Re: [sqlite] EXT :Re: Compiling SQLite3 with MSVC 2010
For gcc try -Wextra and -Wconversion. You'll get tons of warnings. -Wall just does the ones most people are concerned with. Michael D. Black Senior Scientist Advanced Analytics Directorate Advanced GEOINT Solutions Operating Unit Northrop Grumman Information Systems From: sqlite-users-boun...@sqlite.org [sqlite-users-boun...@sqlite.org] on behalf of Igor Korot [ikoro...@gmail.com] Sent: Sunday, November 04, 2012 2:12 PM To: General Discussion of SQLite Database Subject: Re: [sqlite] EXT :Re: Compiling SQLite3 with MSVC 2010 Michael, On Sun, Nov 4, 2012 at 5:59 AM, Black, Michael (IS) michael.bla...@ngc.com wrote: Hmmm...interesting...I'm using VS 2010 Express 32-bit and I would've assumed the warnings would match. I also enabled SQLITE_64BIT_STATS to try and force the first warning but that didn't cause it. Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 16.00.40219.01 for 80x86 Well, I'm getting it with Professional build of 32-bits. That's a bit disconcerting actually but I guess Express is less pedantic than Studio. You can always stick this in to shut it up. There's a yin yang to fixing thesesimple enough to throw a cast in there...but down the road if you make other changes on the right-hand-side datatype that could be of use so you would be suppressing a valid warning. It's a mixed bag. So suppressing warnings from picky compilers is the best way to go. Do you still see then on lower levels? Not with /W3. But it's interesting that -Wall does not produce them with gcc... Thank you. #pragma warning(disable: 4244) // possible loss of data Michael D. Black Senior Scientist Advanced Analytics Directorate Advanced GEOINT Solutions Operating Unit Northrop Grumman Information Systems From: sqlite-users-boun...@sqlite.org [sqlite-users-boun...@sqlite.org] on behalf of Jonas Malaco Filho [jonasmalacofi...@gmail.com] Sent: Saturday, November 03, 2012 3:48 PM To: General Discussion of SQLite Database Subject: EXT :Re: [sqlite] Compiling SQLite3 with MSVC 2010 Actually, on MSVC 2010 I just got the following errors with /W3: -- Rebuild All started: Project: Shell, Configuration: Release x64 -- shell.c sqlite3.c ..\src\sqlite3.c(78502): warning C4244: 'initializing' : conversion from 'sqlite_int64' to 'tRowcnt', possible loss of data ..\src\sqlite3.c(78503): warning C4244: 'initializing' : conversion from 'sqlite_int64' to 'tRowcnt', possible loss of data ..\src\sqlite3.c(78504): warning C4244: 'initializing' : conversion from 'sqlite_int64' to 'tRowcnt', possible loss of data ..\src\sqlite3.c(104145): warning C4244: '=' : conversion from 'i64' to 'double', possible loss of data ..\src\sqlite3.c(104170): warning C4244: '=' : conversion from 'i64' to 'double', possible loss of data Generating code Finished generating code Shell.vcxproj - X:\jonas-malaco-filho\lib\SQLite\Shell\..\bin\x86-64\sqlite3.exe -- Rebuild All started: Project: Shell, Configuration: Release Win32 -- shell.c sqlite3.c ..\src\sqlite3.c(78502): warning C4244: 'initializing' : conversion from 'sqlite_int64' to 'tRowcnt', possible loss of data ..\src\sqlite3.c(78503): warning C4244: 'initializing' : conversion from 'sqlite_int64' to 'tRowcnt', possible loss of data ..\src\sqlite3.c(78504): warning C4244: 'initializing' : conversion from 'sqlite_int64' to 'tRowcnt', possible loss of data ..