[HACKERS] UnixWare 7.1.1b FS
I got early access to the UnixWare 7.1.1b Feature Supplement UnixWare/OpenServer Development Kit (UDK FS for short). PostgreSQL 7.0.2 compiles CLEAN (ish...) and we can now support the C++ stuff. The 3.2 bug in int8.c is GONE. Question: Why do we (for UnixWare) force i486 optimization? Attached is the configure output, gmake output for analysis by y'all... The configure invocation was: CXXFLAGS=-O ./configure --with-perl --with-CC=cc --with-CXX=CC --with-includes=/usr/local/include --with-libs=/usr/local/lib >ler.conf.out 2>&1 & -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 972-414-9812 (voice) Internet: [EMAIL PROTECTED] US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749 creating cache ./config.cache checking host system type... i586-sco-sysv5uw7.1.1 checking echo setting... checking setting template to... unixware checking whether to support locale... disabled checking whether to support cyrillic recode... disabled checking whether to support multibyte... disabled checking setting DEF_PGPORT... 5432 checking setting DEF_MAXBACKENDS... 32 checking setting USE_TCL... disabled checking setting USE_PERL... enabled checking setting USE_ODBC... disabled checking setproctitle... disabled checking setting ASSERT CHECKING... disabled checking for gcc... cc checking whether the C compiler (cc -O -K i486,host,inline,loop_unroll,alloca -Dsvr4 ) works... yes checking whether the C compiler (cc -O -K i486,host,inline,loop_unroll,alloca -Dsvr4 ) is a cross-compiler... no checking whether we are using GNU C... no checking whether cc accepts -g... yes checking how to run the C preprocessor... cc -E - setting CPPFLAGS= -I/usr/local/include -I/opt/include - setting LDFLAGS= -L/usr/local/lib -L/opt/lib checking setting debug compiler flag... using default checking for c++... CC checking whether the C++ compiler (CC -O -L/usr/local/lib -L/opt/lib) works... yes checking whether the C++ compiler (CC -O -L/usr/local/lib -L/opt/lib) is a cross-compiler... no checking whether we are using GNU C++... no checking whether CC accepts -g... yes checking for include in C++... yes checking for namespace std in C++... yes checking for a BSD compatible install... /bin/install -c checking for flex... flex checking for yywrap in -lfl... yes checking whether ln -s works... yes checking whether make sets ${MAKE}... yes checking for ranlib... : checking for find... /bin/find checking for tar... /bin/tar checking for split... /bin/split checking for etags... /usr/local/bin/etags checking for xargs... /bin/xargs checking for gzcat... /usr/local/bin/gzcat checking for perl... perl checking for bison... /usr/local/bin/bison - Using /usr/local/bin/bison -y -d checking for main in -lsfio... no checking for main in -lncurses... no checking for main in -lcurses... yes checking for main in -ltermcap... yes checking for main in -lreadline... yes checking for using_history in -lreadline... yes checking for main in -lbsd... no checking for main in -lutil... yes checking for main in -lm... yes checking for main in -ldl... yes checking for main in -lsocket... yes checking for main in -lnsl... yes checking for main in -lipc... no checking for main in -lIPC... no checking for main in -llc... no checking for main in -ldld... no checking for main in -lln... no checking for main in -lld... yes checking for main in -lcompat... no checking for main in -lBSD... no checking for main in -lcrypt... yes checking for main in -lgen... yes checking for main in -lPW... no checking for ANSI C header files... yes checking for sys/wait.h that is POSIX.1 compatible... yes checking for arpa/inet.h... yes checking for crypt.h... yes checking for dld.h... no checking for endian.h... no checking for float.h... yes checking for fp_class.h... no checking for getopt.h... no checking for history.h... yes checking for ieeefp.h... yes checking for limits.h... yes checking