mysqlfs-commits — Messages from CVS commits, forums, bug tracker, etc

[Mysqlfs-commits] [Help] RE: Using FUSE option problem

[SourceForge.net <no...@so...>]

The following forum message was posted by Anonymous at http://sourceforge.net/projects/mysqlfs/forums/forum/441304/topic/1813046:

As a work around you can specify the options to pass to FUSE like:

mysqlfs -o\-oallow_other -ohost=localhost -ouser=user -opassword=password
-odatabase=mysqlfs  ./fs

note the \-0 escape. this passes -oallow_other to the fuse main

[Mysqlfs-commits] [ mysqlfs-Patches-3154217 ] OPTIMIZATIONS

[SourceForge.net <no...@so...>]

Patches item #3154217, was opened at 2011-01-10 10:23
Message generated for change (Tracker Item Submitted) made by rspadim
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716427&aid=3154217&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Roberto Spadim (rspadim)
Assigned to: Nobody/Anonymous (nobody)
Summary: OPTIMIZATIONS

Initial Comment:
 Hi guys, i was reading SQL code (query.c), and TODO...
I think that mysqlfs developers don't know what is mysqlfs... is it a local filesystem or a network filesystem? (ok i know you know what's mysqlfs, but what's mysql idea? a network or a local database?)
It's very important question. a local filesystem can use shared memory (ext2), a network filesystem should use server based shared memory (smbfs, cifs, nfs). for local filesystem only we can use SQLITE, Berkley, others... for network filesystem we NEED a network database (mysql, oracle, postgres, firebird)... ok? mysql is fast because of MYISAM and fast protocol (we can use compress for a slow connection)

FOR SELECT OPTIMIZATIONS:
http://dev.mysql.com/doc/refman/5.0/en/select.html

When we need speed we could use SELECT optimizations...
for example at 'query.c':
all SELECT query, could use SELECT SQL_CACHE, all SELECT with small results can use SMALL_RESULT (inodes), all big result could use BIG_RESULT (data), if using mysql in local database we don't have network problems but using a network (TCP) we could use SQL_BUFFERED_RESULT (to allow read/write intensive and compressed protocol)
anothers optimizations... don't use NULL values, instead this use negative values, or fixed value (constants), for example for '/' we could use always -1 inode (this allow faster REPAIR TABLE, OPTIMIZE TABLE, CHECK TABLE queries, and allow small tables, since NULL is a `constant` value, we could use a CONSTANT value for each field type) for only unsigned column we can use a signed column and set negative values for NULL


Another interesting think... we should allow per ENGINE optimization... for example a using a INNODB optimizated query may be slow on MYISAM engine or an NDB or a ARIA or a PBXT, MEMORY, etc etc... what i sugest? MYISAM/ARIA (MYISAM with crash safe - see mariadb) or NDB (cluster filesystem!!!)
Some mysql (mariadb) tendencies... 
MYISAM for not important tables (log/select intensive) (no crash safe, but can be used in filesystem with crash safe...) and table lock level
ARIA for important tables (with crash safe, can be used in ext2 filesystem) and page lock level
INNODB for important tables (with crash safe, can be used in devices without filesystem at partition or md (raid) devices) row lock level
NDB (cluster, distribuited, crash safe shared-nothing) very interesting... we HAVE A CLUSTER FILE SYSTEM!!!!


for FSCK code:
why not a 
CHECK TABLE tables...
REPAIR TABLE tables...

for DELETE/PURGE code:
for every purge, or many deletes we SHOULD use OPTIMIZE TABLES tables...

a optimization but with 2 bytes per row (we are using 4096 block size), add at data_blocks table, the data_size field, at each UPDATE/INSERT we should set it as LENGTH(data), and at filesize we can use SUM(data_size) instead of: SELECT seq, data, LENGTH(data) FROM data_blocks, we could use SELECT seq,data,data_length

this is faster... (for a large number of files) since on each update/insert we can calculate it we will not have problem with non-deterministic values
we should but it in fsck too...


this could be translated to:
	     "UPDATE inodes SET size=("
	     	"SELECT seq*%d + LENGTH(data) FROM data_blocks WHERE inode=%ld AND seq=("
			"SELECT MAX(seq) FROM data_blocks WHERE inode=%ld"
		")"
	     ") "
	     "WHERE inode=%ld",
------
	     "UPDATE inodes SET size=("
	     	"SELECT SUM(data_length) FROM data_blocks WHERE inode=%ld)"
	     ") "
	     "WHERE inode=%ld",

another point....
since we use netowork system, TIME functions SHOULD be sincronized, we NEED a DATABASE SIDE DATE/TIME value for UPDATE atime, mtime, utime... Xtime inode values... (NOW() is a good function)
since we use network + database, we have two problems for block size... one is database harddisk (ssd) block size, most ssd use 4096 bytes, some use more, some less, but the main point here is: NETWORK block size, mysql have some variables at database side, i don't remember but i think it's network_buffer or somethink like it, maybe a ****per inode block size**** COULD be implemented?

some queries are very ugly and end with  ';', ok it work...
    delete from data_blocks where inode not in (select inode from inodes);
could be:
    DELETE FROM data_blocks WHERE inode NOT IN (SELECT inode FROM inodes)

i don't understand this:
	snprintf(sql, SQL_MAX, "select inode, sum(OCTET_LENGTH(data)) as size from data_blocks group by inode");
and after:
		snprintf(sql, SQL_MAX, "update inodes set size=%ld where inode=%ld;", size, inode);

why not merge? and why use OCTET_LENGTH and LEGTH for data column?
check:
	UPDATE inodes AS a SET size=( SELECT SUM(LENGTH(data)) FROM data_blocks WHERE inode=a.inode )

ok we can use data_length field... for example:
	UPDATE data_blocks SET data_length=LENGTH(data_length);
	UPDATE inodes AS a SET size=( SELECT SUM(data_length) FROM data_blocks WHERE inode=a.inode )

and why not LOCK tables when making a FSCK?
LOCK TABLE .... FOR WRITE
UNLOCK TABLES



I DIDN'T CHECKED LOCKS! we can lock a file?
what about using per row lock, page lock or table lock at mysql server side?
some updates should be transaction safe, for example update two data_block, should lock rows and after change it and unlock rows
for all engines- lock table, change data_block rows, unlock table
for row/page lock - we should check per engine lock:

check this INNODB locking:
	SELECT counter_field FROM child_codes FOR UPDATE;
	UPDATE child_codes SET counter_field = counter_field + 1;
	http://dev.mysql.com/doc/refman/5.0/en/innodb-locking-reads.html

	Note
		Locking of rows for update using SELECT FOR UPDATE only applies when autocommit is disabled (either by beginning transaction with START TRANSACTION or by setting autocommit to 0. If autocommit is enabled, the rows matching the specification are not locked.

if we want implement a lock inside mysql, we should implement somethink like:
create table locks(
inode,
seq,
offset,
size,
connection_id,
uptime_at_lock_time,
LOCK_ID) ENGINE=memory or NDB or MYISAM (don't need to be fail safe if at fsck we TRUNCATE locks)

and at mysqlfs init function we could check SHOW PROCESSLIST check if connection_id is OK (a wathdog work better), and if uptime is bigger than current uptime, some UNIQUE lock time should be used too, maybe per mysqlfs mount point maybe LOCK_ID= option (per mysqlfs client lock)
the problem? with mysql we can't call another CLIENT with default mysql features (we can use plugin, ok dba don't like plugins...) but we could use a message table with a watchdog thread at mysqlfs side

create table watch_dog(
LOCK_ID,
expire_time
)

and a thread only to update it
UPDATE watch_dog SET expire_time=NOW() + some time... WHERE LOCK_ID=(mysqlfs option lock_id maybe MAC ADDRESS? a string or a int field?)

if the watch_dog isn't working we know that a lock is dead with:
LOCK TABLE locks...
DELETE FROM locks WHERE LOCK_ID IN (SELECT LOCK_ID FROM watch_dog WHERE expire_time<NOW())
SELECT COUNT(*) FROM locks WHERE inode=(INODE) AND seq IN (seqs to lock) LOCK_ID!=(CURRENT CONNECTION LOCK ID)

if COUNT(*)>0 we have a lock
if not we should lock it...