\src\sqlite3.c(104145): warning C4244: '=' : conversion from 'i64' to 'double', possible loss of data ..\src\sqlite3.c(104170): warning C4244: '=' : conversion from 'i64' to 'double', possible loss of data Generating code Finished generating code Shell.vcxproj - X:\jonas-malaco-filho\lib\SQLite\Shell\..\bin\x86\sqlite3.exe == Rebuild All: 2 succeeded, 0 failed, 0 skipped == *Jonas Malaco Filho* 2012/11/3 Black, Michael (IS) michael.bla...@ngc.com You probably have the warning level turned up high. Up to level 3 it compiles without warnings. Level 4 starts complaining loudly. Michael D. Black Senior Scientist Advanced Analytics Directorate Advanced GEOINT Solutions Operating Unit Northrop Grumman Information Systems From: sqlite-users-boun...@sqlite.org [sqlite-users-boun...@sqlite.org] on behalf of Igor Korot [ikoro...@gmail.com] Sent: Friday, November 02, 2012 6:25 PM To: General Discussion of SQLite Database Subject: EXT :Re: [sqlite] Compiling SQLite3 with MSVC 2010 Richard, On Fri, Nov 2, 2012 at 4:14 PM, Richard Hipp d...@sqlite.org wrote: On Fri, Nov 2, 2012 at 7:05 PM, Igor Korot ikoro...@gmail.com wrote: Hi, ALL, Is anybody trying to compile SQLite with MSVC 2010? Tests 9e and 9f at http://www.sqlite.org/checklists/3071400#c9 were performed using MSVC 2010. I am getting a lot of warnings. Is there any interest in fixing
[sqlite] Required help on Sqlite+TCL -testsuit
Hi, I am New to sqlite and TCL. I need information regarding, 1. Where i can download TCL Testsuit to test sqlite ? I am using Cygwin environment , testsits are available for cygwin environment ? Does the same run on Linux too ? 2. How to install / compile and run these testsuits ? A reply is really appreciated... Thanks in advance Brijesh -- View this message in context: http://sqlite.1065341.n5.nabble.com/Required-help-on-Sqlite-TCL-testsuit-tp65374.html Sent from the SQLite mailing list archive at Nabble.com. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Required help on Sqlite+TCL -testsuit
I have the sqlite-amalgamation-3071401 and binaries for the TCL also extracted to the root directory respective folders... -- View this message in context: http://sqlite.1065341.n5.nabble.com/Required-help-on-Sqlite-TCL-testsuit-tp65374p65375.html Sent from the SQLite mailing list archive at Nabble.com. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Required help on Sqlite+TCL -testsuit
On Mon, Nov 5, 2012 at 5:38 AM, brijesh_philips brijeshk_...@yahoo.comwrote: Hi, I am New to sqlite and TCL. I need information regarding, 1. Where i can download TCL Testsuit to test sqlite ? I am using Cygwin environment , testsits are available for cygwin environment ? Does the same run on Linux too ? 2. How to install / compile and run these testsuits ? A reply is really appreciated... Download and install ActiveTcl. http://www.activestate.com/activetcl To compile and test using MSVC, first download and install the SQLite source tree, then type:: nmake /f makefile.msc test To compile and test using GCC, install mingw and msys (not cygwin) and then type: sh configure; make test Thanks in advance Brijesh -- View this message in context: http://sqlite.1065341.n5.nabble.com/Required-help-on-Sqlite-TCL-testsuit-tp65374.html Sent from the SQLite mailing list archive at Nabble.com. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
[sqlite] Anybody using REGEXP in the ICU extension?
