Note: This is a public test instance of Red Hat Bugzilla. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback at bugzilla.redhat.com.
Bug 578484 - Please add poco to EPEL-5
Summary: Please add poco to EPEL-5
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Fedora
Classification: Fedora
Component: poco
Version: rawhide
Hardware: All
OS: Linux
low
medium
Target Milestone: ---
Assignee: Matěj Cepl
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On: 579763
Blocks: 578480
TreeView+ depends on / blocked
 
Reported: 2010-03-31 13:18 UTC by Matěj Cepl
Modified: 2018-04-11 11:45 UTC (History)
3 users (show)

Fixed In Version: poco-1.3.5-8.el5
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2010-04-14 09:31:56 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
current state of the patch (2.06 KB, patch)
2010-04-03 10:46 UTC, Matěj Cepl
no flags Details | Diff

Description Matěj Cepl 2010-03-31 13:18:50 UTC
Description of problem:
Hi, I am packaging spectrum Jabber gateway (bug 578480) which uses poco. I
would love to get spectrum to EPEL-5 as well, and so I would need poco in
EPEL-5 as well. Would you do it, or should I take over maintainership of poco
in EPEL-5?

Thank you

Comment 1 Matěj Cepl 2010-03-31 13:30:11 UTC
Silly me, of course poco already is at EPEL-5.

Well, actually ... http://cvs.fedoraproject.org/viewvc/rpms/poco/EL-5/ is there but it seems to be kind of empty. What's going on here?

Comment 2 Maxim Udushlivy 2010-03-31 19:13:13 UTC
Hallo!

A very simple thing is going on here: insufficient dependencies. First, POCO requires newer version of sqlite (last time I checked it was at 3.3.6 here which POCO refused to compile against); second, iodbc is missing.

iodbc may be added to EPEL, but sqlite has to be updated at the "RHEL" level.

Comment 3 Matěj Cepl 2010-04-02 10:43:27 UTC
(In reply to comment #2)
> second, iodbc is missing.

So just to make sure iodbc is not an issue http://koji.fedoraproject.org/koji/taskinfo?taskID=2090916 Now I have to think about that sqlite

Comment 4 Matěj Cepl 2010-04-02 11:43:51 UTC
OK, so just to document my (non-)progress:

** Compiling src/SessionImpl.cpp (debug, shared)
g++ -I/usr/include/libiodbc -Iinclude -I/usr/src/redhat/BUILD/poco-1.3.5-all/CppUnit/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Foundation/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/XML/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Util/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Net/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Crypto/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/NetSSL_OpenSSL/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/ODBC/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/MySQL/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Zip/include -Wall -Wno-sign-compare -DPOCO_BUILD_HOST=tikanga.ceplovi.cz -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m32 -march=i386 -mtune=generic -fasynchronous-unwind-tables -D_XOPEN_SOURCE=500 -D_REENTRANT -D_THREAD_SAFE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DSQLITE_THREADSAFE=1 -DSQLITE_DISABLE_LFS -DSQLITE_OMIT_UTF16 -DSQLITE_OMIT_PROGRESS_CALLBACK -DSQLITE_OMIT_COMPLETE -DSQLITE_OMIT_TCL_VARIABLE -DSQLITE_OMIT_DEPRECATED  -g -D_DEBUG -fPIC -c src/SessionImpl.cpp -o /usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite/obj/Linux/i686/debug_shared/SessionImpl.o
src/SessionImpl.cpp: In member function 'void Poco::Data::SQLite::SessionImpl::open()':
src/SessionImpl.cpp:118: error: 'sqlite3_extended_result_codes' was not declared in this scope
make[1]: *** [/usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite/obj/Linux/i686/debug_shared/SessionImpl.o] Error 1
make[1]: Leaving directory `/usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite'
make: *** [Data/SQLite-libexec] Error 2
error: Bad exit status from /var/tmp/rpm-tmp.28060 (%build)


RPM build errors:
    Bad exit status from /var/tmp/rpm-tmp.28060 (%build)
[matej@tikanga rpmbuild]$

Comment 5 Matěj Cepl 2010-04-02 11:44:27 UTC
(BTW, upstream should fix their autofoo to declare correctly which version of sqlite they want, but that's another point)

Comment 6 Michal Schmidt 2010-04-02 12:29:25 UTC
If unavailable "sqlite3_extended_result_codes" function is the only problem, then I think we're fine. From a quick look at the source poco does not seem to take the extended error codes into account anyway (ie. it does handle any of the codes SQLITE_IOERR_* specifically, it only knows the basic SQLITE_IOERR). I'd say you can safely comment out the call.

More info here:
http://www.sqlite.org/c3ref/c_ioerr_access.html

Comment 7 Michal Schmidt 2010-04-02 12:30:19 UTC
(In reply to comment #6)
> (ie. it does handle any of the codes SQLITE_IOERR_* specifically

Should have said: It does NOT handle any ...