for netdb.h... yes checking for netinet/in.h... yes checking for readline.h... yes checking for readline/history.h... yes checking for readline/readline.h... yes checking for sys/select.h... yes checking for termios.h... yes checking for unistd.h... yes checking for values.h... yes checking for sys/param.h... yes checking for pwd.h... yes checking for working const... yes checking for inline... inline checking for preprocessor stringizing operator... yes checking for uid_t in sys/types.h... yes checking for mode_t... yes checking for off_t... yes checking for size_t... yes checking whether time.h and sys/time.h may both be included... yes checking whether struct tm is in sys/time.h or time.h... time.h checking for tm_zone in struct tm... no checking for tzname... yes checking for signed types... yes checking for volatile... yes checking for type of last arg to accept... size_t checking for int timezone... yes checking for gettimeofday args... 2 args checking for union semun... no checking for fcntl(F_SETLK)... yes checking for 8-bit clean memcmp... y
[HACKERS] Re: [COMMITTERS] pgsql/doc (FAQ_MSWIN INSTALL_MSWIN)
Peter Eisentraut wrote: > OTOH, support for Windows NT with Cygwin is not "exceptional" compared to > any other supported system. You run configure; make; make install, with > GNU make and GCC. The only thing this INSTALL_WIN file does is to tell > people where to download Cygwin and cgyipc, which is no more different > than FAQ_Solaris telling people to download GNU packages from > www.sunfreeware.com or FAQ_HPUX advising to download various vendor > patches. OK. I *should* have looked on my development box before spouting off mail, eh? ;) Sorry for the noise... - Thomas
Re: [HACKERS] New build fails: cannot find postmaster.opts.default
Kevin O'Gorman writes: > I'm getting a complaint from pg_ctl that it cannot find > postmaster.opts.default. > > What am I missing? The postmaster.opts.default file. :-) Seriously, I'm thinking this check is not necessary, a missing file should be treated like an empty file. Objections? -- Peter Eisentraut [EMAIL PROTECTED] http://yi.org/peter-e/
Re: [HACKERS] New build fails: cannot find postmaster.opts.default
On Fri, Oct 20, 2000 at 04:15:11PM -0700, Kevin O'Gorman wrote: > I've just built 7.1 from a slightly old point in the tree: > October 9. Regression tests pass, but postmaster won't > start. > > I'm getting a complaint from pg_ctl that it cannot find > postmaster.opts.default. > > What am I missing? > touch postmaster.opts.default in the directory pg_ctl complains about. An empty file works fine. Ross -- Open source code is like a natural resource, it's the result of providing food and sunshine to programmers, and then staying out of their way. [...] [It] is not going away because it has utility for both the developers and users independent of economic motivations. Jim Flynn, Sunnyvale, Calif.
[HACKERS] New build fails: cannot find postmaster.opts.default
I've just built 7.1 from a slightly old point in the tree: October 9. Regression tests pass, but postmaster won't start. I'm getting a complaint from pg_ctl that it cannot find postmaster.opts.default. What am I missing? ++ kevin -- Kevin O'Gorman (805) 650-6274 mailto:[EMAIL PROTECTED] Permanent e-mail forwarder: mailto:Kevin.O'[EMAIL PROTECTED] At school: mailto:[EMAIL PROTECTED] Web: http://www.cs.ucsb.edu/~kogorman/index.html Web: http://trixie.kosman.via.ayuda.com/~kevin/index.html "There is a freedom lying beyond circumstance, derived from the direct intuition that life can be grounded upon its absorption in what is changeless amid change" -- Alfred North Whitehead
[HACKERS] Navigating time-warps in the CVS tree (was re the rule system)