Is anybody using the REGEXP operator that is included with the ICU extension of SQLite? The reason I ask is that the current ICU REGEXP implementation is broken. We are wanting to fix it. But the fix runs a serious risk of breaking any legacy applications that depend on the old broken behavior. The brokenness is this: The ICU REGEXP operator assumes that every pattern string begins with ^ and ends with $. In other words, the pattern must match the entire string. This is how LIKE and GLOB work. But we are thinking that most users expect a REGEXP pattern to match any substring of the input, unless the ^ prefix and/or $ suffix are used. ICU experts: The fix we want to make is to use uregex_find() instead of uregex_matches(). Will any legacy code break if we fix the ICU REGEXP operator to do what programmers normally expect? -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] SQLite on Windows Phone 8
Has anybody tried to install SQLite for Windows Phone 8? Any success? I can install it. It shows up fine under References, but then what? The Object Browser shows nothing. What namespace can I reference? What classes can I use? \Peter Original message: ** Topic: [sqlite] SQLite on Windows Phone 8 Hi Visual Studio 2012 for Windows Phone offers installation of SQLite for Windows Phone 8 as an extension. To enable programming against this extension a wrapper component or library needs to be installed. I am using sqlite-net installed via NuGet. But when compiling SQLite.cs, the compiler complains that the namespace Community.CsharpSqlite is not found. The same procedure works just fine in Visual Studio 2012 for Windows 8. For Windows Phone a DLL is needed that includes the Community.CsharpSqlite namespace. I have three questions: 1. Has anyone allready build this DLL (Community.CsharpSqlite.WP.dll), that is needed for sqlite-net? 2. How do I compile the DLL myself? 3. Are there other ways to utilize the SQLite for Windows Phone 8 extension? I eagerly awaits your comments and answers. \Peter Naldal ** ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] SQLite on Windows Phone 8
Peter Naldal wrote: Original message: ** Topic: [sqlite] SQLite on Windows Phone 8 Hi Visual Studio 2012 for Windows Phone offers installation of SQLite for Windows Phone 8 as an extension. To enable programming against this extension a wrapper component or library needs to be installed. I am using sqlite-net installed via NuGet. But when compiling SQLite.cs, the compiler complains that the namespace Community.CsharpSqlite is not found. The same procedure works just fine in Visual Studio 2012 for Windows 8. For Windows Phone a DLL is needed that includes the Community.CsharpSqlite namespace. I have three questions: 1. Has anyone allready build this DLL (Community.CsharpSqlite.WP.dll), that is needed for sqlite-net? 2. How do I compile the DLL myself? 3. Are there other ways to utilize the SQLite for Windows Phone 8 extension? I eagerly awaits your comments and answers. \Peter Naldal ** Did you see the response I sent previously? http://www.mail-archive.com/sqlite-users%40sqlite.org/msg73619.html -- Joe Mistachkin ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Anybody using REGEXP in the ICU extension?
I use REGEXP operator from ICU extension but the patch will not break my applications. 2012/11/5 Richard Hipp d...@sqlite.org Is anybody using the REGEXP operator that is included with the ICU extension of SQLite? The reason I ask is that the current ICU REGEXP implementation is broken. We are wanting to fix it. But the fix runs a serious risk of breaking any legacy applications that depend on the old broken behavior. The brokenness is this: The ICU REGEXP operator assumes that every pattern string begins with ^ and ends with $. In other words, the pattern must match the entire string. This is how LIKE and GLOB work. But we are thinking that most users expect a REGEXP pattern to match any substring of the input, unless the ^ prefix and/or $ suffix are used. ICU experts: The fix we want to make is to use uregex_find() instead of uregex_matches(). Will any legacy code break if we fix the ICU REGEXP operator to do what programmers normally expect? -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users -- Best regards, Alexey Pechnikov. http://pechnikov.tel/ ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] SQLite on Windows Phone 8