UNLOCK TABLES...
there's some ideas... but for a good and fast lock we could put it at data_blocks table... and a !=0 lock field with index...


nice?
some features are easy to implement, some not, but it's nice, not?
thanks guys, i'm using it at office for a cluster filesystem today (using NDB)! :D it works but it's not write intensive... just for reads
i could help at: 
roberto at spadim dot com dot br
bye

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716427&aid=3154217&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Patches-3074652 ] Extended attribute (xattr) support (fixed)

[SourceForge.net <no...@so...>]

Patches item #3074652, was opened at 2010-09-24 09:45
Message generated for change (Tracker Item Submitted) made by jj101k
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716427&aid=3074652&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Jim D (jj101k)
Assigned to: Nobody/Anonymous (nobody)
Summary: Extended attribute (xattr) support (fixed)

Initial Comment:
As before, this adds:

.setxattr
.getxattr
.listxattr
.removexattr

It also:
- logs when it opens the log file (useful for "is it logging???" questions), but it's not difficult to strip that out; 
- includes a query.h/log.h include fix to support loading from a separate .c file (again, easy to strip).
- includes a generated SQL file rather than a hand-edited one (needed for the extra table)

Caveats:
- A small number of commented out log statements (should really have set them for the debug level)
- Adds a dependency to the inode delete routine
- (As before) it would make sense for someone who groks conventional limits on name and value length to tweak the schema, as I couldn't find it firmly documented anywhere, but the existing limits should be fine
- (As before) would return EIO for an excessively large name or value (as it would fail at the db level); possibly this should be ERANGE

Changes since last time (a year ago!) are:
- Now logs every error condition return I could find in xattr.c, except the ones which are more strictly warnings.
- Apparently under some circumstances which are apparently load-related, MySQL-client would overwrite the integer "size" with random data when returning NULL. It now checks for that and returns 0.
- Updated SQL to avoid returning NULL in the first place (listxattr)

Notes:
This has been in production use for some time

Usage (once mounted):
touch myfileordirectory
setfattr -n user.foo -v bar myfileordirectory
getfattr -d myfileordirectory

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716427&aid=3074652&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-2994981 ] make error query.c

[SourceForge.net <no...@so...>]

Bugs item #2994981, was opened at 2010-05-01 02:09
Message generated for change (Tracker Item Submitted) made by 
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=2994981&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: darzanebor ()
Assigned to: Nobody/Anonymous (nobody)
Summary: make error query.c

Initial Comment:
2.6.32-21x86_64 GNU/Linux
...
make[2]: ./make/mysqlfs-0.3.99.2\'
if gcc -DHAVE_CONFIG_H -I. -I. -I.   -I/usr/local/include -I/usr/include/mysql  -g -O2 -I/usr/local/include -Wall -Werror -I/usr/include/mysql -Wall -Werror -MT query.o -MD -MP -MF \".deps/query.Tpo\" -c -o query.o query.c; \\
	then mv -f \".deps/query.Tpo\" \".deps/query.Po\"; else rm -f \".deps/query.Tpo\"; exit 1; fi
cc1: warnings being treated as errors
query.c: In function ‘query_truncate’:
query.c:298: error: format ‘%lld’ expects type ‘long long int’, but argument 4 has type ‘off_t’
query.c:298: error: format ‘%lld’ expects type ‘long long int’, but argument 4 has type ‘off_t’
query.c: In function ‘write_one_block’:
query.c:795: error: format ‘%llu’ expects type ‘long long unsigned int’, but argument 4 has type ‘off_t’
query.c:795: error: format ‘%llu’ expects type ‘long long unsigned int’, but argument 4 has type ‘off_t’
query.c:799: error: format ‘%llu’ expects type ‘long long unsigned int’, but argument 4 has type ‘long unsigned int’
query.c:799: error: format ‘%llu’ expects type ‘long long unsigned int’, but argument 4 has type ‘long unsigned int’
make[2]: *** [query.o] Error 1


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=2994981&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Patches-2887810 ] Extended attribute (xattr) support

[SourceForge.net <no...@so...>]

Patches item #2887810, was opened at 2009-10-28 09:50
Message generated for change (Comment added) made by nobody
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716427&aid=2887810&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: Extended attribute (xattr) support

Initial Comment:
Adds:
.setxattr
.getxattr
.listxattr
.removexattr

This is the raw patch from the quick implementation I've made. Of note: 

- the bulk of the code is in a separate file, xattr.c; this makes no attempt to separate queries from the top level. It should be relatively easy to #ifdef the code out.
- it adds a few "#include"s to log.h and query.h to support loading those from a file which hasn't already "#include"d everything
- schema.sql should probably be loaded and re-dumped by whoever dumped it last time, as the 5.1 format used would add unnecessary noise to the commit - all I actually did was add one table.
- some commented out log statements; possibly excessive logging compared to the rest of the project
- adds a dependancy to the inode-delete routine which might be syntactically invalid on older mysql? This is to effectively simulate foreign key ON DELETE CASCADE behaviour, as extended attributes must go in a separate table but logically should be removed when the main attributes (mode, etc.) go.
- it would make sense for someone who groks conventional limits on name and value length to tweak the schema, as I couldn't find it firmly documented anywhere, but the existing limits should be fine
- would return EIO for an excessively large name or value (as it would fail at the db level); possibly this should be ERANGE

This tests fine with setfattr and getfattr -d --match=.*

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-10-28 09:57

Message:
Also of note: I didn't get around to documenting the functions, but they
are consistent with setxattr(2), getxattr(2) removexattr(2) and
listxattr(2).

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716427&aid=2887810&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Patches-2887810 ] Extended attribute (xattr) support

[SourceForge.net <no...@so...>]

Patches item #2887810, was opened at 2009-10-28 09:50
Message generated for change (Tracker Item Submitted) made by nobody
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716427&aid=2887810&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: Extended attribute (xattr) support

Initial Comment:
Adds:
.setxattr
.getxattr
.listxattr
.removexattr

This is the raw patch from the quick implementation I've made. Of note: 

- the bulk of the code is in a separate file, xattr.c; this makes no attempt to separate queries from the top level. It should be relatively easy to #ifdef the code out.
- it adds a few "#include"s to log.h and query.h to support loading those from a file which hasn't already "#include"d everything
- schema.sql should probably be loaded and re-dumped by whoever dumped it last time, as the 5.1 format used would add unnecessary noise to the commit - all I actually did was add one table.
- some commented out log statements; possibly excessive logging compared to the rest of the project
- adds a dependancy to the inode-delete routine which might be syntactically invalid on older mysql? This is to effectively simulate foreign key ON DELETE CASCADE behaviour, as extended attributes must go in a separate table but logically should be removed when the main attributes (mode, etc.) go.
- it would make sense for someone who groks conventional limits on name and value length to tweak the schema, as I couldn't find it firmly documented anywhere, but the existing limits should be fine
- would return EIO for an excessively large name or value (as it would fail at the db level); possibly this should be ERANGE

This tests fine with setfattr and getfattr -d --match=.*

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716427&aid=2887810&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1849644 ] configure pthread test failing

[SourceForge.net <no...@so...>]

Bugs item #1849644, was opened at 2007-12-12 21:35
Message generated for change (Settings changed) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Closed
>Resolution: Fixed
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: chickenandporn (chickenandporn)
Summary: configure pthread test failing

Initial Comment:
Trying to compile of fc7 or fc7 I get this output:

checking for pthread_create in -lpthread... no
configure: error: Please install pthreads library first.

But it is installed with glibc-headers-2.6-4 and glibc-devel.

I am unable to compile this on a modern Fedora distro!

respond to: da...@ge...

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-07-15 01:53

Message:
Configure changed a bit, should be more responsive, but does hide which
libraries are required for pthread et al.  The error message is accurate,
though.  Fixed in 0.4.0, reopen if I missed your situation.

----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-06-28 04:25

Message:
Sure, commit whatever you like. I don't require you to ask for approval
before every check-in. Now it's your project ;-)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 01:31

Message:
OK.  Testing with a mysql-devel but missing fuse-devel shows that the LDADD
and such are being given a "-L${FUSE_LIB_DIR}", but with no fuse, that
means that LDADD += "-L ", which seems to break the later test builds. 
I've protected those behind a "if test x${FUSE_LIB_DIR} != x".  Done the
same for the MySQL stuff, since it also uses its own fallback to find MySQL
if the user hasn't give one.

I also notice that letting it run normally gets a "... -l mysqlclient_r
... -l mysqlclient" which seems redundant.  AC_SEARCH_LIBS allows it to
check without adding -lmysqlclient first, which might mislead the user but
accurately avoids duplicate mysqlclient libs.

Index: configure.in
===================================================================
--- configure.in	(revision 47)
+++ configure.in	(working copy)
@@ -48,11 +48,14 @@
     break
   fi
 done
+if test x${FUSE_LIB_DIR} != x
+then
 FUSE_LDFLAGS="-L${FUSE_LIB_DIR}"
 FUSE_CFLAGS="-I${FUSE_INC_DIR}"
 LDFLAGS="$LDFLAGS $FUSE_LDFLAGS"
 CFLAGS="$CFLAGS $FUSE_CFLAGS"
 CPPFLAGS="$CPPFLAGS $FUSE_CFLAGS"
+fi
 
 MYSQL_CONFIG=$(which mysql_config)
 if test "$MYSQL_CONFIG"; then
@@ -72,17 +75,20 @@
       break
     fi
   done
+  if test x${MYSQL_LIB_DIR} != x
+  then
   MYSQL_LDFLAGS="-L${MYSQL_LIB_DIR}"
   MYSQL_CFLAGS="-I${MYSQL_INC_DIR}"
 fi
+fi
 LDFLAGS="$LDFLAGS $MYSQL_LDFLAGS"
 CFLAGS="$CFLAGS $CFLAGS_ADD $MYSQL_CFLAGS"
 CPPFLAGS="$CPPFLAGS $MYSQL_CFLAGS"
 
 dnl Checks for libraries.
-AC_CHECK_LIB(pthread, pthread_create,, AC_MSG_ERROR([Please install
pthreads library first.]))
-AC_CHECK_LIB(mysqlclient, mysql_init,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
-AC_CHECK_LIB(fuse, fuse_main,, AC_MSG_ERROR([Please install fuse library
first.]))
+AC_SEARCH_LIBS(mysql_init, mysqlclient,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
+AC_SEARCH_LIBS(pthread_create, pthread,, AC_MSG_ERROR([Please install
pthreads library first.]))
+AC_SEARCH_LIBS(fuse_main, fuse,, AC_MSG_ERROR([Please install fuse
library first.]))
 
 dnl Checks for header files. (mac -- and BSD? -- have statfs in mount.h)
 AC_CHECK_HEADERS(stdio.h sys/param.h sys/mount.h)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 01:07

Message:
The error seems to be that "mysql_config" is part of "mysql", not part of
"mysql-devel", so the configure finds a "mysql_config" and interrogates
it:

MYSQL_CONFIG=$(which mysql_config)
if test "$MYSQL_CONFIG"; then
  MYSQL_LDFLAGS=$(${MYSQL_CONFIG} --libs_r)
...
...
LDFLAGS="$LDFLAGS $MYSQL_LDFLAGS"


Problem is, that generates a LDFLAGS that doesn't actually work because
the linker cannot find -lmysqlclient_r (which is part of mysql-devel).  It
just so happens that the first compilation is pthreads, not mysql.