Comment 8 Matěj Cepl 2010-04-02 12:42:06 UTC
[matej@tikanga poco-1.3.5-all]$ cat TODO_Fedora 
+ grep -ri sqlite3_extended_result_code .
./Data/SQLite/src/SessionImpl.cpp:	sqlite3_extended_result_codes(_pDB, 1);
+ grep -r SQLITE_IOERR .
./Data/SQLite/src/Utility.cpp:	case SQLITE_IOERR:
+ grep -r SQLITE_L .
./Data/SQLite/src/SQLiteStatementImpl.cpp:			// When we get SQLITE_LOCKED, we must reset the statement before trying
./Data/SQLite/src/SQLiteStatementImpl.cpp:			// again. SQLITE_LOCKED is only returned for the first call to sqlite3_step,
./Data/SQLite/src/SQLiteStatementImpl.cpp:		case SQLITE_LOCKED:
./Data/SQLite/src/SQLiteStatementImpl.cpp:		case SQLITE_LOCKED_SHAREDCACHE:
./Data/SQLite/src/Utility.cpp:	case SQLITE_LOCKED:
./Data/SQLite/src/Utility.cpp:	case SQLITE_LOCKED_SHAREDCACHE:
[matej@tikanga poco-1.3.5-all]$

Comment 9 Michal Schmidt 2010-04-02 12:56:03 UTC
Alright, SQLITE_LOCKED_SHAREDCACHE is also one of the extended error codes, but poco always handles it identically to SQLITE_LOCKED.

Comment 10 Matěj Cepl 2010-04-02 13:15:04 UTC
OK, so we have progressed:

** Compiling src/SQLiteStatementImpl.cpp (debug, shared)
g++ -I/usr/include/libiodbc -Iinclude -I/usr/src/redhat/BUILD/poco-1.3.5-all/CppUnit/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Foundation/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/XML/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Util/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Net/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Crypto/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/NetSSL_OpenSSL/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/ODBC/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/MySQL/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Zip/include -Wall -Wno-sign-compare -DPOCO_BUILD_HOST=tikanga.ceplovi.cz -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m32 -march=i386 -mtune=generic -fasynchronous-unwind-tables -D_XOPEN_SOURCE=500 -D_REENTRANT -D_THREAD_SAFE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DSQLITE_THREADSAFE=1 -DSQLITE_DISABLE_LFS -DSQLITE_OMIT_UTF16 -DSQLITE_OMIT_PROGRESS_CALLBACK -DSQLITE_OMIT_COMPLETE -DSQLITE_OMIT_TCL_VARIABLE -DSQLITE_OMIT_DEPRECATED  -g -D_DEBUG -fPIC -c src/SQLiteStatementImpl.cpp -o /usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite/obj/Linux/i686/debug_shared/SQLiteStatementImpl.o
src/SQLiteStatementImpl.cpp: In member function 'void Poco::Data::SQLite::SQLiteStatementImpl::compileImplImpl()':
src/SQLiteStatementImpl.cpp:106: error: 'sqlite3_prepare_v2' was not declared in this scope
src/SQLiteStatementImpl.cpp: In member function 'virtual void Poco::Data::SQLite::SQLiteStatementImpl::bindImpl()':
src/SQLiteStatementImpl.cpp:173: warning: comparison between signed and unsigned integer expressions
src/SQLiteStatementImpl.cpp: In member function 'virtual bool Poco::Data::SQLite::SQLiteStatementImpl::hasNext()':
src/SQLiteStatementImpl.cpp:225: error: 'SQLITE_LOCKED_SHAREDCACHE' was not declared in this scope
src/SQLiteStatementImpl.cpp: In member function 'virtual void Poco::Data::SQLite::SQLiteStatementImpl::next()':
src/SQLiteStatementImpl.cpp:254: warning: comparison between signed and unsigned integer expressions
make[1]: *** [/usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite/obj/Linux/i686/debug_shared/SQLiteStatementImpl.o] Error 1
make[1]: Leaving directory `/usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite'
make: *** [Data/SQLite-libexec] Error 2
error: Bad exit status from /var/tmp/rpm-tmp.47776 (%build)


RPM build errors:
    Bad exit status from /var/tmp/rpm-tmp.47776 (%build)
[matej@tikanga rpmbuild]$

Comment 11 Michal Schmidt 2010-04-02 13:56:04 UTC
sqlite3_prepare_v2 can be replaced by sqlite3_prepare, but more care is needed, because it changes the semantics of subsequent calls slightly.