Hello Peter, This is a common configuration. Please visit the C#SQLite project here: http://code.google.com/p/csharp-sqlite/ and post questions to the usergroup here: https://groups.google.com/forum/?fromgroups#!forum/csharp-sqlite Please keep in mind the following: This is not an official version of SQLite There is completely NO affiliation with SQLite.org Bugs should not be reported to their ticket tracking system Thank you, Noah Hart -- View this message in context: http://sqlite.1065341.n5.nabble.com/SQLite-on-Windows-Phone-8-tp65316p65382.html Sent from the SQLite mailing list archive at Nabble.com. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Problem with create … from … order by on largish data
Is there a predicted next release date? Or more to the point, does anyone have a guess about when the fix for that 32-bit int overflow error will be in general release? -Tod On Nov 3, 2012, at 3:18 PM, Tod Olson t...@uchicago.edu wrote: On Nov 3, 2012, at 7:32 AM, Richard Hipp d...@sqlite.org wrote: On Fri, Nov 2, 2012 at 6:44 PM, Tod Olson t...@uchicago.edu wrote: I'm having a problem with a create … from … order by when my data starts approaching 2GB. I'm using SQLite 3.7.14 on FreeBSD 8.3-RELEASE-p3. The processing starts with an unsorted table, created thus: CREATE TABLE all_headings (key, heading); Then it creates a sorted version of the table: create table headings as select * from all_headings order by key; This is fine on small data, but when I load 1.8GB of data (8.8 million rows) the second CREATE fails, reporting a disk I/O error. You might be running out of /tmp space. Do you have plenty of /tmp space available for use by the sorter. Plenty, 14GB of free space available to /tmp (it's all one big partition). You might also be running into the 32-bit integer overflow bug that was fixed at http://www.sqlite.org/src/info/e24ba5bee4 though normally that requires a great deal more than 1.8GB of data. Yes! I compiled up that version and it solves the problem. I eagerly await the release of SQLite version 3.7.15. Thank you for your help. -Tod Please enable error logging using SQLITE_CONFIG_LOG ( http://www.sqlite.org/c3ref/c_config_getmalloc.html#sqliteconfiglog) and rerun your query and see if that provides any additional clues. If I remove the order by clause, the create succeeds. (SQLite was compiled with large file support, and I could create a 4GB database using .import so it's not a file system limitation, and the /tmp space is plenty large.) [At that point it looks like pre-sorting the data before loading has some appeal, but the code maintainer prefers to treat SQLite as the authority on sorting rather to mess with the many versions of sort(1) on the various UNIXes and Windows. I understand his point.] So trying to understand the error with the ORDER BY clause, I loaded up the unsorted all_headings table and then trussed sqlite3 running the CREATE TABLE…ORDER BY. Before the error, there's a lot of lseek()/read() of the .db file, and a lot of lseek()/write() to temp file (in /tmp, I assume this is the sort space). Then there's a read() of the temp file, which returns an error. From truss: read(5,0x800f64108,-1834983915) ERR#22 'Invalid argument' man 2 read says read's type signature is: ssize_t read(int d, void *buf, size_t nbytes); and it says this about read returning errno 22: [EINVAL] The pointer associated with d was negative. [EINVAL] The value nbytes is greater than INT_MAX. The pointer doesn't look negative, but that nbytes argument looks possibly a problem. size_t is 64-bit on this system, but INT_MAX = 2147483647, or the max for a 32-bit signed int. Though truss doesn't know signed from unsigned valued, the size_t nbytes value that truss reports is greater than MAX_INT. So I think that explains the error. The main question is: is there anything to be done to get that CREATE TABLE … ORDER BY to work? Based on the truss output, I suspect not, but maybe someone here has run into the problem before. Thanks, -Tod Tod Olson t...@uchicago.edu Systems Librarian University of Chicago Library ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Problem with create … from … order by on largish data