I would recommend swapping the order.

At the same time, I would recommend moving the pthreads check to
AC_SEARCH_LIBS rather than AC_CHECK_LIB so that OSes with it already
without a -lpthreads are properly detected as "happy without -lpthreads"
(or the EDG variant: -K thread )

Whaddaya say to that?

--- configure.in	(revision 47)
+++ configure.in	(working copy)
@@ -80,8 +80,8 @@
 CPPFLAGS="$CPPFLAGS $MYSQL_CFLAGS"
 
 dnl Checks for libraries.
-AC_CHECK_LIB(pthread, pthread_create,, AC_MSG_ERROR([Please install
pthreads library first.]))
 AC_CHECK_LIB(mysqlclient, mysql_init,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
+AC_SEARCH_LIBS(pthread_create, pthread,, AC_MSG_ERROR([Please install
pthreads library first.]))
 AC_CHECK_LIB(fuse, fuse_main,, AC_MSG_ERROR([Please install fuse library
first.]))

...
checking for mysql_init in -lmysqlclient... no
configure: error: Please install mysqlclient library first.
make: *** [config.status] Error 1



----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-06-28 00:48

Message:
Hi Allan, simply uninstall mysql-devel and fuse-devel and you should be
able to trigger this error with ./configure. IIRC there's something wrong
with CFLAGS or LDFLAGS handling - when mysql-devel is missing it makes
CFLAGS invalid which in turn breaks pthread test. It's indeed my fault but
I never got around to fix it. Feel free to do so ;-)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 17:57

Message:
Hey Dan, if you are willing to get back on the diagnostic side, can you
attach a config.log below, by expanding the "Attached file" and uploading a
config.log from your dev environment?

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 16:07

Message:
Hi Dan; I'd like to address your issue in a next release; I'm a bit new
here as well.  Can I send you another tarball to try?  As well, I don't
expect this to magically resolve the issue, are you willing to banter a bit
to resolve the issue on your platform?

----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-03-03 01:16

Message:
IIRC the message is misleading - in fact it means you don't have
mysql-devel (or libmysql-dev or how is it called on debian) package
installed. You'll also need a development package for fuse (fuse-devel or
libfuse-dev).

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-03-03 00:12

Message:
I have the same problem, "configure: error: Please install pthreads library
first."
if i do a locate pthread i get
"/lib/libpthread-2.7.so
/lib/libpthread.so.0
/lib/tls/i686/cmov/libpthread-2.7.so
/lib/tls/i686/cmov/libpthread.so.0
/usr/NX/lib/perl/include/bits/pthreadtypes.ph
/usr/include/pthread.h
/usr/include/bits/pthreadtypes.h
/usr/lib/libgpgme-pthread.so.11
/usr/lib/libgpgme-pthread.so.11.6.4
/usr/lib/libpthread.a
/usr/lib/libpthread.so
/usr/lib/libpthread_nonshared.a
/usr/lib/perl/5.8.8/bits/pthreadtypes.ph
/usr/lib/xen/libpthread.a
/usr/lib/xen/libpthread_nonshared.a
/usr/share/man/man7/pthreads.7.gz
"

Am I doing something wrong? I'm using ubuntu 8.10 and i have installed
allmost anything with "thread" as package name..
any help?

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-12-12 22:01

Message:
Logged In: NO 

I meant fc6 or fc7, 32bit or 64bit

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-2096705 ] syntax error or argument missing in query.c

[SourceForge.net <no...@so...>]

Bugs item #2096705, was opened at 2008-09-06 09:24
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=2096705&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
>Status: Closed
Resolution: Works For Me
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: chickenandporn (chickenandporn)
Summary: syntax error or argument missing in query.c

Initial Comment:
on compiling query.c, the compiler crashes because is waiting for the string "%lld" in a sprintf command, and after the string only 2 arguments are given (a long-typed one and an integer), so the format string is mistyped or an argument is missing.

lines 235, 593 and 597 in file query.c

----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-07-15 01:54

Message:
Checked the lines in question, I can't see the error in them.  Compiler
cannot either, it passes by the line happily.

Need the original poster to give more details, such as a dump of the
message and a version that's affected.  :(

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 18:58

Message:
I'm marking this as "works for me" -- because it does -- but I'm looking
for your response where it doesn't "work for you".  Maybe we need to
augment the project for portability.

I'd like to close this after 2009-07-31 if we don't see a response.  I
don't mean to be draconian, I'd like to resolve this if we can still get
feedback.

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 18:49

Message:
Just checking this, line 235 involves off_t length which is indirectly
defined as a __int64_t on my platform.  line 593 is similar.  The only
dodgy one might be line 597: length + size + 1 -- where size is a "size_t",
indirectly defined as an unsigned long.  The compiler should convert the
unsigned long to a long long for the math, it should be fine.

What's your platform?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=2096705&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1849644 ] configure pthread test failing

[SourceForge.net <no...@so...>]

Bugs item #1849644, was opened at 2007-12-12 21:35
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
>Status: Closed
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: chickenandporn (chickenandporn)
Summary: configure pthread test failing

Initial Comment:
Trying to compile of fc7 or fc7 I get this output:

checking for pthread_create in -lpthread... no
configure: error: Please install pthreads library first.

But it is installed with glibc-headers-2.6-4 and glibc-devel.

I am unable to compile this on a modern Fedora distro!

respond to: da...@ge...

----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-07-15 01:53

Message:
Configure changed a bit, should be more responsive, but does hide which
libraries are required for pthread et al.  The error message is accurate,
though.  Fixed in 0.4.0, reopen if I missed your situation.

----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-06-28 04:25

Message:
Sure, commit whatever you like. I don't require you to ask for approval
before every check-in. Now it's your project ;-)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 01:31

Message:
OK.  Testing with a mysql-devel but missing fuse-devel shows that the LDADD
and such are being given a "-L${FUSE_LIB_DIR}", but with no fuse, that
means that LDADD += "-L ", which seems to break the later test builds. 
I've protected those behind a "if test x${FUSE_LIB_DIR} != x".  Done the
same for the MySQL stuff, since it also uses its own fallback to find MySQL
if the user hasn't give one.

I also notice that letting it run normally gets a "... -l mysqlclient_r
... -l mysqlclient" which seems redundant.  AC_SEARCH_LIBS allows it to
check without adding -lmysqlclient first, which might mislead the user but
accurately avoids duplicate mysqlclient libs.

Index: configure.in
===================================================================
--- configure.in	(revision 47)
+++ configure.in	(working copy)
@@ -48,11 +48,14 @@
     break
   fi
 done
+if test x${FUSE_LIB_DIR} != x
+then
 FUSE_LDFLAGS="-L${FUSE_LIB_DIR}"
 FUSE_CFLAGS="-I${FUSE_INC_DIR}"
 LDFLAGS="$LDFLAGS $FUSE_LDFLAGS"
 CFLAGS="$CFLAGS $FUSE_CFLAGS"
 CPPFLAGS="$CPPFLAGS $FUSE_CFLAGS"
+fi
 
 MYSQL_CONFIG=$(which mysql_config)
 if test "$MYSQL_CONFIG"; then
@@ -72,17 +75,20 @@
       break
     fi
   done
+  if test x${MYSQL_LIB_DIR} != x
+  then
   MYSQL_LDFLAGS="-L${MYSQL_LIB_DIR}"
   MYSQL_CFLAGS="-I${MYSQL_INC_DIR}"
 fi
+fi
 LDFLAGS="$LDFLAGS $MYSQL_LDFLAGS"
 CFLAGS="$CFLAGS $CFLAGS_ADD $MYSQL_CFLAGS"
 CPPFLAGS="$CPPFLAGS $MYSQL_CFLAGS"
 
 dnl Checks for libraries.
-AC_CHECK_LIB(pthread, pthread_create,, AC_MSG_ERROR([Please install
pthreads library first.]))
-AC_CHECK_LIB(mysqlclient, mysql_init,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
-AC_CHECK_LIB(fuse, fuse_main,, AC_MSG_ERROR([Please install fuse library
first.]))
+AC_SEARCH_LIBS(mysql_init, mysqlclient,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
+AC_SEARCH_LIBS(pthread_create, pthread,, AC_MSG_ERROR([Please install
pthreads library first.]))
+AC_SEARCH_LIBS(fuse_main, fuse,, AC_MSG_ERROR([Please install fuse
library first.]))
 
 dnl Checks for header files. (mac -- and BSD? -- have statfs in mount.h)
 AC_CHECK_HEADERS(stdio.h sys/param.h sys/mount.h)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 01:07

Message:
The error seems to be that "mysql_config" is part of "mysql", not part of
"mysql-devel", so the configure finds a "mysql_config" and interrogates
it:

MYSQL_CONFIG=$(which mysql_config)
if test "$MYSQL_CONFIG"; then
  MYSQL_LDFLAGS=$(${MYSQL_CONFIG} --libs_r)
...
...
LDFLAGS="$LDFLAGS $MYSQL_LDFLAGS"


Problem is, that generates a LDFLAGS that doesn't actually work because
the linker cannot find -lmysqlclient_r (which is part of mysql-devel).  It
just so happens that the first compilation is pthreads, not mysql.