Tom Lane wrote: > I'm not sure whether to recommend that you work from current CVS sources > or not. A couple weeks ago that's what I would have said, but Vadim is > halfway through integrating WAL changes and I'm not sure how stable the > tip really is. You could try the tip, and if it blows up fall back to > a dated retrieval from about 7-Oct. Or you could investigate the way > that the 7.0.* rewriter handles the rtable list for multiple queries, > but that's probably not a real profitable use of your time. > > regards, tom lane Well, I tried the tip of the tree today, and initdb fails to complete, so I tried going back to '7 Oct 2000 10:00:00 PST' and it's better, but regression tests fail on the rule system. It makes the server die. Since rules are what I want, this won't do. I'm not familiar enough with CVS or your changelog system well enough to know a good way to find a time-point that might be stable enough for me. How would I find out where I need to be?? ++ kevin -- Kevin O'Gorman (805) 650-6274 mailto:[EMAIL PROTECTED] Permanent e-mail forwarder: mailto:Kevin.O'[EMAIL PROTECTED] At school: mailto:[EMAIL PROTECTED] Web: http://www.cs.ucsb.edu/~kogorman/index.html Web: http://trixie.kosman.via.ayuda.com/~kevin/index.html "There is a freedom lying beyond circumstance, derived from the direct intuition that life can be grounded upon its absorption in what is changeless amid change" -- Alfred North Whitehead
Re: make depend (Re: [HACKERS] Coming attractions: VPATH build; makevariables issue)
Tom Lane writes: > Do you have an idea how much it'd bloat the tarball to do that? current depsincrease gzipped 7.1 MB 35591 bytes 0.5% unpacked29MB1309669 bytes 4% Should be okay... -- Peter Eisentraut [EMAIL PROTECTED] http://yi.org/peter-e/
Re: [HACKERS] Now 376175 lines of code
Ross J. Reedstrom writes: > For example, no one seems to have commented on the -8% of inline > comments reported by Peter's c_count! Funny math, indeed. If you had actually done the math ;-) you would have noticed that the percentage of the inline comments is negative because those lines have both comments and code, therefore the total has to exclude these lines once when adding comments and code. -- Peter Eisentraut [EMAIL PROTECTED] http://yi.org/peter-e/
Re: [HACKERS] what is CVS?
Robert Kernell <[EMAIL PROTECTED]> writes: > > Kevin O'Gorman wrote: > > > > > > I've been unable to follow the directions > > > in the Programmer's Guide > > > for getting to the anonymous CVS server. > > > > > > I'm running RedHat 6.1, and CVS 1.10 which > > What is CVS? An open-source network-transparent version control system. http://www.cvshome.org/ -- Trond Eivind Glomsrød Red Hat, Inc.
Re: [HACKERS] is there a way to DROP foreign key constraint ?
On Fri, 20 Oct 2000, Hannu Krosing wrote: > I'm unable to find the complementary function to > > ALTER TABLE t ADD FOREIGN KEY(id) REFERENCES pkt(pk); Currently, ALTER TABLE ... DROP table constraint definition doesn't exist. > I would try DROP TRIGGER, but I've also been unable to > find a way to name the constraint ;( Umm, put the constraint name in the table constraint definition? >From the SQL spec: ::= [ ] [ ] ::= CONSTRAINT Currently, you need to drop the three triggers from pg_trigger that are associated with the constraint. Easiest way to find them if you don't have a constraint name is to look at the tg_args.
Re: [HACKERS] Now 376175 lines of code
On Fri, Oct 20, 2000 at 01:30:25PM -0400, Bruce Momjian wrote: > I compute the code count with: > > find . -name \*.[chyl] | xargs cat| wc -l Right, that solves the problem others might be seeing, with the command line getting expanded and silently chopped off. For example, no one seems to have commented on the -8% of inline comments reported by Peter's c_count! Funny math, indeed. Ross -- Open source code is like a natural resource, it's the result of providing food and sunshine to programmers, and then staying out of their way. [...] [It] is not going away because it has utility for both the developers and users independent of economic motivations. Jim Flynn, Sunnyvale, Calif.
