The way I understand it, someone has been maintaining a 2PC patch (perhaps even since the 7.3.x days?) However, there seems to be little interest in the patch. It didn't get reviewed for long. And finally, when it's about to be integrated in the later phase of going to 8.0, the nested transaction patch got in. This conflicts heavily with 2PC. The 2PC patch maintainer didn't have time to adjust the patch, and so 2PC didn't make it to 8.0. Instead, nested transaction does.
2PC will probably be integrated in 8.1?
So if you really need 2PC, join the pgsql-hackers mailing list and test/comment on the 2PC patch. More requested features will get in first. That's the way it works.
What BYTEA encoding issues? You store *raw bytes* in BYTEA. As for quoting, you can use '\\000' to '\\377' to represent byte 0 - 255. You can store *anything* in BYTEA. So again, what encoding issues?
The features are getting in there... You make me laugh. Where are the *basic* things like triggers, stored procedures (only alpha), check constraints, etc? People have been waiting for *years*. MySQL development is getting slower and slower...
Problem with PG is that it is an old design that has been worked with for quite a while. Dude, MySQL design is even *older* (ISAM, for one). Postgres has relatively newer features like object relational.
Postgres will probably support it in the long run (2-3 years?) after they rework their internationalization support. This will allow you to define a collation ordering for your columns.
But on the other hand, perhaps you should not rely on case insensitivity, as it will limit you to only MySQL and SQL Server. Many other DB's act case sensitively (Oracle, PostgreSQL, DB2,...)
Debian (stable) is geared towards server, the whole Gentoo thing is geared towards desktop or experimental.
Debian has always had the philosophy of free distribution and legal safety, I've seen none of this in Gentoo. I love the Debian philosophy.
Debian is a mature distribution with a strict QA, I still don't believe Gentoo has a decent QA "department" at all.
None of the datacenters/dedicated servers facility that I know offer Gentoo, for each one you mention supporting Gentoo, I can name 25 supporting Debian.
Gentoo has bleeding edge stuffs, that's why I don't want it.
Debian has complete support forum (debianplanet), a portal (debianplanet), ten times the number of mailing lists than Gentoo, local user groups, not to mention SEVERAL newsletters with real content.
Debian has Knoppix, etc based on it.
Debian has at least twice the number of worldwide mirrors compared to Gentoo;
As for "versions", you can upgrade from between Debian versions pretty much seamlessly.
Flexible, Coherent backups: check. Postgres 8.0 does this via PITR.
Runs on VMS. Generally irrelevant considering VMS is no longer manufactured. Any other platforms Ingres runs that Postgres doesn't? On the other hand, there are people that already port Postgres to PDA/Zaurus/etc.
I haven't done any reading regarding Ingres' other features, but it will be interesting to see whether Ingres has [the alternative of] stuffs like BYTEA/TEXT ("inline" blobs), PL's in many languages (Perl, C, Ruby, Python, Tcl, Java, Mono C#, PHP, PL/PGSQL), MVCC, partial index (index on only some rows of a table), regex, nested transaction/savepoint, full text search, object relational features (like table inheritance), and a bunch of convenient data types like arrays, geometry types, IPv4/IPv6, arbitrary precision numbers, etc.
Postgres... It has a reputation for slowness; but, then again, so does Ingres.
Postgres *had* the reputation for slowness. That was back in the days of 6.4-6.5. Nowadays it's generally quite fast. You can even compare it head to head with MySQL in many cases. But when it comes to features, Postgres just blows MySQL away...
Couldn't agree more. There are just some classes of programs that Java won't be able to do. CGI scripts for example. Or programs that have to be called several times in a second (like using it with tcpserver).
I will still wait for a year or two though. The PHP team is notorious for breaking compatibility even between minor versions. I had several days of headaches a few months ago dealing with 4.2-4.3 migration. Defaults are changed/switched, path info not working, other changes of behaviours, some warnings becoming fatal error, etc etc. Don't even want to think about 4.3-5.0 for now...
Come on, over 100 comments and nobody said regexp operator?
I can't stand re.replace('...'), re.match('...'), etc. That's why I like Perl and Ruby, which provides with re operators such as//, m(), qr()/%r(), etc.
I have been using BestCrypt for several years and it's great. Unfortunately it's the only solid product available on Linux and Windows that I know of.
Why can't you store long filenames on Linux? BestCrypt just provides a block device on which you can use any filesystem. FAT32/VFAT is fine for Linux-Windows work. I've formatted a Bestcrypt volume with fat, vfat, iso9660, ext2, ext3, reiserfs, all without any problem.
Another alternative is to use VMware and then use ext3 on Bestcrypt and serve Windows needs with Samba.
Win2k's is 128GB and I was bitten by this once. I bought a 160GB drive, created one big partition with Redhat 7.3, and formatted it as NTFS under Win2k. Win2k displays it as 160GB but actually when the drive is near full, old data was overwritten by the new one!