On Mon, Nov 5, 2012 at 3:58 PM, Tod Olson t...@uchicago.edu wrote: Is there a predicted next release date? Or more to the point, does anyone have a guess about when the fix for that 32-bit int overflow error will be in general release? http://www.sqlite.org/draft/ http://www.sqlite.org/draft/releaselog/3_7_15.html -Tod On Nov 3, 2012, at 3:18 PM, Tod Olson t...@uchicago.edu wrote: On Nov 3, 2012, at 7:32 AM, Richard Hipp d...@sqlite.org wrote: On Fri, Nov 2, 2012 at 6:44 PM, Tod Olson t...@uchicago.edu wrote: I'm having a problem with a create … from … order by when my data starts approaching 2GB. I'm using SQLite 3.7.14 on FreeBSD 8.3-RELEASE-p3. The processing starts with an unsorted table, created thus: CREATE TABLE all_headings (key, heading); Then it creates a sorted version of the table: create table headings as select * from all_headings order by key; This is fine on small data, but when I load 1.8GB of data (8.8 million rows) the second CREATE fails, reporting a disk I/O error. You might be running out of /tmp space. Do you have plenty of /tmp space available for use by the sorter. Plenty, 14GB of free space available to /tmp (it's all one big partition). You might also be running into the 32-bit integer overflow bug that was fixed at http://www.sqlite.org/src/info/e24ba5bee4 though normally that requires a great deal more than 1.8GB of data. Yes! I compiled up that version and it solves the problem. I eagerly await the release of SQLite version 3.7.15. Thank you for your help. -Tod Please enable error logging using SQLITE_CONFIG_LOG ( http://www.sqlite.org/c3ref/c_config_getmalloc.html#sqliteconfiglog) and rerun your query and see if that provides any additional clues. If I remove the order by clause, the create succeeds. (SQLite was compiled with large file support, and I could create a 4GB database using .import so it's not a file system limitation, and the /tmp space is plenty large.) [At that point it looks like pre-sorting the data before loading has some appeal, but the code maintainer prefers to treat SQLite as the authority on sorting rather to mess with the many versions of sort(1) on the various UNIXes and Windows. I understand his point.] So trying to understand the error with the ORDER BY clause, I loaded up the unsorted all_headings table and then trussed sqlite3 running the CREATE TABLE…ORDER BY. Before the error, there's a lot of lseek()/read() of the .db file, and a lot of lseek()/write() to temp file (in /tmp, I assume this is the sort space). Then there's a read() of the temp file, which returns an error. From truss: read(5,0x800f64108,-1834983915) ERR#22 'Invalid argument' man 2 read says read's type signature is: ssize_t read(int d, void *buf, size_t nbytes); and it says this about read returning errno 22: [EINVAL] The pointer associated with d was negative. [EINVAL] The value nbytes is greater than INT_MAX. The pointer doesn't look negative, but that nbytes argument looks possibly a problem. size_t is 64-bit on this system, but INT_MAX = 2147483647, or the max for a 32-bit signed int. Though truss doesn't know signed from unsigned valued, the size_t nbytes value that truss reports is greater than MAX_INT. So I think that explains the error. The main question is: is there anything to be done to get that CREATE TABLE … ORDER BY to work? Based on the truss output, I suspect not, but maybe someone here has run into the problem before. Thanks, -Tod Tod Olson t...@uchicago.edu Systems Librarian University of Chicago Library ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Problem with create … from … order by on largish data
Great! Thank you for the pointers. -Tod On Nov 5, 2012, at 3:21 PM, Richard Hipp d...@sqlite.org wrote: On Mon, Nov 5, 2012 at 3:58 PM, Tod Olson t...@uchicago.edu wrote: Is there a predicted next release date? Or more to the point, does anyone have a guess about when the fix for that 32-bit int overflow error will be in general release? http://www.sqlite.org/draft/ http://www.sqlite.org/draft/releaselog/3_7_15.html -Tod On Nov 3, 2012, at 3:18 PM, Tod Olson t...@uchicago.edu wrote: On Nov 3, 2012, at 7:32 AM, Richard Hipp d...@sqlite.org wrote: On Fri, Nov 2, 2012 at 6:44 PM, Tod Olson t...@uchicago.edu wrote: I'm having a problem with a create … from … order by when my data starts approaching 2GB. I'm using SQLite 3.7.14 on FreeBSD 8.3-RELEASE-p3. The processing starts with an unsorted table, created thus: CREATE TABLE all_headings (key, heading); Then it creates a sorted version of the table: create table headings as select * from all_headings order by key; This is fine on small data, but when I load 1.8GB of data (8.8 million rows) the second CREATE fails, reporting a disk I/O error. You might be running out of /tmp space. Do you have plenty of /tmp space available for use by the sorter. Plenty, 14GB of free space available to /tmp (it's all one big partition). You might also be running into the 32-bit integer overflow bug that was fixed at http://www.sqlite.org/src/info/e24ba5bee4 though normally that requires