Re: [HACKERS] Now 376175 lines of code
On Fri, 20 Oct 2000, Hannu Krosing wrote: > Peter Eisentraut wrote: > > > > Bruce Momjian writes: > > > > > FYI, it is 376k lines of C code, not bytes. > > > > How did you calculate that? I get this using c_count over all .c and .h > > files: > > > > 20903 lines had comments25.4 % > > 6603 comments are inline -8.0 % > > 11911 lines were blank 14.5 % > > 7287 lines for preprocessor 8.9 % > > 48716 lines containing code 59.3 % > > 82214 total lines 100.0 % > > > > Surely we don't have 294000 lines of Java, C++, Shell, and Perl??? > > doing the following in version 6.5.3 in src/backend > > [hannu@hu backend]$ cat */*.[ch] */*/*.[ch] */*/*/*.[ch]| wc > > gives > > 208284 658632 5249304 > > So you (or c_count ;) must be missing some files > > in src/ ther result was > [hannu@hu src]$ cat */*.[ch] */*/*.[ch] */*/*/*.[ch] */*/*/*/*.[ch]| wc > 311469 1069935 8440682 Just now downloaded from ftp.postgresql.org: $ tar -zxvf postgresql-6.5.3.tar.gz $ cd postgresql-6.5.3 $ wc `find -name "*.[ch]"` 318131 1089740 8585092 total $ wc `find -name "*"` 756810 3037982 25583644 total $ cd src $ wc `find -name "*.[ch]"` 311469 1069935 8440682 total $ wc `find -name "*"` 519318 2024262 16656475 total $ tar -zxvf postgresql-7.0.2.tar.gz $ cd postgresql-7.0.2 $ wc `find -name "*.[ch]"` 368502 126 9910813 total $ wc `find -name "*"` 756810 3037982 25583644 total $ cd src $ wc `find -name "*.[ch]"` 361297 1240788 9751161 total $ wc `find -name "*"` 596772 2360555 18574015 total Karel
Re: [HACKERS] Now 376175 lines of code
I compute the code count with: find . -name \*.[chyl] | xargs cat| wc -l -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup.| Drexel Hill, Pennsylvania 19026
[HACKERS] Re: [COMMITTERS] pgsql/doc (FAQ_MSWIN INSTALL_MSWIN)
Thomas Lockhart writes: > Hmm. It is a fact that M$Windows is not quite the same as *any* of the > other boxes we support. > We currently have a section in the printed/html docs covering installing > on M$ for client-side libraries, and we should move the installation > instructions for the server into these docs also. Note that there's a difference here: Support for a plain-old Windows 95 box with MS Visual C++ compiler is certainly very different. There are even separate makefiles (project files?) for that. OTOH, support for Windows NT with Cygwin is not "exceptional" compared to any other supported system. You run configure; make; make install, with GNU make and GCC. The only thing this INSTALL_WIN file does is to tell people where to download Cygwin and cgyipc, which is no more different than FAQ_Solaris telling people to download GNU packages from www.sunfreeware.com or FAQ_HPUX advising to download various vendor patches. -- Peter Eisentraut [EMAIL PROTECTED] http://yi.org/peter-e/
Re: [HACKERS] Now 376175 lines of code
Peter Eisentraut wrote: > > Bruce Momjian writes: > > > FYI, it is 376k lines of C code, not bytes. > > How did you calculate that? I get this using c_count over all .c and .h > files: > > 20903 lines had comments25.4 % > 6603 comments are inline -8.0 % > 11911 lines were blank 14.5 % > 7287 lines for preprocessor 8.9 % > 48716 lines containing code 59.3 % > 82214 total lines 100.0 % > > Surely we don't have 294000 lines of Java, C++, Shell, and Perl??? doing the following in version 6.5.3 in src/backend [hannu@hu backend]$ cat */*.[ch] */*/*.[ch] */*/*/*.[ch]| wc gives 208284 658632 5249304 So you (or c_count ;) must be missing some files in src/ ther result was [hannu@hu src]$ cat */*.[ch] */*/*.[ch] */*/*/*.[ch] */*/*/*/*.[ch]| wc 311469 1069935 8440682 - Hannu
Re: [HACKERS] Now 376175 lines of code
> Bruce Momjian writes: > > > FYI, it is 376k lines of C code, not bytes. > > How did you calculate that? I get this using c_count over all .c and .h > files: > > 20903 lines had comments25.4 % > 6603 comments are inline -8.0 % > 11911 lines were blank 14.5 % > 7287 lines for preprocessor 8.9 % > 48716 lines containing code 59.3 % > 82214 total lines 100.0 % > > Surely we don't have 294000 lines of Java, C++, Shell, and Perl??? I just counted lines, not line content. Not sure which is more meaningful. Our comments are as important as the code, sometimes, though they do not add functionality to the application. I am not inclined to inflate numbers, but I am not sure the 59% number is accurate either. Opinions? -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 853-3000 + If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup.| Drexel Hill, Pennsylvania 19026
[HACKERS] what is CVS?
