little harsh, and Don apologized. > Did Peter's actions in checking in > code lose you one minutes work, either time? I seem to recall that your not > tracking the CVS (which you shouldn't). Wrong -- Don is one of the core (or lead) developers porting the ArsDigita Community System from Oracle to PostgreSQL -- and in order to do this he has indeed been tracking the CVS -- and in fact he is running the pre-beta PostgreSQL 7 right now on a site with the pre-pre-beta ACS port to PostgreSQL running on the beta AOLserver 3.0. He also is a major maintainer of the AOLserver driver for postgresql -- which could be directly impacted by these changes. So, even though Don hasn't been a heavy contributor here as yet, I believe that he has a right to let his position be known -- although a little more gently, perhaps. Why does he need to do this? Two words: Referential Integrity, which is heavily used by the ACS. I also track the current CVS -- but for a totally different reason, as I want to be able to release RPMs of the beta release the same day as the beta release -- thus, I am doing trial builds of RPM's against the CVS. However, this current issue doesn't impact me in the slightest -- which is why I have not and will not say anything about it. -- Lamar Owen WGCR Internet Radio 1 Peter 4:11
On Wed, Feb 02, 2000 at 01:24:07PM -0500, Lamar Owen wrote: > "Ross J. Reedstrom" wrote: > > Ah Don, that's a little harsh, isn't it? > > Yes, it was a little harsh, and Don apologized. > > > Did Peter's actions in checking in > > code lose you one minutes work, either time? I seem to recall that your not > > tracking the CVS (which you shouldn't). > > Wrong -- Don is one of the core (or lead) developers porting the (details of what Don's up to) Ah, now it's time for me to apologize. As I said, *I'm* the kibitzer here, so I'll shut up now. Ross -- Ross J. Reedstrom, Ph.D., <reedstrm@rice.edu> NSBRI Research Scientist/Programmer Computer and Information Technology Institute Rice University, 6100 S. Main St., Houston, TX 77005