a great deal more than 1.8GB of data. Yes! I compiled up that version and it solves the problem. I eagerly await the release of SQLite version 3.7.15. Thank you for your help. -Tod Please enable error logging using SQLITE_CONFIG_LOG ( http://www.sqlite.org/c3ref/c_config_getmalloc.html#sqliteconfiglog) and rerun your query and see if that provides any additional clues. If I remove the order by clause, the create succeeds. (SQLite was compiled with large file support, and I could create a 4GB database using .import so it's not a file system limitation, and the /tmp space is plenty large.) [At that point it looks like pre-sorting the data before loading has some appeal, but the code maintainer prefers to treat SQLite as the authority on sorting rather to mess with the many versions of sort(1) on the various UNIXes and Windows. I understand his point.] So trying to understand the error with the ORDER BY clause, I loaded up the unsorted all_headings table and then trussed sqlite3 running the CREATE TABLE…ORDER BY. Before the error, there's a lot of lseek()/read() of the .db file, and a lot of lseek()/write() to temp file (in /tmp, I assume this is the sort space). Then there's a read() of the temp file, which returns an error. From truss: read(5,0x800f64108,-1834983915) ERR#22 'Invalid argument' man 2 read says read's type signature is: ssize_t read(int d, void *buf, size_t nbytes); and it says this about read returning errno 22: [EINVAL] The pointer associated with d was negative. [EINVAL] The value nbytes is greater than INT_MAX. The pointer doesn't look negative, but that nbytes argument looks possibly a problem. size_t is 64-bit on this system, but INT_MAX = 2147483647, or the max for a 32-bit signed int. Though truss doesn't know signed from unsigned valued, the size_t nbytes value that truss reports is greater than MAX_INT. So I think that explains the error. The main question is: is there anything to be done to get that CREATE TABLE … ORDER BY to work? Based on the truss output, I suspect not, but maybe someone here has run into the problem before. Thanks, -Tod Tod Olson t...@uchicago.edu Systems Librarian University of Chicago Library ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] light weight write barriers
On Mon, Nov 05, 2012 at 09:03:48PM +0100, Pavel Machek wrote: Well, using data journalling with ext3/4 may do what you want. If you don't do any fsync, the changes will get written every 5 seconds when the automatic journal sync happens (and sub-4k writes will also get Hmm. But that would need setting journalling mode per-file, no? Like, make it journal data for all the databases, but keep normal mode for rest of system... You can do that, using chattr +j file.db. It's apparently not a well known feature of ext3/4 - Ted ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] light weight write barriers
On Mon, Nov 5, 2012 at 5:04 PM, Theodore Ts'o ty...@mit.edu wrote: On Mon, Nov 05, 2012 at 09:03:48PM +0100, Pavel Machek wrote: Well, using data journalling with ext3/4 may do what you want. If you don't do any fsync, the changes will get written every 5 seconds when the automatic journal sync happens (and sub-4k writes will also get Hmm. But that would need setting journalling mode per-file, no? Like, make it journal data for all the databases, but keep normal mode for rest of system... You can do that, using chattr +j file.db. It's apparently not a well known feature of ext3/4 Per the docs: Only the superuser or a process possessing the CAP_SYS_RESOURCE capability can set or clear this attribute. That prevents most applications that run SQLite from being able to take advantage of this, since most such applications lack elevated privileges. - Ted ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] light weight write barriers
On Mon, Nov 05, 2012 at 05:37:02PM -0500, Richard Hipp wrote: Per the docs: Only the superuser or a process possessing the CAP_SYS_RESOURCE capability can set or clear this attribute. That prevents most applications that run SQLite from being able to take advantage of this, since most such applications lack elevated privileges. If this feature would prove useful to sqllite, that's something we could address. I could image making this available to processes that belong to a specific group that would be specified in the superblock or as a mount option. (We already have something like that which allows a specific uid or gid to use the reserved space in the superblock.) - Ted ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