> Kevin O'Gorman wrote: > > > > I've been unable to follow the directions > > in the Programmer's Guide > > for getting to the anonymous CVS server. > > > > I'm running RedHat 6.1, and CVS 1.10 which What is CVS? Bob Kernell Research Scientist Surface Validation Group AS&M, Inc. email: [EMAIL PROTECTED] tel: 757-827-4631
Re: [HACKERS] Now 376175 lines of code
Bruce Momjian writes: > FYI, it is 376k lines of C code, not bytes. How did you calculate that? I get this using c_count over all .c and .h files: 20903 lines had comments25.4 % 6603 comments are inline -8.0 % 11911 lines were blank 14.5 % 7287 lines for preprocessor 8.9 % 48716 lines containing code 59.3 % 82214 total lines 100.0 % Surely we don't have 294000 lines of Java, C++, Shell, and Perl??? -- Peter Eisentraut [EMAIL PROTECTED] http://yi.org/peter-e/
Re: [HACKERS] Unable to access CVS server
Kevin O'Gorman wrote: > > I've been unable to follow the directions > in the Programmer's Guide > for getting to the anonymous CVS server. > > I'm running RedHat 6.1, and CVS 1.10 which > comes with it. I get as far as entering > the 'postgresql' password, but it gets > rejected every time. > > Any hints? I ran into the same problem a while ago. Both username and password in docs are WRONG unfortunately I'm away from my regular computer and so I can't look up the right ones .. - Hannu
Re: [HACKERS] to_char() dumps core
On Fri, 20 Oct 2000, Tatsuo Ishii wrote: > In 7.0.2 > >select to_char(sum(n),'999') from t1; > > causes backend dump a core if n is a float/numeric ...data type AND if > sum(n) returns NULL. This seems due to a bad null pointer handling for > aruguments of pass-by-reference data types. I think just a simple > null pointer checking at very top of each function (for example > float4_to_char()) would solve the problem. Comments? In the 7.1devel it's correct, but here it's bug, IMHO it bear on changes in the 7.1's fmgr, because code is same in both versions for this. On Monday, I try fix it for 7.0.3 Karel > test=# create table t1(f float); > CREATE > test=# select to_char(sum(f),'999') from t1; > pqReadData() -- backend closed the channel unexpectedly. > This probably means the backend terminated abnormally > before or while processing the request. > The connection to the server was lost. Attempting reset: Failed. >
Re: [HACKERS] Re: pg_dump docs
On Fri, 20 Oct 2000, Thomas Lockhart wrote: > Ah, the build has been failing for at least the last few days due to > small problems in new content. Since I receive ~700 logs of doc builds > each year (well, that is the annual rate but I've only stepped up to > twice daily since ~April), I get sloppy about looking through them > carefully, and instead tend to look for the *length* of the log as a > measure of success while rarely examining the end of the log to see how > it actually went. In this case I missed the failure. > > btw, the build log is updated and posted at > > http://www.postgresql.org/devel-corner/docs/docbuild.log > > Vince, could we get a cross reference to this on the developer's page? I was gonna ask you if you wanted that. Anyway it's now there. And before anyone says it, yes I know about the duplicate link. :) Vince. -- == Vince Vielhaber -- KA8CSHemail: [EMAIL PROTECTED]http://www.pop4.net 128K ISDN from $22.00/mo - 56K Dialup from $16.00/mo at Pop4 Networking Online Campground Directoryhttp://www.camping-usa.com Online Giftshop Superstorehttp://www.cloudninegifts.com ==
[HACKERS] is there a way to DROP foreign key constraint ?
