Обсуждение: BUG #14203: pgindent requires pg_bsd_indent 1.2, but only 1.3 available for download
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sthomas@datawareventures.com writes:
> I hacked pgindent to accept 1.3. This time it ran, but changed lots and lots
> of files. (I assume because pg_bsd_indent 1.3 makes different choices than
> pg_bsd_indent 1.2.)
> Does anyone know where I can find pg_bsd_indent 1.2?

I'm not sure why we're not keeping the older tarballs online, but there's
not that much difference between 1.2 and 1.3 --- see attached diff.

I think your problem may be in expecting that 9.3.x is pgindent-clean.
We don't normally re-pgindent stable branches, so unless committers are
careful to manually indent whatever they back-patch, branches will drift
away from a clean state.  If you're trying to validate that you have a
non-broken build of pgindent, I'd suggest checking out the immediately
post-pgindent state from our git repo.  The 9.3 run was at commit
9af4159fc.

            regards, tom lane

diff -cr pg_bsd_indent-1.2/Makefile pg_bsd_indent-1.3/Makefile
*** pg_bsd_indent-1.2/Makefile    Wed Oct 12 12:17:12 2011
--- pg_bsd_indent-1.3/Makefile    Thu Jan 30 22:32:50 2014
***************
*** 31,35 ****
--- 31,38 ----
  clean:
      rm -f *.o $(TARGET) log core

+ distclean: clean
+     rm -f tags ID
+
  install: $(TARGET)
      install -s -o bin -g bin $(TARGET) /usr/local/bin
diff -cr pg_bsd_indent-1.2/args.c pg_bsd_indent-1.3/args.c
*** pg_bsd_indent-1.2/args.c    Mon Aug 27 09:25:53 2012
--- pg_bsd_indent-1.3/args.c    Thu Jan 30 22:09:31 2014
***************
*** 83,89 ****
  #include <string.h>
  #include "indent_globs.h"

! #define INDENT_PG_VERSION    "1.2"

  /* profile types */
  #define    PRO_SPECIAL    1    /* special case */
--- 83,89 ----
  #include <string.h>
  #include "indent_globs.h"

! #define INDENT_PG_VERSION    "1.3"

  /* profile types */
  #define    PRO_SPECIAL    1    /* special case */
Only in pg_bsd_indent-1.2: indent.bsd.patch
diff -cr pg_bsd_indent-1.2/indent.c pg_bsd_indent-1.3/indent.c
*** pg_bsd_indent-1.2/indent.c    Mon Nov 14 19:30:01 2005
--- pg_bsd_indent-1.3/indent.c    Thu Jan 30 22:06:43 2014
***************
*** 550,555 ****
--- 550,562 ----
          case newline:
              if (ps.last_token != comma || ps.p_l_follow > 0
                  || !ps.leave_comma || ps.block_init || !break_comma || s_com != e_com) {
+         /* don't newline on cpp lines like #endif */
+         if (buf_ptr[0] == '#')
+         {
+             ps.just_saw_decl = 0;
+             prefix_blankline_requested = 0;
+             postfix_blankline_requested = 0;
+         }
                  dump_line();
                  ps.want_blank = false;
              }
Only in pg_bsd_indent-1.3: netbsd.patch

Thanks, Tom. You're right, my mistake was assuming that 9.3.3 was
pgindent-clean. Now I understand that that's not necessarily the case!

And thanks for the pg_bsd_indent patch. There's not much of a difference
between the two so I won't worry.

On Mon, Jun 20, 2016 at 3:50 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:

> sthomas@datawareventures.com writes:
> > I hacked pgindent to accept 1.3. This time it ran, but changed lots and
> lots
> > of files. (I assume because pg_bsd_indent 1.3 makes different choices
> than
> > pg_bsd_indent 1.2.)
> > Does anyone know where I can find pg_bsd_indent 1.2?
>
> I'm not sure why we're not keeping the older tarballs online, but there's
> not that much difference between 1.2 and 1.3 --- see attached diff.
>
> I think your problem may be in expecting that 9.3.x is pgindent-clean.
> We don't normally re-pgindent stable branches, so unless committers are
> careful to manually indent whatever they back-patch, branches will drift
> away from a clean state.  If you're trying to validate that you have a
> non-broken build of pgindent, I'd suggest checking out the immediately
> post-pgindent state from our git repo.  The 9.3 run was at commit
> 9af4159fc.
>
>                         regards, tom lane
>
>


--
Stephen Thomas
COO, Dataware Ventures
http://www.datawareventures.com

On Mon, Jun 20, 2016 at 03:50:27PM -0400, Tom Lane wrote:
> sthomas@datawareventures.com writes:
> > I hacked pgindent to accept 1.3. This time it ran, but changed lots and lots
> > of files. (I assume because pg_bsd_indent 1.3 makes different choices than
> > pg_bsd_indent 1.2.)
> > Does anyone know where I can find pg_bsd_indent 1.2?
>
> I'm not sure why we're not keeping the older tarballs online, but there's
> not that much difference between 1.2 and 1.3 --- see attached diff.

I can put the 1.2 version online, but I figured only HEAD would use
pgindent, and I wasn't sure how to specify which major versions require
which pg_bsd_indent version.

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

+ As you are, so once was I. As I am, so you will be. +
+                     Ancient Roman grave inscription +