I would recommend swapping the order.

At the same time, I would recommend moving the pthreads check to
AC_SEARCH_LIBS rather than AC_CHECK_LIB so that OSes with it already
without a -lpthreads are properly detected as "happy without -lpthreads"
(or the EDG variant: -K thread )

Whaddaya say to that?

--- configure.in	(revision 47)
+++ configure.in	(working copy)
@@ -80,8 +80,8 @@
 CPPFLAGS="$CPPFLAGS $MYSQL_CFLAGS"
 
 dnl Checks for libraries.
-AC_CHECK_LIB(pthread, pthread_create,, AC_MSG_ERROR([Please install
pthreads library first.]))
 AC_CHECK_LIB(mysqlclient, mysql_init,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
+AC_SEARCH_LIBS(pthread_create, pthread,, AC_MSG_ERROR([Please install
pthreads library first.]))
 AC_CHECK_LIB(fuse, fuse_main,, AC_MSG_ERROR([Please install fuse library
first.]))

...
checking for mysql_init in -lmysqlclient... no
configure: error: Please install mysqlclient library first.
make: *** [config.status] Error 1



----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-06-28 00:48

Message:
Hi Allan, simply uninstall mysql-devel and fuse-devel and you should be
able to trigger this error with ./configure. IIRC there's something wrong
with CFLAGS or LDFLAGS handling - when mysql-devel is missing it makes
CFLAGS invalid which in turn breaks pthread test. It's indeed my fault but
I never got around to fix it. Feel free to do so ;-)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 17:57

Message:
Hey Dan, if you are willing to get back on the diagnostic side, can you
attach a config.log below, by expanding the "Attached file" and uploading a
config.log from your dev environment?

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 16:07

Message:
Hi Dan; I'd like to address your issue in a next release; I'm a bit new
here as well.  Can I send you another tarball to try?  As well, I don't
expect this to magically resolve the issue, are you willing to banter a bit
to resolve the issue on your platform?

----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-03-03 01:16

Message:
IIRC the message is misleading - in fact it means you don't have
mysql-devel (or libmysql-dev or how is it called on debian) package
installed. You'll also need a development package for fuse (fuse-devel or
libfuse-dev).

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-03-03 00:12

Message:
I have the same problem, "configure: error: Please install pthreads library
first."
if i do a locate pthread i get
"/lib/libpthread-2.7.so
/lib/libpthread.so.0
/lib/tls/i686/cmov/libpthread-2.7.so
/lib/tls/i686/cmov/libpthread.so.0
/usr/NX/lib/perl/include/bits/pthreadtypes.ph
/usr/include/pthread.h
/usr/include/bits/pthreadtypes.h
/usr/lib/libgpgme-pthread.so.11
/usr/lib/libgpgme-pthread.so.11.6.4
/usr/lib/libpthread.a
/usr/lib/libpthread.so
/usr/lib/libpthread_nonshared.a
/usr/lib/perl/5.8.8/bits/pthreadtypes.ph
/usr/lib/xen/libpthread.a
/usr/lib/xen/libpthread_nonshared.a
/usr/share/man/man7/pthreads.7.gz
"

Am I doing something wrong? I'm using ubuntu 8.10 and i have installed
allmost anything with "thread" as package name..
any help?

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-12-12 22:01

Message:
Logged In: NO 

I meant fc6 or fc7, 32bit or 64bit

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1681567 ] After copy the file is changed.

[SourceForge.net <no...@so...>]

Bugs item #1681567, was opened at 2007-03-15 16:23
Message generated for change (Settings changed) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1681567&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
>Status: Closed
Resolution: Works For Me
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: chickenandporn (chickenandporn)
Summary: After copy the file is changed.

Initial Comment:
Hello,

I'm working with mysqlfs, and very interested. It could be a backup-back-end. 

But now I run into some problems:

after I mount it (with root privileges!) I've got a mountpoint:

mysqlfs on /tmp/mysqlfs type fuse (rw,nosuid,nodev)

Now I copy a file to it:

cp ~/download/linux-2.6.20.2.tar.bz2 /tmp/mysqlfs

this goes fine. But now I want to unpack it:
cd /tmp
tar -xvf mysqlfs/linux-2.6.20.2.tar.bz2 

I get the error:

tar: This does not look like a tar archive
tar: Skipping to the next header
tar: Archive contains obsolescent base-64 headers
tar: Error exit delayed from previous error

Now this is not ok. Then I do:

diff ~/download/linux-2.6.20.2.tar.bz2 \
     mysqlfs/linux-2.6.20.2.tar.bz2

Binary files differ.

How is this possible?

Is this a new error or....?

Stef Bon





----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-07-15 01:52

Message:
Verified works fine in the regression-testing in 0.4.0

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 03:07

Message:
The bug is a bit old, but I hope the original poster comes back with
details.  No response in a month (2009-07-28), let's consider closing this,
although I'd prefer the original poster coming back, re-trying, falling
back in love with mysqlfs, and disappearing into the sunset, a happy
ending.  Maybe a postcard.

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 03:04

Message:
I just added revision 48, which includes a testcase that mounts the
filesystem and copies a file onto it, then compares the checksum.  Tested
with linux-2.6.30.tar.bz2, checksum was a match.

I'm shifting this to a "works for me", but I hope you're willing to come
back and offer more details, or check out rev 48 of the SVN and try it: 
./configure --with-testfile=linux-2.6.20.2.tar.bz2 && make check

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1681567&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Feature Requests-1773343 ] Improve usage

[SourceForge.net <no...@so...>]

Feature Requests item #1773343, was opened at 2007-08-13 18:23
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716428&aid=1773343&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
>Status: Closed
>Resolution: Fixed
Priority: 5
Private: No
Submitted By: Stephen More (mores)
Assigned to: Nobody/Anonymous (nobody)
Summary: Improve usage

Initial Comment:
Usage should also contain:

mysqlfs -osocket=/tmp/mysql5.sock -ouser=mysqlfs -opassword=password -odatabase=mysqlfs  /var/mysqlfs



----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-07-01 09:30

Message:
Resolved in commit r50-51, for release on v0.4.0.  We can also take the
same parameters now as the mysql command-line client, making it easier to
remember the options.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716428&aid=1773343&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1681567 ] After copy the file is changed.

[SourceForge.net <no...@so...>]

Bugs item #1681567, was opened at 2007-03-15 16:23
Message generated for change (Settings changed) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1681567&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: Works For Me
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
>Assigned to: chickenandporn (chickenandporn)
Summary: After copy the file is changed.

Initial Comment:
Hello,

I'm working with mysqlfs, and very interested. It could be a backup-back-end. 

But now I run into some problems:

after I mount it (with root privileges!) I've got a mountpoint:

mysqlfs on /tmp/mysqlfs type fuse (rw,nosuid,nodev)

Now I copy a file to it:

cp ~/download/linux-2.6.20.2.tar.bz2 /tmp/mysqlfs

this goes fine. But now I want to unpack it:
cd /tmp
tar -xvf mysqlfs/linux-2.6.20.2.tar.bz2 

I get the error:

tar: This does not look like a tar archive
tar: Skipping to the next header
tar: Archive contains obsolescent base-64 headers
tar: Error exit delayed from previous error

Now this is not ok. Then I do:

diff ~/download/linux-2.6.20.2.tar.bz2 \
     mysqlfs/linux-2.6.20.2.tar.bz2

Binary files differ.

How is this possible?

Is this a new error or....?

Stef Bon





----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 03:07

Message:
The bug is a bit old, but I hope the original poster comes back with
details.  No response in a month (2009-07-28), let's consider closing this,
although I'd prefer the original poster coming back, re-trying, falling
back in love with mysqlfs, and disappearing into the sunset, a happy
ending.  Maybe a postcard.

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 03:04

Message:
I just added revision 48, which includes a testcase that mounts the
filesystem and copies a file onto it, then compares the checksum.  Tested
with linux-2.6.30.tar.bz2, checksum was a match.