I'm unable to find the complementary function to ALTER TABLE t ADD FOREIGN KEY(id) REFERENCES pkt(pk); I would try DROP TRIGGER, but I've also been unable to find a way to name the constraint ;( the built-in help is both inadequate and inconsistent 8<-8<-8<-8<-8<-8<- amphora2=# \h alter table Command: ALTER TABLE Description: Modifies table properties Syntax: ALTER TABLE table [ * ] ADD [ COLUMN ] column type ALTER TABLE table [ * ] ALTER [ COLUMN ] column { SET DEFAULT value | DROP DEFAULT } ALTER TABLE table [ * ] RENAME [ COLUMN ] column TO newcolumn ALTER TABLE table RENAME TO newtable ALTER TABLE table ADD table constraint definition amphora2=# \h create table Command: CREATE TABLE Description: Creates a new table Syntax: CREATE [ TEMPORARY | TEMP ] TABLE table ( column type [ NULL | NOT NULL ] [ UNIQUE ] [ DEFAULT value ] [column_constraint_clause | PRIMARY KEY } [ ... ] ] [, ... ] [, PRIMARY KEY ( column [, ...] ) ] [, CHECK ( condition ) ] [, table_constraint_clause ] ) [ INHERITS ( inherited_table [, ...] ) ] amphora2=# \h column_constraint_clause No help available for 'column_constraint_clause'. Try \h with no arguments to see available help. amphora2=# \h column constraint definition No help available for 'column constraint definition'. Try \h with no arguments to see available help. 8<-8<-8<-8<-8<-8<-
[HACKERS] Re: RE. COBOL FILES
Orlandi wrote: > > I am user of Microfocus Cobol Netexpress 3.1 and I use > IDXFORMAT " 4 " for my files. > I know that I can use the format Btrieve ANSI and to install Btrieve > in my Linux server with Format : IDXFORMAT " 6 " , and in the > case of Btrieve I can just recompile the programs, without > needing to alter the source code , I also know about that. > The one that I need to know is: > What should I do to use Postgre as manager of files in the > Linux server ? Bom dia Jairo. Unfortunately I have no experience with Microfocus Cobol or Btrieve. I'm copying this to the -hackers mailing list and perhaps someone will have some suggestions in this area. In general, Postgres is an RDBMS which is highly compatible with other RDBMS products. So if you have an RDBMS solution already, then the port should be pretty easy. If you don't already have an RDBMS in your system, then you will want to look into the general capabilities and interface options to make sure that you have a clear path for your application migration. Good luck! - Thomas
[HACKERS] to_char() dumps core
In 7.0.2 select to_char(sum(n),'999') from t1; causes backend dump a core if n is a float/numeric ...data type AND if sum(n) returns NULL. This seems due to a bad null pointer handling for aruguments of pass-by-reference data types. I think just a simple null pointer checking at very top of each function (for example float4_to_char()) would solve the problem. Comments? test=# create table t1(f float); CREATE test=# select to_char(sum(f),'999') from t1; pqReadData() -- backend closed the channel unexpectedly. This probably means the backend terminated abnormally before or while processing the request. The connection to the server was lost. Attempting reset: Failed.