Is Win2k's limitation artificial? I'd hate that.
Well, anyway, I've said goodbye to Windows as my desktop.
Before you cough on Ruby and said "2" + 2 in Ruby will equal 4, have you even tried it? Ruby (and Python) are _strongly-typed_ languages, unlike Perl (and PHP) which are weakly-typed. Ruby protects you from mixing string and numbers in operations like +. You'll have to say "2".to_i + 2.
And having used all four of those for projects large and small, I can say with confidence that I prefer strongly typed language. Weakly typed language is more dangerous and error-prone.
They should've patented the virus warning message instead. You know, when a virus/worm sends the server a copy of itself, and antivirus on the server returns a bounce saying the original message contains virus and has been removed/quarantined/whatever by Acme(TM) VirusNoMore 1.23.
I'm so pissed off by these messages. The antivirus maker _knows_ the Sender is faked. But they send it anyway. They're basically spammers!
If this is patented, then no other companies can use the same process, which is fine by me.
The article says that Intel's HT doesn't improve performance much. Isn't this expected, considering that IIRC FreeBSD's kernel threads still suck and most of the programs are single threaded anyway?
Of course, the answer lies in GUID, which is guaranteed to be unique even across the universe. So no more painful name switching of Firebird to Firefox to Firephoebe, just use
Basically, the upgrade inertia is largely due to modules. For me here's the list of modules that are currently 1.3.x only:
* ChiliASP/mod_casp. I don't know whether they have supported Apache2 now, but frankly I intend to get rid of it of mod_mono (which already supports Apache2). I truly hate this piece of crap!
* mod_frontpage. Also haven't checked out lately.
* a couple of C modules I wrote. I really hate C though, and I intend to rewrite these in Ruby/mod_ruby.
Slashdot Mirror
The way I understand it, someone has been maintaining a 2PC patch (perhaps even since the 7.3.x days?) However, there seems to be little interest in the patch. It didn't get reviewed for long. And finally, when it's about to be integrated in the later phase of going to 8.0, the nested transaction patch got in. This conflicts heavily with 2PC. The 2PC patch maintainer didn't have time to adjust the patch, and so 2PC didn't make it to 8.0. Instead, nested transaction does.
2PC will probably be integrated in 8.1?
So if you really need 2PC, join the pgsql-hackers mailing list and test/comment on the 2PC patch. More requested features will get in first. That's the way it works.
What BYTEA encoding issues? You store *raw bytes* in BYTEA. As for quoting, you can use '\\000' to '\\377' to represent byte 0 - 255. You can store *anything* in BYTEA. So again, what encoding issues?
There's ltree in contrib/. It's pretty much the same thing, but with different syntax.
The features are getting in there... You make me laugh. Where are the *basic* things like triggers, stored procedures (only alpha), check constraints, etc? People have been waiting for *years*. MySQL development is getting slower and slower...
Problem with PG is that it is an old design that has been worked with for quite a while. Dude, MySQL design is even *older* (ISAM, for one). Postgres has relatively newer features like object relational.
Not only you can store binary data easily in Postgres, you can even *index* it (even if it's thousands of bytes long).
Can you do that with MySQL???
Postgres will probably support it in the long run (2-3 years?) after they rework their internationalization support. This will allow you to define a collation ordering for your columns. But on the other hand, perhaps you should not rely on case insensitivity, as it will limit you to only MySQL and SQL Server. Many other DB's act case sensitively (Oracle, PostgreSQL, DB2, ...)
Yes, yes, parent is a troll. But...
Debian (stable) is geared towards server, the whole Gentoo thing is geared towards desktop or experimental.
Debian has always had the philosophy of free distribution and legal safety, I've seen none of this in Gentoo. I love the Debian philosophy.
Debian is a mature distribution with a strict QA, I still don't believe Gentoo has a decent QA "department" at all.
None of the datacenters/dedicated servers facility that I know offer Gentoo, for each one you mention supporting Gentoo, I can name 25 supporting Debian.
Gentoo has bleeding edge stuffs, that's why I don't want it.
Debian has complete support forum (debianplanet), a portal (debianplanet), ten times the number of mailing lists than Gentoo, local user groups, not to mention SEVERAL newsletters with real content.
Debian has Knoppix, etc based on it.
Debian has at least twice the number of worldwide mirrors compared to Gentoo;
As for "versions", you can upgrade from between Debian versions pretty much seamlessly.
I thought Woody still uses 2.2 kernel? Doesn't a jump to another kernel series merit a major version upgrade?
PITR: check. Postgres 8.0 now has it.
Tablespaces: check. Postgres 8.0 now has it.
Flexible, Coherent backups: check. Postgres 8.0 does this via PITR.