I'm shifting this to a "works for me", but I hope you're willing to come
back and offer more details, or check out rev 48 of the SVN and try it: 
./configure --with-testfile=linux-2.6.20.2.tar.bz2 && make check

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1681567&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-2096705 ] syntax error or argument missing in query.c

[SourceForge.net <no...@so...>]

Bugs item #2096705, was opened at 2008-09-06 09:24
Message generated for change (Settings changed) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=2096705&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: Works For Me
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
>Assigned to: chickenandporn (chickenandporn)
Summary: syntax error or argument missing in query.c

Initial Comment:
on compiling query.c, the compiler crashes because is waiting for the string "%lld" in a sprintf command, and after the string only 2 arguments are given (a long-typed one and an integer), so the format string is mistyped or an argument is missing.

lines 235, 593 and 597 in file query.c

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 18:58

Message:
I'm marking this as "works for me" -- because it does -- but I'm looking
for your response where it doesn't "work for you".  Maybe we need to
augment the project for portability.

I'd like to close this after 2009-07-31 if we don't see a response.  I
don't mean to be draconian, I'd like to resolve this if we can still get
feedback.

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 18:49

Message:
Just checking this, line 235 involves off_t length which is indirectly
defined as a __int64_t on my platform.  line 593 is similar.  The only
dodgy one might be line 597: length + size + 1 -- where size is a "size_t",
indirectly defined as an unsigned long.  The compiler should convert the
unsigned long to a long long for the math, it should be fine.

What's your platform?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=2096705&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1849644 ] configure pthread test failing

[SourceForge.net <no...@so...>]

Bugs item #1849644, was opened at 2007-12-13 10:35
Message generated for change (Comment added) made by ludvigm
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
>Assigned to: chickenandporn (chickenandporn)
Summary: configure pthread test failing

Initial Comment:
Trying to compile of fc7 or fc7 I get this output:

checking for pthread_create in -lpthread... no
configure: error: Please install pthreads library first.

But it is installed with glibc-headers-2.6-4 and glibc-devel.

I am unable to compile this on a modern Fedora distro!

respond to: da...@ge...

----------------------------------------------------------------------

>Comment By: Michal Ludvig (ludvigm)
Date: 2009-06-28 16:25

Message:
Sure, commit whatever you like. I don't require you to ask for approval
before every check-in. Now it's your project ;-)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 13:31

Message:
OK.  Testing with a mysql-devel but missing fuse-devel shows that the LDADD
and such are being given a "-L${FUSE_LIB_DIR}", but with no fuse, that
means that LDADD += "-L ", which seems to break the later test builds. 
I've protected those behind a "if test x${FUSE_LIB_DIR} != x".  Done the
same for the MySQL stuff, since it also uses its own fallback to find MySQL
if the user hasn't give one.

I also notice that letting it run normally gets a "... -l mysqlclient_r
... -l mysqlclient" which seems redundant.  AC_SEARCH_LIBS allows it to
check without adding -lmysqlclient first, which might mislead the user but
accurately avoids duplicate mysqlclient libs.

Index: configure.in
===================================================================
--- configure.in	(revision 47)
+++ configure.in	(working copy)
@@ -48,11 +48,14 @@
     break
   fi
 done
+if test x${FUSE_LIB_DIR} != x
+then
 FUSE_LDFLAGS="-L${FUSE_LIB_DIR}"
 FUSE_CFLAGS="-I${FUSE_INC_DIR}"
 LDFLAGS="$LDFLAGS $FUSE_LDFLAGS"
 CFLAGS="$CFLAGS $FUSE_CFLAGS"
 CPPFLAGS="$CPPFLAGS $FUSE_CFLAGS"
+fi
 
 MYSQL_CONFIG=$(which mysql_config)
 if test "$MYSQL_CONFIG"; then
@@ -72,17 +75,20 @@
       break
     fi
   done
+  if test x${MYSQL_LIB_DIR} != x
+  then
   MYSQL_LDFLAGS="-L${MYSQL_LIB_DIR}"
   MYSQL_CFLAGS="-I${MYSQL_INC_DIR}"
 fi
+fi
 LDFLAGS="$LDFLAGS $MYSQL_LDFLAGS"
 CFLAGS="$CFLAGS $CFLAGS_ADD $MYSQL_CFLAGS"
 CPPFLAGS="$CPPFLAGS $MYSQL_CFLAGS"
 
 dnl Checks for libraries.
-AC_CHECK_LIB(pthread, pthread_create,, AC_MSG_ERROR([Please install
pthreads library first.]))
-AC_CHECK_LIB(mysqlclient, mysql_init,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
-AC_CHECK_LIB(fuse, fuse_main,, AC_MSG_ERROR([Please install fuse library
first.]))
+AC_SEARCH_LIBS(mysql_init, mysqlclient,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
+AC_SEARCH_LIBS(pthread_create, pthread,, AC_MSG_ERROR([Please install
pthreads library first.]))
+AC_SEARCH_LIBS(fuse_main, fuse,, AC_MSG_ERROR([Please install fuse
library first.]))
 
 dnl Checks for header files. (mac -- and BSD? -- have statfs in mount.h)
 AC_CHECK_HEADERS(stdio.h sys/param.h sys/mount.h)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 13:07

Message:
The error seems to be that "mysql_config" is part of "mysql", not part of
"mysql-devel", so the configure finds a "mysql_config" and interrogates
it:

MYSQL_CONFIG=$(which mysql_config)
if test "$MYSQL_CONFIG"; then
  MYSQL_LDFLAGS=$(${MYSQL_CONFIG} --libs_r)
...
...
LDFLAGS="$LDFLAGS $MYSQL_LDFLAGS"


Problem is, that generates a LDFLAGS that doesn't actually work because
the linker cannot find -lmysqlclient_r (which is part of mysql-devel).  It
just so happens that the first compilation is pthreads, not mysql.

I would recommend swapping the order.

At the same time, I would recommend moving the pthreads check to
AC_SEARCH_LIBS rather than AC_CHECK_LIB so that OSes with it already
without a -lpthreads are properly detected as "happy without -lpthreads"
(or the EDG variant: -K thread )

Whaddaya say to that?

--- configure.in	(revision 47)
+++ configure.in	(working copy)
@@ -80,8 +80,8 @@
 CPPFLAGS="$CPPFLAGS $MYSQL_CFLAGS"
 
 dnl Checks for libraries.
-AC_CHECK_LIB(pthread, pthread_create,, AC_MSG_ERROR([Please install
pthreads library first.]))
 AC_CHECK_LIB(mysqlclient, mysql_init,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
+AC_SEARCH_LIBS(pthread_create, pthread,, AC_MSG_ERROR([Please install
pthreads library first.]))
 AC_CHECK_LIB(fuse, fuse_main,, AC_MSG_ERROR([Please install fuse library
first.]))

...
checking for mysql_init in -lmysqlclient... no
configure: error: Please install mysqlclient library first.
make: *** [config.status] Error 1



----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-06-28 12:48

Message:
Hi Allan, simply uninstall mysql-devel and fuse-devel and you should be
able to trigger this error with ./configure. IIRC there's something wrong
with CFLAGS or LDFLAGS handling - when mysql-devel is missing it makes
CFLAGS invalid which in turn breaks pthread test. It's indeed my fault but
I never got around to fix it. Feel free to do so ;-)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 05:57

Message:
Hey Dan, if you are willing to get back on the diagnostic side, can you
attach a config.log below, by expanding the "Attached file" and uploading a
config.log from your dev environment?

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 04:07

Message:
Hi Dan; I'd like to address your issue in a next release; I'm a bit new
here as well.  Can I send you another tarball to try?  As well, I don't
expect this to magically resolve the issue, are you willing to banter a bit
to resolve the issue on your platform?

----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-03-03 14:16

Message:
IIRC the message is misleading - in fact it means you don't have
mysql-devel (or libmysql-dev or how is it called on debian) package
installed. You'll also need a development package for fuse (fuse-devel or
libfuse-dev).

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-03-03 13:12

Message:
I have the same problem, "configure: error: Please install pthreads library
first."
if i do a locate pthread i get
"/lib/libpthread-2.7.so
/lib/libpthread.so.0
/lib/tls/i686/cmov/libpthread-2.7.so
/lib/tls/i686/cmov/libpthread.so.0
/usr/NX/lib/perl/include/bits/pthreadtypes.ph
/usr/include/pthread.h
/usr/include/bits/pthreadtypes.h
/usr/lib/libgpgme-pthread.so.11
/usr/lib/libgpgme-pthread.so.11.6.4
/usr/lib/libpthread.a
/usr/lib/libpthread.so
/usr/lib/libpthread_nonshared.a
/usr/lib/perl/5.8.8/bits/pthreadtypes.ph
/usr/lib/xen/libpthread.a
/usr/lib/xen/libpthread_nonshared.a
/usr/share/man/man7/pthreads.7.gz
"

Am I doing something wrong? I'm using ubuntu 8.10 and i have installed
allmost anything with "thread" as package name..
any help?

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-12-13 11:01

Message:
Logged In: NO 

I meant fc6 or fc7, 32bit or 64bit

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1681567 ] After copy the file is changed.

[SourceForge.net <no...@so...>]

Bugs item #1681567, was opened at 2007-03-15 16:23
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1681567&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: Works For Me
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: After copy the file is changed.

Initial Comment:
Hello,

I'm working with mysqlfs, and very interested. It could be a backup-back-end. 