Re: make depend (Re: [HACKERS] Coming attractions: VPATH build; make variables issue)
Peter Eisentraut <[EMAIL PROTECTED]> writes: > What we could do is ship the dependencies (.deps/*.P) in the tarball. > That would require running an entire build before making a tarball, but it > would be a nice service to users. Hm. It might be handy for people not using gcc, since they'd have no easy way to build dependencies for themselves. Do you have an idea how much it'd bloat the tarball to do that? Isn't the basic idea to write Makefile targets to remake dependency files when they are out of date with code? Won't those targets involve implicit rules for going, for example, from *.c -> *.d (or whatever convention you use for dependency files)? Don't these Makefiles also have a list of srcs to be built, e.g., a make variable that defines a list of *.c filename? If so, can't you just implement a depend: target as ${DEPEND_FILES}+=${SRCS:%c=%d} depend: ${DEPEND_FILES} .SUFFIXES: .c .d .c.d: gcc -M ... .include "${DEPEND_FILES}" For gmake users, all the magic happens automatically. Prior to distribution, just do make depend to get all the *.d files to include in the tarball. For non-gmake users, all the *.d files already exist in the source. If they make changes, they can run make depend manually. Sorry if this is what you had in mind already, but the discussion seemed to imply that you can't have it both ways. Cheers, Brook
Re: [HACKERS] Re: pg_dump docs
> [Not to list] Back on list; thanks though for protecting me from ridicule ;) > >For developers (-hackers developers, not application developers using > >the current release), the "current docs" correspond to the docs built > >nightly (actually twice a day), which reflect the current development > >tree. > If http://www.postgresql.org/devel-corner/docs/user/ is supposed to be > based on CVS, then I must have done something wrong; the new pg_restore > entry has not appeared (it's now > 24 hours old). > Any hints would be appreciated. Ah, the build has been failing for at least the last few days due to small problems in new content. Since I receive ~700 logs of doc builds each year (well, that is the annual rate but I've only stepped up to twice daily since ~April), I get sloppy about looking through them carefully, and instead tend to look for the *length* of the log as a measure of success while rarely examining the end of the log to see how it actually went. In this case I missed the failure. btw, the build log is updated and posted at http://www.postgresql.org/devel-corner/docs/docbuild.log Vince, could we get a cross reference to this on the developer's page? The current problem is in using underscores in the "id" field of header tags; this is an illegal character in this context per DocBook (a feature not at all obvious, but which can be seen by omission in our other docs). pg_restore.sgml also had quite a few ^M's at the end of lines; I've got a utility to clean those up so I applied those fixes also. While tracking down another problem with replicated ID fields in runtime.sgml, I caught a duplicated section and removed the apparently older version. I've also applied a couple of fixes suggested by Laser Henry. Things now build without errors on my local machine, and should do the same on postgresql.org. - Thomas
AW: AW: [HACKERS] Backup, restore & pg_dump
> > Yes, writes are only necessary when "too many dirty pages" > > are in the buffer pool. Those writes can be done by a page flusher > > on demand or during checkpoint (don't know if we need checkpoint, > > but you referred to doing checkpoints). > > How else to know from where in log to start redo and how far go back > for undo ? I don't know, but if your checkpoint algorithm does not need to block other activity, that would be great. The usual way would involve: writing all dirty pages to disk during checkpoint block all modifying activity One other thing I would like to ask, is O_SYNC not available on all platforms ? Then you could avoid the (or some) fsync calls in xlog.c ? And is there a possibility to add -F mode without fsyncs to xlog.c ? Andreas
Re: [HACKERS] INHERITS doesn't offer enough functionality
> The point is: this is classic, but noone does it > like this if your really have a larger hierarchy of > classes. You'll not get any good performance, when > solving an association in your oo > program, because the framework has to query against > each table: 6 tables - 6 queries !!! :-( > > With the PostgreSQL approach one can send ONE query > against the tables and one would get one result ... > which will be much faster (I hope so ... that has to > be prooved ..).= You'll still have to do 6 queries in postgres because it does not return fields in sub-classes. Imagine the root of the hierarchy is abstract with no fields. You query this class and you get 100 tuples with no columns! This is the aspect I'm hoping to fix but I'm waiting for Tom to re-do the query data structures before I do changes that are thrown away. > Actually one should think about: why do I really want to > have inheritance in the oo-rdbms ? Actually I could put > all columns (of all classes in this hierarchy into one table > and that's it). Ouch. That way lies madness.