Comment 12 Matěj Cepl 2010-04-03 10:46:55 UTC
Created attachment 404290 [details]
current state of the patch

** Compiling src/SQLiteStatementImpl.cpp (debug, shared)
g++ -I/usr/include/libiodbc -Iinclude -I/usr/src/redhat/BUILD/poco-1.3.5-all/CppUnit/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Foundation/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/XML/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Util/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Net/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Crypto/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/NetSSL_OpenSSL/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/ODBC/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Data/MySQL/include -I/usr/src/redhat/BUILD/poco-1.3.5-all/Zip/include -Wall -Wno-sign-compare -DPOCO_BUILD_HOST=tikanga.ceplovi.cz -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m32 -march=i386 -mtune=generic -fasynchronous-unwind-tables -D_XOPEN_SOURCE=500 -D_REENTRANT -D_THREAD_SAFE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DSQLITE_THREADSAFE=1 -DSQLITE_DISABLE_LFS -DSQLITE_OMIT_UTF16 -DSQLITE_OMIT_PROGRESS_CALLBACK -DSQLITE_OMIT_COMPLETE -DSQLITE_OMIT_TCL_VARIABLE -DSQLITE_OMIT_DEPRECATED  -g -D_DEBUG -fPIC -c src/SQLiteStatementImpl.cpp -o /usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite/obj/Linux/i686/debug_shared/SQLiteStatementImpl.o
src/SQLiteStatementImpl.cpp: In member function 'void Poco::Data::SQLite::SQLiteStatementImpl::compileImplImpl()':
src/SQLiteStatementImpl.cpp:104: error: 'SQLITE_CONFIG_MEMSTATUS' was not declared in this scope
src/SQLiteStatementImpl.cpp:104: error: 'sqlite3_config' was not declared in this scope
src/SQLiteStatementImpl.cpp: In member function 'virtual void Poco::Data::SQLite::SQLiteStatementImpl::bindImpl()':
src/SQLiteStatementImpl.cpp:175: warning: comparison between signed and unsigned integer expressions
src/SQLiteStatementImpl.cpp: In member function 'virtual void Poco::Data::SQLite::SQLiteStatementImpl::next()':
src/SQLiteStatementImpl.cpp:256: warning: comparison between signed and unsigned integer expressions
make[1]: *** [/usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite/obj/Linux/i686/debug_shared/SQLiteStatementImpl.o] Error 1
make[1]: Leaving directory `/usr/src/redhat/BUILD/poco-1.3.5-all/Data/SQLite'
make: *** [Data/SQLite-libexec] Error 2
error: Bad exit status from /var/tmp/rpm-tmp.42529 (%build)


RPM build errors:
    Bad exit status from /var/tmp/rpm-tmp.42529 (%build)
[matej@tikanga build]$

Comment 13 Matěj Cepl 2010-04-03 10:47:50 UTC
Taking the bug, no-one should be punished for this obsession.

Comment 14 Matěj Cepl 2010-04-06 13:52:52 UTC
Package as is in http://koji.fedoraproject.org/koji/taskinfo?taskID=2097187 (and patch commited to CVS should theoretically build; it does locally), but for missing libiodbc in EPEL-5 (see bug 579763 for that).

Comment 15 Michal Schmidt 2010-04-06 14:33:07 UTC
A couple of notes about the current poco-1.3.5-RH-old-SQLite.patch:

+// Attempt to make this build compatible with more recent versions of SQLite
+// API.
+#ifndef SQLITE_LOCKED_SHAREDCACHE
+       #define SQLITE_LOCKED_SHAREDCACHE      (SQLITE_LOCKED | (1<<8) )
+#endif

There are two reasons you're guaranteed never to receive this error code:
 - It does not exist in sqlite-3.3.6 (the version in RHEL-5).
 - You keep SQlite extended result codes disabled and this is one of them.
So I find removing the two occurrences of "SQLITE_LOCKED_SHAREDCACHE" from poco source code cleaner than defining the macro.
But it's not really wrong either way, so whatever.


-               rc = sqlite3_prepare_v2(_pDB, pSql, -1, &pStmt, &pLeftover);
+               rc = sqlite3_prepare(_pDB, pSql, -1, &pStmt, &pLeftover);

Yeah, as I said before, this change needs to be handled with care. It changes the way sqlite3_step() reports errors and I don't see  SQLiteStatementImpl::hasNext() changed appropriately. It won't be a big change, but it must be tested in error situations.

Comment 16 Fedora Update System 2010-04-14 09:19:15 UTC
poco-1.3.5-8.el5 has been submitted as an update for Fedora EPEL 5.
http://admin.fedoraproject.org/updates/poco-1.3.5-8.el5

Comment 17 Matěj Cepl 2010-04-14 09:30:59 UTC
I am aware of this comment, and I will address it in the separate bug 582147 now when we have the package in.

Comment 18 Fedora Update System 2010-06-30 14:27:04 UTC
poco-1.3.5-8.el5 has been pushed to the Fedora EPEL 5 stable repository.  If problems still persist, please make note of it in this bug report.


Note You need to log in before you can comment on or make changes to this bug.