But now I run into some problems:

after I mount it (with root privileges!) I've got a mountpoint:

mysqlfs on /tmp/mysqlfs type fuse (rw,nosuid,nodev)

Now I copy a file to it:

cp ~/download/linux-2.6.20.2.tar.bz2 /tmp/mysqlfs

this goes fine. But now I want to unpack it:
cd /tmp
tar -xvf mysqlfs/linux-2.6.20.2.tar.bz2 

I get the error:

tar: This does not look like a tar archive
tar: Skipping to the next header
tar: Archive contains obsolescent base-64 headers
tar: Error exit delayed from previous error

Now this is not ok. Then I do:

diff ~/download/linux-2.6.20.2.tar.bz2 \
     mysqlfs/linux-2.6.20.2.tar.bz2

Binary files differ.

How is this possible?

Is this a new error or....?

Stef Bon





----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 03:07

Message:
The bug is a bit old, but I hope the original poster comes back with
details.  No response in a month (2009-07-28), let's consider closing this,
although I'd prefer the original poster coming back, re-trying, falling
back in love with mysqlfs, and disappearing into the sunset, a happy
ending.  Maybe a postcard.

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 03:04

Message:
I just added revision 48, which includes a testcase that mounts the
filesystem and copies a file onto it, then compares the checksum.  Tested
with linux-2.6.30.tar.bz2, checksum was a match.

I'm shifting this to a "works for me", but I hope you're willing to come
back and offer more details, or check out rev 48 of the SVN and try it: 
./configure --with-testfile=linux-2.6.20.2.tar.bz2 && make check

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1681567&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1681567 ] After copy the file is changed.

[SourceForge.net <no...@so...>]

Bugs item #1681567, was opened at 2007-03-15 16:23
Message generated for change (Settings changed) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1681567&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
>Resolution: Works For Me
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: After copy the file is changed.

Initial Comment:
Hello,

I'm working with mysqlfs, and very interested. It could be a backup-back-end. 

But now I run into some problems:

after I mount it (with root privileges!) I've got a mountpoint:

mysqlfs on /tmp/mysqlfs type fuse (rw,nosuid,nodev)

Now I copy a file to it:

cp ~/download/linux-2.6.20.2.tar.bz2 /tmp/mysqlfs

this goes fine. But now I want to unpack it:
cd /tmp
tar -xvf mysqlfs/linux-2.6.20.2.tar.bz2 

I get the error:

tar: This does not look like a tar archive
tar: Skipping to the next header
tar: Archive contains obsolescent base-64 headers
tar: Error exit delayed from previous error

Now this is not ok. Then I do:

diff ~/download/linux-2.6.20.2.tar.bz2 \
     mysqlfs/linux-2.6.20.2.tar.bz2

Binary files differ.

How is this possible?

Is this a new error or....?

Stef Bon





----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 03:04

Message:
I just added revision 48, which includes a testcase that mounts the
filesystem and copies a file onto it, then compares the checksum.  Tested
with linux-2.6.30.tar.bz2, checksum was a match.

I'm shifting this to a "works for me", but I hope you're willing to come
back and offer more details, or check out rev 48 of the SVN and try it: 
./configure --with-testfile=linux-2.6.20.2.tar.bz2 && make check

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1681567&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1681567 ] After copy the file is changed.

[SourceForge.net <no...@so...>]

Bugs item #1681567, was opened at 2007-03-15 16:23
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1681567&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: After copy the file is changed.

Initial Comment:
Hello,

I'm working with mysqlfs, and very interested. It could be a backup-back-end. 

But now I run into some problems:

after I mount it (with root privileges!) I've got a mountpoint:

mysqlfs on /tmp/mysqlfs type fuse (rw,nosuid,nodev)

Now I copy a file to it:

cp ~/download/linux-2.6.20.2.tar.bz2 /tmp/mysqlfs

this goes fine. But now I want to unpack it:
cd /tmp
tar -xvf mysqlfs/linux-2.6.20.2.tar.bz2 

I get the error:

tar: This does not look like a tar archive
tar: Skipping to the next header
tar: Archive contains obsolescent base-64 headers
tar: Error exit delayed from previous error

Now this is not ok. Then I do:

diff ~/download/linux-2.6.20.2.tar.bz2 \
     mysqlfs/linux-2.6.20.2.tar.bz2

Binary files differ.

How is this possible?

Is this a new error or....?

Stef Bon





----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 03:04

Message:
I just added revision 48, which includes a testcase that mounts the
filesystem and copies a file onto it, then compares the checksum.  Tested
with linux-2.6.30.tar.bz2, checksum was a match.

I'm shifting this to a "works for me", but I hope you're willing to come
back and offer more details, or check out rev 48 of the SVN and try it: 
./configure --with-testfile=linux-2.6.20.2.tar.bz2 && make check

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1681567&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1849644 ] configure pthread test failing

[SourceForge.net <no...@so...>]

Bugs item #1849644, was opened at 2007-12-12 21:35
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: configure pthread test failing

Initial Comment:
Trying to compile of fc7 or fc7 I get this output:

checking for pthread_create in -lpthread... no
configure: error: Please install pthreads library first.

But it is installed with glibc-headers-2.6-4 and glibc-devel.

I am unable to compile this on a modern Fedora distro!

respond to: da...@ge...

----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 01:31

Message:
OK.  Testing with a mysql-devel but missing fuse-devel shows that the LDADD
and such are being given a "-L${FUSE_LIB_DIR}", but with no fuse, that
means that LDADD += "-L ", which seems to break the later test builds. 
I've protected those behind a "if test x${FUSE_LIB_DIR} != x".  Done the
same for the MySQL stuff, since it also uses its own fallback to find MySQL
if the user hasn't give one.

I also notice that letting it run normally gets a "... -l mysqlclient_r
... -l mysqlclient" which seems redundant.  AC_SEARCH_LIBS allows it to
check without adding -lmysqlclient first, which might mislead the user but
accurately avoids duplicate mysqlclient libs.

Index: configure.in
===================================================================
--- configure.in	(revision 47)
+++ configure.in	(working copy)
@@ -48,11 +48,14 @@
     break
   fi
 done
+if test x${FUSE_LIB_DIR} != x
+then
 FUSE_LDFLAGS="-L${FUSE_LIB_DIR}"
 FUSE_CFLAGS="-I${FUSE_INC_DIR}"
 LDFLAGS="$LDFLAGS $FUSE_LDFLAGS"
 CFLAGS="$CFLAGS $FUSE_CFLAGS"
 CPPFLAGS="$CPPFLAGS $FUSE_CFLAGS"
+fi
 
 MYSQL_CONFIG=$(which mysql_config)
 if test "$MYSQL_CONFIG"; then
@@ -72,17 +75,20 @@
       break
     fi
   done
+  if test x${MYSQL_LIB_DIR} != x
+  then
   MYSQL_LDFLAGS="-L${MYSQL_LIB_DIR}"
   MYSQL_CFLAGS="-I${MYSQL_INC_DIR}"
 fi
+fi
 LDFLAGS="$LDFLAGS $MYSQL_LDFLAGS"
 CFLAGS="$CFLAGS $CFLAGS_ADD $MYSQL_CFLAGS"
 CPPFLAGS="$CPPFLAGS $MYSQL_CFLAGS"
 
 dnl Checks for libraries.
-AC_CHECK_LIB(pthread, pthread_create,, AC_MSG_ERROR([Please install
pthreads library first.]))
-AC_CHECK_LIB(mysqlclient, mysql_init,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
-AC_CHECK_LIB(fuse, fuse_main,, AC_MSG_ERROR([Please install fuse library
first.]))
+AC_SEARCH_LIBS(mysql_init, mysqlclient,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
+AC_SEARCH_LIBS(pthread_create, pthread,, AC_MSG_ERROR([Please install
pthreads library first.]))
+AC_SEARCH_LIBS(fuse_main, fuse,, AC_MSG_ERROR([Please install fuse
library first.]))
 
 dnl Checks for header files. (mac -- and BSD? -- have statfs in mount.h)
 AC_CHECK_HEADERS(stdio.h sys/param.h sys/mount.h)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 01:07

Message:
The error seems to be that "mysql_config" is part of "mysql", not part of
"mysql-devel", so the configure finds a "mysql_config" and interrogates
it:

MYSQL_CONFIG=$(which mysql_config)
if test "$MYSQL_CONFIG"; then
  MYSQL_LDFLAGS=$(${MYSQL_CONFIG} --libs_r)
...
...
LDFLAGS="$LDFLAGS $MYSQL_LDFLAGS"


Problem is, that generates a LDFLAGS that doesn't actually work because
the linker cannot find -lmysqlclient_r (which is part of mysql-devel).  It
just so happens that the first compilation is pthreads, not mysql.

I would recommend swapping the order.

At the same time, I would recommend moving the pthreads check to
AC_SEARCH_LIBS rather than AC_CHECK_LIB so that OSes with it already
without a -lpthreads are properly detected as "happy without -lpthreads"
(or the EDG variant: -K thread )

Whaddaya say to that?

--- configure.in	(revision 47)
+++ configure.in	(working copy)
@@ -80,8 +80,8 @@
 CPPFLAGS="$CPPFLAGS $MYSQL_CFLAGS"
 
 dnl Checks for libraries.
-AC_CHECK_LIB(pthread, pthread_create,, AC_MSG_ERROR([Please install
pthreads library first.]))
 AC_CHECK_LIB(mysqlclient, mysql_init,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
+AC_SEARCH_LIBS(pthread_create, pthread,, AC_MSG_ERROR([Please install
pthreads library first.]))
 AC_CHECK_LIB(fuse, fuse_main,, AC_MSG_ERROR([Please install fuse library
first.]))

