Обсуждение: Broken source tree

I am (not) finding a missing file src/interfaces/libpq/fe-print.c in the
current CVSup'd cvs source tree, and psql is having trouble building
because it is missing PQprint().

I assume they are related and that someone intended PQprint() to move
from fe-exec.c to a new file fe-print.c.

Is this the case? If so, where is fe-print.c??

Assuming that I can get the tree to build, what is the state of the
tree? Do the regression tests pass? Are there any known breakages?? I'd
like to move my type conversion development up from my 980408 rev-locked
tree but not if it won't run :/

                       - Tom

>
> I am (not) finding a missing file src/interfaces/libpq/fe-print.c in the
> current CVSup'd cvs source tree, and psql is having trouble building
> because it is missing PQprint().
>
> I assume they are related and that someone intended PQprint() to move
> from fe-exec.c to a new file fe-print.c.
>
> Is this the case? If so, where is fe-print.c??
>
> Assuming that I can get the tree to build, what is the state of the
> tree? Do the regression tests pass? Are there any known breakages?? I'd
> like to move my type conversion development up from my 980408 rev-locked
> tree but not if it won't run :/

Fixed.  It was a new file from the Tom Lane patch that I missed a 'cvs
add'.  Don't know that status, but I think things still run with this
new patch.

--
Bruce Momjian                          |  830 Blythe Avenue
maillist@candle.pha.pa.us              |  Drexel Hill, Pennsylvania 19026
  +  If your life is a hard drive,     |  (610) 353-9879(w)
  +  Christ can be your backup.        |  (610) 853-3000(h)

> I am (not) finding a missing file src/interfaces/libpq/fe-print.c in the
> current CVSup'd cvs source tree, and psql is having trouble building
> because it is missing PQprint().
> I assume they are related and that someone intended PQprint() to move
> from fe-exec.c to a new file fe-print.c.

Yes, I moved the print functions out of fe-exec, because they bulked it
up to an unreasonable size, and because I felt that an application not
using them should not be forced to link them in.

Looks like Bruce forgot to check in the new file when he applied the
libpq patches I sent him.  Sorry about the glitch.

> Assuming that I can get the tree to build, what is the state of the
> tree? Do the regression tests pass? Are there any known breakages??

As far as the libpq changes go: the regression tests pass.  The gripes
I've been sending about missing regression tests have to do with things
that still didn't work after the regression tests all passed ;-).  AFAIK
there is nothing broken now.

As to what other people may have been breaking, this deponent sayeth
not.

            regards, tom lane

> Yes, I moved the print functions out of fe-exec,
> As far as the libpq changes go: the regression tests pass.
> AFAIK there is nothing broken now.

Thanks Bruce I have the file now :)

There is one compiler warning in fe-print.c about a missing declaration
for ioctl(); the includes should have <sys/ioctl.h> as did the old
fe-exec.c ??

                   - Tom

>
> > Yes, I moved the print functions out of fe-exec,
> > As far as the libpq changes go: the regression tests pass.
> > AFAIK there is nothing broken now.
>
> Thanks Bruce I have the file now :)
>
> There is one compiler warning in fe-print.c about a missing declaration
> for ioctl(); the includes should have <sys/ioctl.h> as did the old
> fe-exec.c ??

Fixed.  fe-misc.c was missing errno.h on my OS too.  I want to make
those backend changes so people can start testing the new features.

--
Bruce Momjian                          |  830 Blythe Avenue
maillist@candle.pha.pa.us              |  Drexel Hill, Pennsylvania 19026
  +  If your life is a hard drive,     |  (610) 353-9879(w)
  +  Christ can be your backup.        |  (610) 853-3000(h)