Runs on VMS. Generally irrelevant considering VMS is no longer manufactured. Any other platforms Ingres runs that Postgres doesn't? On the other hand, there are people that already port Postgres to PDA/Zaurus/etc.
I haven't done any reading regarding Ingres' other features, but it will be interesting to see whether Ingres has [the alternative of] stuffs like BYTEA/TEXT ("inline" blobs), PL's in many languages (Perl, C, Ruby, Python, Tcl, Java, Mono C#, PHP, PL/PGSQL), MVCC, partial index (index on only some rows of a table), regex, nested transaction/savepoint, full text search, object relational features (like table inheritance), and a bunch of convenient data types like arrays, geometry types, IPv4/IPv6, arbitrary precision numbers, etc.
Postgres... It has a reputation for slowness; but, then again, so does Ingres.
Postgres *had* the reputation for slowness. That was back in the days of 6.4-6.5. Nowadays it's generally quite fast. You can even compare it head to head with MySQL in many cases. But when it comes to features, Postgres just blows MySQL away...
Couldn't agree more. There are just some classes of programs that Java won't be able to do. CGI scripts for example. Or programs that have to be called several times in a second (like using it with tcpserver).
...though only for tests that Parrot can implement right now. See http://lambda-the-ultimate.org/node/view/141
Is there a similar utility for Linux?
I will still wait for a year or two though. The PHP team is notorious for breaking compatibility even between minor versions. I had several days of headaches a few months ago dealing with 4.2-4.3 migration. Defaults are changed/switched, path info not working, other changes of behaviours, some warnings becoming fatal error, etc etc. Don't even want to think about 4.3-5.0 for now...
Come on, over 100 comments and nobody said regexp operator?
//, m(), qr()/%r(), etc.
I can't stand re.replace('...'), re.match('...'), etc. That's why I like Perl and Ruby, which provides with re operators such as
Python's raw string still sucks for regex.
I have been using BestCrypt for several years and it's great. Unfortunately it's the only solid product available on Linux and Windows that I know of.
Why can't you store long filenames on Linux? BestCrypt just provides a block device on which you can use any filesystem. FAT32/VFAT is fine for Linux-Windows work. I've formatted a Bestcrypt volume with fat, vfat, iso9660, ext2, ext3, reiserfs, all without any problem.
Another alternative is to use VMware and then use ext3 on Bestcrypt and serve Windows needs with Samba.
I have flashed my BIOS with the latest and the BIOS has shown the right size (160G).
Besides, Microsoft is known to put artificial limitation, for example 32GB for FAT32.
Win2k's is 128GB and I was bitten by this once. I bought a 160GB drive, created one big partition with Redhat 7.3, and formatted it as NTFS under Win2k. Win2k displays it as 160GB but actually when the drive is near full, old data was overwritten by the new one!
Is Win2k's limitation artificial? I'd hate that.
Well, anyway, I've said goodbye to Windows as my desktop.
Before you cough on Ruby and said "2" + 2 in Ruby will equal 4, have you even tried it? Ruby (and Python) are _strongly-typed_ languages, unlike Perl (and PHP) which are weakly-typed. Ruby protects you from mixing string and numbers in operations like +. You'll have to say "2".to_i + 2.
And having used all four of those for projects large and small, I can say with confidence that I prefer strongly typed language. Weakly typed language is more dangerous and error-prone.
They should've patented the virus warning message instead. You know, when a virus/worm sends the server a copy of itself, and antivirus on the server returns a bounce saying the original message contains virus and has been removed/quarantined/whatever by Acme(TM) VirusNoMore 1.23.
I'm so pissed off by these messages. The antivirus maker _knows_ the Sender is faked. But they send it anyway. They're basically spammers!
If this is patented, then no other companies can use the same process, which is fine by me.
The article says that Intel's HT doesn't improve performance much. Isn't this expected, considering that IIRC FreeBSD's kernel threads still suck and most of the programs are single threaded anyway?
Of course, the answer lies in GUID, which is guaranteed to be unique even across the universe. So no more painful name switching of Firebird to Firefox to Firephoebe, just use
Mozilla 43c23aa3-7c29-4ce2-96ca-23c751efe5ff 0.9
Basically, the upgrade inertia is largely due to modules. For me here's the list of modules that are currently 1.3.x only: * ChiliASP/mod_casp. I don't know whether they have supported Apache2 now, but frankly I intend to get rid of it of mod_mono (which already supports Apache2). I truly hate this piece of crap! * mod_frontpage. Also haven't checked out lately. * a couple of C modules I wrote. I really hate C though, and I intend to rewrite these in Ruby/mod_ruby.
Hm, doesn't sound as good, does it?
Actually, the report is wrong. They just don't read the resumes carefully. The other 68% actually know *Javanese* Java, not Sun Java.