...
checking for mysql_init in -lmysqlclient... no
configure: error: Please install mysqlclient library first.
make: *** [config.status] Error 1



----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-06-28 00:48

Message:
Hi Allan, simply uninstall mysql-devel and fuse-devel and you should be
able to trigger this error with ./configure. IIRC there's something wrong
with CFLAGS or LDFLAGS handling - when mysql-devel is missing it makes
CFLAGS invalid which in turn breaks pthread test. It's indeed my fault but
I never got around to fix it. Feel free to do so ;-)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 17:57

Message:
Hey Dan, if you are willing to get back on the diagnostic side, can you
attach a config.log below, by expanding the "Attached file" and uploading a
config.log from your dev environment?

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 16:07

Message:
Hi Dan; I'd like to address your issue in a next release; I'm a bit new
here as well.  Can I send you another tarball to try?  As well, I don't
expect this to magically resolve the issue, are you willing to banter a bit
to resolve the issue on your platform?

----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-03-03 01:16

Message:
IIRC the message is misleading - in fact it means you don't have
mysql-devel (or libmysql-dev or how is it called on debian) package
installed. You'll also need a development package for fuse (fuse-devel or
libfuse-dev).

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-03-03 00:12

Message:
I have the same problem, "configure: error: Please install pthreads library
first."
if i do a locate pthread i get
"/lib/libpthread-2.7.so
/lib/libpthread.so.0
/lib/tls/i686/cmov/libpthread-2.7.so
/lib/tls/i686/cmov/libpthread.so.0
/usr/NX/lib/perl/include/bits/pthreadtypes.ph
/usr/include/pthread.h
/usr/include/bits/pthreadtypes.h
/usr/lib/libgpgme-pthread.so.11
/usr/lib/libgpgme-pthread.so.11.6.4
/usr/lib/libpthread.a
/usr/lib/libpthread.so
/usr/lib/libpthread_nonshared.a
/usr/lib/perl/5.8.8/bits/pthreadtypes.ph
/usr/lib/xen/libpthread.a
/usr/lib/xen/libpthread_nonshared.a
/usr/share/man/man7/pthreads.7.gz
"

Am I doing something wrong? I'm using ubuntu 8.10 and i have installed
allmost anything with "thread" as package name..
any help?

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-12-12 22:01

Message:
Logged In: NO 

I meant fc6 or fc7, 32bit or 64bit

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1849644 ] configure pthread test failing

[SourceForge.net <no...@so...>]

Bugs item #1849644, was opened at 2007-12-12 21:35
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: configure pthread test failing

Initial Comment:
Trying to compile of fc7 or fc7 I get this output:

checking for pthread_create in -lpthread... no
configure: error: Please install pthreads library first.

But it is installed with glibc-headers-2.6-4 and glibc-devel.

I am unable to compile this on a modern Fedora distro!

respond to: da...@ge...

----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 01:07

Message:
The error seems to be that "mysql_config" is part of "mysql", not part of
"mysql-devel", so the configure finds a "mysql_config" and interrogates
it:

MYSQL_CONFIG=$(which mysql_config)
if test "$MYSQL_CONFIG"; then
  MYSQL_LDFLAGS=$(${MYSQL_CONFIG} --libs_r)
...
...
LDFLAGS="$LDFLAGS $MYSQL_LDFLAGS"


Problem is, that generates a LDFLAGS that doesn't actually work because
the linker cannot find -lmysqlclient_r (which is part of mysql-devel).  It
just so happens that the first compilation is pthreads, not mysql.

I would recommend swapping the order.

At the same time, I would recommend moving the pthreads check to
AC_SEARCH_LIBS rather than AC_CHECK_LIB so that OSes with it already
without a -lpthreads are properly detected as "happy without -lpthreads"
(or the EDG variant: -K thread )

Whaddaya say to that?

--- configure.in	(revision 47)
+++ configure.in	(working copy)
@@ -80,8 +80,8 @@
 CPPFLAGS="$CPPFLAGS $MYSQL_CFLAGS"
 
 dnl Checks for libraries.
-AC_CHECK_LIB(pthread, pthread_create,, AC_MSG_ERROR([Please install
pthreads library first.]))
 AC_CHECK_LIB(mysqlclient, mysql_init,, AC_MSG_ERROR([Please install
mysqlclient library first.]))
+AC_SEARCH_LIBS(pthread_create, pthread,, AC_MSG_ERROR([Please install
pthreads library first.]))
 AC_CHECK_LIB(fuse, fuse_main,, AC_MSG_ERROR([Please install fuse library
first.]))

...
checking for mysql_init in -lmysqlclient... no
configure: error: Please install mysqlclient library first.
make: *** [config.status] Error 1



----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-06-28 00:48

Message:
Hi Allan, simply uninstall mysql-devel and fuse-devel and you should be
able to trigger this error with ./configure. IIRC there's something wrong
with CFLAGS or LDFLAGS handling - when mysql-devel is missing it makes
CFLAGS invalid which in turn breaks pthread test. It's indeed my fault but
I never got around to fix it. Feel free to do so ;-)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 17:57

Message:
Hey Dan, if you are willing to get back on the diagnostic side, can you
attach a config.log below, by expanding the "Attached file" and uploading a
config.log from your dev environment?

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 16:07

Message:
Hi Dan; I'd like to address your issue in a next release; I'm a bit new
here as well.  Can I send you another tarball to try?  As well, I don't
expect this to magically resolve the issue, are you willing to banter a bit
to resolve the issue on your platform?

----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-03-03 01:16

Message:
IIRC the message is misleading - in fact it means you don't have
mysql-devel (or libmysql-dev or how is it called on debian) package
installed. You'll also need a development package for fuse (fuse-devel or
libfuse-dev).

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-03-03 00:12

Message:
I have the same problem, "configure: error: Please install pthreads library
first."
if i do a locate pthread i get
"/lib/libpthread-2.7.so
/lib/libpthread.so.0
/lib/tls/i686/cmov/libpthread-2.7.so
/lib/tls/i686/cmov/libpthread.so.0
/usr/NX/lib/perl/include/bits/pthreadtypes.ph
/usr/include/pthread.h
/usr/include/bits/pthreadtypes.h
/usr/lib/libgpgme-pthread.so.11
/usr/lib/libgpgme-pthread.so.11.6.4
/usr/lib/libpthread.a
/usr/lib/libpthread.so
/usr/lib/libpthread_nonshared.a
/usr/lib/perl/5.8.8/bits/pthreadtypes.ph
/usr/lib/xen/libpthread.a
/usr/lib/xen/libpthread_nonshared.a
/usr/share/man/man7/pthreads.7.gz
"

Am I doing something wrong? I'm using ubuntu 8.10 and i have installed
allmost anything with "thread" as package name..
any help?

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-12-12 22:01

Message:
Logged In: NO 

I meant fc6 or fc7, 32bit or 64bit

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1849644 ] configure pthread test failing

[SourceForge.net <no...@so...>]

Bugs item #1849644, was opened at 2007-12-13 10:35
Message generated for change (Comment added) made by ludvigm
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: configure pthread test failing

Initial Comment:
Trying to compile of fc7 or fc7 I get this output:

checking for pthread_create in -lpthread... no
configure: error: Please install pthreads library first.

But it is installed with glibc-headers-2.6-4 and glibc-devel.

I am unable to compile this on a modern Fedora distro!

respond to: da...@ge...

----------------------------------------------------------------------

>Comment By: Michal Ludvig (ludvigm)
Date: 2009-06-28 12:48

Message:
Hi Allan, simply uninstall mysql-devel and fuse-devel and you should be
able to trigger this error with ./configure. IIRC there's something wrong
with CFLAGS or LDFLAGS handling - when mysql-devel is missing it makes
CFLAGS invalid which in turn breaks pthread test. It's indeed my fault but
I never got around to fix it. Feel free to do so ;-)

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 05:57

Message:
Hey Dan, if you are willing to get back on the diagnostic side, can you
attach a config.log below, by expanding the "Attached file" and uploading a
config.log from your dev environment?

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-28 04:07

Message:
Hi Dan; I'd like to address your issue in a next release; I'm a bit new
here as well.  Can I send you another tarball to try?  As well, I don't
expect this to magically resolve the issue, are you willing to banter a bit
to resolve the issue on your platform?

----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-03-03 14:16

Message:
IIRC the message is misleading - in fact it means you don't have
mysql-devel (or libmysql-dev or how is it called on debian) package
installed. You'll also need a development package for fuse (fuse-devel or
libfuse-dev).

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-03-03 13:12

Message:
I have the same problem, "configure: error: Please install pthreads library
first."
if i do a locate pthread i get
"/lib/libpthread-2.7.so
/lib/libpthread.so.0
/lib/tls/i686/cmov/libpthread-2.7.so
/lib/tls/i686/cmov/libpthread.so.0
/usr/NX/lib/perl/include/bits/pthreadtypes.ph
/usr/include/pthread.h
/usr/include/bits/pthreadtypes.h
/usr/lib/libgpgme-pthread.so.11
/usr/lib/libgpgme-pthread.so.11.6.4
/usr/lib/libpthread.a
/usr/lib/libpthread.so
/usr/lib/libpthread_nonshared.a
/usr/lib/perl/5.8.8/bits/pthreadtypes.ph
/usr/lib/xen/libpthread.a
/usr/lib/xen/libpthread_nonshared.a
/usr/share/man/man7/pthreads.7.gz
"

Am I doing something wrong? I'm using ubuntu 8.10 and i have installed
allmost anything with "thread" as package name..
any help?

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-12-13 11:01

Message:
Logged In: NO 

I meant fc6 or fc7, 32bit or 64bit

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-2096705 ] syntax error or argument missing in query.c

[SourceForge.net <no...@so...>]

Bugs item #2096705, was opened at 2008-09-06 09:24
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=2096705&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
>Resolution: Works For Me
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: syntax error or argument missing in query.c

Initial Comment:
on compiling query.c, the compiler crashes because is waiting for the string "%lld" in a sprintf command, and after the string only 2 arguments are given (a long-typed one and an integer), so the format string is mistyped or an argument is missing.

lines 235, 593 and 597 in file query.c

----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 18:58

Message:
I'm marking this as "works for me" -- because it does -- but I'm looking
for your response where it doesn't "work for you".  Maybe we need to
augment the project for portability.

I'd like to close this after 2009-07-31 if we don't see a response.  I
don't mean to be draconian, I'd like to resolve this if we can still get
feedback.

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 18:49

Message:
Just checking this, line 235 involves off_t length which is indirectly
defined as a __int64_t on my platform.  line 593 is similar.  The only
dodgy one might be line 597: length + size + 1 -- where size is a "size_t",
indirectly defined as an unsigned long.  The compiler should convert the
unsigned long to a long long for the math, it should be fine.

What's your platform?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=2096705&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-2096705 ] syntax error or argument missing in query.c

[SourceForge.net <no...@so...>]

Bugs item #2096705, was opened at 2008-09-06 09:24
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=2096705&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: syntax error or argument missing in query.c

Initial Comment:
on compiling query.c, the compiler crashes because is waiting for the string "%lld" in a sprintf command, and after the string only 2 arguments are given (a long-typed one and an integer), so the format string is mistyped or an argument is missing.

lines 235, 593 and 597 in file query.c

----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 18:49

Message:
Just checking this, line 235 involves off_t length which is indirectly
defined as a __int64_t on my platform.  line 593 is similar.  The only
dodgy one might be line 597: length + size + 1 -- where size is a "size_t",
indirectly defined as an unsigned long.  The compiler should convert the
unsigned long to a long long for the math, it should be fine.

What's your platform?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=2096705&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1849644 ] configure pthread test failing

[SourceForge.net <no...@so...>]

Bugs item #1849644, was opened at 2007-12-12 21:35
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: configure pthread test failing

Initial Comment:
Trying to compile of fc7 or fc7 I get this output:

checking for pthread_create in -lpthread... no
configure: error: Please install pthreads library first.

But it is installed with glibc-headers-2.6-4 and glibc-devel.

I am unable to compile this on a modern Fedora distro!

respond to: da...@ge...

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 17:57

Message:
Hey Dan, if you are willing to get back on the diagnostic side, can you
attach a config.log below, by expanding the "Attached file" and uploading a
config.log from your dev environment?

----------------------------------------------------------------------

Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 16:07

Message:
Hi Dan; I'd like to address your issue in a next release; I'm a bit new
here as well.  Can I send you another tarball to try?  As well, I don't
expect this to magically resolve the issue, are you willing to banter a bit
to resolve the issue on your platform?

----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-03-03 01:16

Message:
IIRC the message is misleading - in fact it means you don't have
mysql-devel (or libmysql-dev or how is it called on debian) package
installed. You'll also need a development package for fuse (fuse-devel or
libfuse-dev).

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-03-03 00:12

Message:
I have the same problem, "configure: error: Please install pthreads library
first."
if i do a locate pthread i get
"/lib/libpthread-2.7.so
/lib/libpthread.so.0
/lib/tls/i686/cmov/libpthread-2.7.so
/lib/tls/i686/cmov/libpthread.so.0
/usr/NX/lib/perl/include/bits/pthreadtypes.ph
/usr/include/pthread.h
/usr/include/bits/pthreadtypes.h
/usr/lib/libgpgme-pthread.so.11
/usr/lib/libgpgme-pthread.so.11.6.4
/usr/lib/libpthread.a
/usr/lib/libpthread.so
/usr/lib/libpthread_nonshared.a
/usr/lib/perl/5.8.8/bits/pthreadtypes.ph
/usr/lib/xen/libpthread.a
/usr/lib/xen/libpthread_nonshared.a
/usr/share/man/man7/pthreads.7.gz
"

Am I doing something wrong? I'm using ubuntu 8.10 and i have installed
allmost anything with "thread" as package name..
any help?

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-12-12 22:01

Message:
Logged In: NO 

I meant fc6 or fc7, 32bit or 64bit

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1849644 ] configure pthread test failing

[SourceForge.net <no...@so...>]

Bugs item #1849644, was opened at 2007-12-12 21:35
Message generated for change (Comment added) made by chickenandporn
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: configure pthread test failing

Initial Comment:
Trying to compile of fc7 or fc7 I get this output:

checking for pthread_create in -lpthread... no
configure: error: Please install pthreads library first.

But it is installed with glibc-headers-2.6-4 and glibc-devel.

I am unable to compile this on a modern Fedora distro!

respond to: da...@ge...

----------------------------------------------------------------------

>Comment By: chickenandporn (chickenandporn)
Date: 2009-06-27 16:07

Message:
Hi Dan; I'd like to address your issue in a next release; I'm a bit new
here as well.  Can I send you another tarball to try?  As well, I don't
expect this to magically resolve the issue, are you willing to banter a bit
to resolve the issue on your platform?

----------------------------------------------------------------------

Comment By: Michal Ludvig (ludvigm)
Date: 2009-03-03 01:16

Message:
IIRC the message is misleading - in fact it means you don't have
mysql-devel (or libmysql-dev or how is it called on debian) package
installed. You'll also need a development package for fuse (fuse-devel or
libfuse-dev).

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-03-03 00:12

Message:
I have the same problem, "configure: error: Please install pthreads library
first."
if i do a locate pthread i get
"/lib/libpthread-2.7.so
/lib/libpthread.so.0
/lib/tls/i686/cmov/libpthread-2.7.so
/lib/tls/i686/cmov/libpthread.so.0
/usr/NX/lib/perl/include/bits/pthreadtypes.ph
/usr/include/pthread.h
/usr/include/bits/pthreadtypes.h
/usr/lib/libgpgme-pthread.so.11
/usr/lib/libgpgme-pthread.so.11.6.4
/usr/lib/libpthread.a
/usr/lib/libpthread.so
/usr/lib/libpthread_nonshared.a
/usr/lib/perl/5.8.8/bits/pthreadtypes.ph
/usr/lib/xen/libpthread.a
/usr/lib/xen/libpthread_nonshared.a
/usr/share/man/man7/pthreads.7.gz
"

Am I doing something wrong? I'm using ubuntu 8.10 and i have installed
allmost anything with "thread" as package name..
any help?

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-12-12 22:01

Message:
Logged In: NO 

I meant fc6 or fc7, 32bit or 64bit

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

[Mysqlfs-commits] [ mysqlfs-Bugs-1849644 ] configure pthread test failing

[SourceForge.net <no...@so...>]

Bugs item #1849644, was opened at 2007-12-13 10:35
Message generated for change (Comment added) made by ludvigm
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: configure pthread test failing

Initial Comment:
Trying to compile of fc7 or fc7 I get this output:

checking for pthread_create in -lpthread... no
configure: error: Please install pthreads library first.

But it is installed with glibc-headers-2.6-4 and glibc-devel.

I am unable to compile this on a modern Fedora distro!

respond to: da...@ge...

----------------------------------------------------------------------

>Comment By: Michal Ludvig (ludvigm)
Date: 2009-03-03 14:16

Message:
IIRC the message is misleading - in fact it means you don't have
mysql-devel (or libmysql-dev or how is it called on debian) package
installed. You'll also need a development package for fuse (fuse-devel or
libfuse-dev).

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2009-03-03 13:12

Message:
I have the same problem, "configure: error: Please install pthreads library
first."
if i do a locate pthread i get
"/lib/libpthread-2.7.so
/lib/libpthread.so.0
/lib/tls/i686/cmov/libpthread-2.7.so
/lib/tls/i686/cmov/libpthread.so.0
/usr/NX/lib/perl/include/bits/pthreadtypes.ph
/usr/include/pthread.h
/usr/include/bits/pthreadtypes.h
/usr/lib/libgpgme-pthread.so.11
/usr/lib/libgpgme-pthread.so.11.6.4
/usr/lib/libpthread.a
/usr/lib/libpthread.so
/usr/lib/libpthread_nonshared.a
/usr/lib/perl/5.8.8/bits/pthreadtypes.ph
/usr/lib/xen/libpthread.a
/usr/lib/xen/libpthread_nonshared.a
/usr/share/man/man7/pthreads.7.gz
"

Am I doing something wrong? I'm using ubuntu 8.10 and i have installed
allmost anything with "thread" as package name..
any help?

----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2007-12-13 11:01

Message:
Logged In: NO 

I meant fc6 or fc7, 32bit or 64bit

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=716425&aid=1849644&group_id=129981