Bug#893439: stretch-pu: package gnucash/1:2.6.15-1+deb9u1

[Julien Cristau]

Control: reassign -1 pu: libdbi/0.9.0-4+deb9u2
Control: tag -1 confirmed

On Fri, Nov 09, 2018 at 07:29:32AM +0100, László Böszörményi wrote:
> On Sat, Oct 6, 2018 at 7:07 PM Adam D. Barratt  
> wrote:
> >
> > László: ping?
> >
> > On Mon, 2018-04-02 at 15:20 +0200, Julien Cristau wrote:
> > > On Mon, Apr  2, 2018 at 14:51:54 +0300, Adrian Bunk wrote:
> > > > On Mon, Apr 02, 2018 at 01:05:39PM +0200, Julien Cristau wrote:
> > > > > On Sun, Mar 18, 2018 at 22:07:25 +0200, Adrian Bunk wrote:
> > [...]
> > > > > libdbi 0.9.0-4+deb9u1 broke gnucash tests, runtime issues
> > > > > > with this backend were so far not reported but are not
> > > > > > unlikely.
> > [...]
> > > So the other option here would be to revert the libdbi change.  As
> > > far as I can tell from #880896 it wasn't prompted by a specific
> > > problem, so a revert there might be the safest course of action and
> > > sidesteps the gnucash issue.  László, any thoughts?
>  Indeed, that change is better reverted. I've proposed a patch on
> #884119 [1]. Can I upload it to Stretch or should file a separate PU
> proposal?
> 
Please go ahead and upload to stretch.  Repurposing this bug for the
libdbi update.

Cheers,
Julien

Bug#893439: stretch-pu: package gnucash/1:2.6.15-1+deb9u1

[Adrian Bunk]

On Fri, Nov 09, 2018 at 07:29:32AM +0100, László Böszörményi wrote:
> On Sat, Oct 6, 2018 at 7:07 PM Adam D. Barratt  
> wrote:
> >
> > László: ping?
> >
> > On Mon, 2018-04-02 at 15:20 +0200, Julien Cristau wrote:
> > > On Mon, Apr  2, 2018 at 14:51:54 +0300, Adrian Bunk wrote:
> > > > On Mon, Apr 02, 2018 at 01:05:39PM +0200, Julien Cristau wrote:
> > > > > On Sun, Mar 18, 2018 at 22:07:25 +0200, Adrian Bunk wrote:
> > [...]
> > > > > libdbi 0.9.0-4+deb9u1 broke gnucash tests, runtime issues
> > > > > > with this backend were so far not reported but are not
> > > > > > unlikely.
> > [...]
> > > So the other option here would be to revert the libdbi change.  As
> > > far as I can tell from #880896 it wasn't prompted by a specific
> > > problem, so a revert there might be the safest course of action and
> > > sidesteps the gnucash issue.  László, any thoughts?
>  Indeed, that change is better reverted. I've proposed a patch on
> #884119 [1]. Can I upload it to Stretch or should file a separate PU
> proposal?

Adam?

Thanks
Adrian

Bug#893439: stretch-pu: package gnucash/1:2.6.15-1+deb9u1

[GCS]

On Sat, Oct 6, 2018 at 7:07 PM Adam D. Barratt  wrote:
>
> László: ping?
>
> On Mon, 2018-04-02 at 15:20 +0200, Julien Cristau wrote:
> > On Mon, Apr  2, 2018 at 14:51:54 +0300, Adrian Bunk wrote:
> > > On Mon, Apr 02, 2018 at 01:05:39PM +0200, Julien Cristau wrote:
> > > > On Sun, Mar 18, 2018 at 22:07:25 +0200, Adrian Bunk wrote:
> [...]
> > > > libdbi 0.9.0-4+deb9u1 broke gnucash tests, runtime issues
> > > > > with this backend were so far not reported but are not
> > > > > unlikely.
> [...]
> > So the other option here would be to revert the libdbi change.  As
> > far as I can tell from #880896 it wasn't prompted by a specific
> > problem, so a revert there might be the safest course of action and
> > sidesteps the gnucash issue.  László, any thoughts?
 Indeed, that change is better reverted. I've proposed a patch on
#884119 [1]. Can I upload it to Stretch or should file a separate PU
proposal?

Regards,
Laszlo/GCS
[1] https://bugs.debian.org/884119#17

Bug#893439: stretch-pu: package gnucash/1:2.6.15-1+deb9u1

[Adam D. Barratt]

László: ping?

On Mon, 2018-04-02 at 15:20 +0200, Julien Cristau wrote:
> On Mon, Apr  2, 2018 at 14:51:54 +0300, Adrian Bunk wrote:
> 
> > Control: tags -1 -moreinfo
> > 
> > On Mon, Apr 02, 2018 at 01:05:39PM +0200, Julien Cristau wrote:
> > > Control: tag -1 moreinfo
> > > 
> > > On Sun, Mar 18, 2018 at 22:07:25 +0200, Adrian Bunk wrote:
[...]
> > > libdbi 0.9.0-4+deb9u1 broke gnucash tests, runtime issues
> > > > with this backend were so far not reported but are not
> > > > unlikely.
[...]
> So the other option here would be to revert the libdbi change.  As
> far as I can tell from #880896 it wasn't prompted by a specific
> problem, so a revert there might be the safest course of action and
> sidesteps the gnucash issue.  László, any thoughts?
> 
> Cheers,
> Julien

Bug#893439: stretch-pu: package gnucash/1:2.6.15-1+deb9u1

[Adrian Bunk]

openbsc also FTBFS in stretch due to the libdbi change: #880233

cu
Adrian

On Mon, Apr 02, 2018 at 03:20:56PM +0200, Julien Cristau wrote:
> On Mon, Apr  2, 2018 at 14:51:54 +0300, Adrian Bunk wrote:
> 
> > Control: tags -1 -moreinfo
> > 
> > On Mon, Apr 02, 2018 at 01:05:39PM +0200, Julien Cristau wrote:
> > > Control: tag -1 moreinfo
> > > 
> > > On Sun, Mar 18, 2018 at 22:07:25 +0200, Adrian Bunk wrote:
> > > 
> > > > Package: release.debian.org
> > > > Severity: normal
> > > > Tags: stretch
> > > > User: release.debian@packages.debian.org
> > > > Usertags: pu
> > > > 
> > > > libdbi 0.9.0-4+deb9u1 broke gnucash tests, runtime issues
> > > > with this backend were so far not reported but are not unlikely.
> > > 
> > > How comprehensive are automated tests for this package,
> > 
> > Comprehensive enough that gnucash in stretch does FTBFS since the 
> > stretch-pu update of libdbi.
> > 
> > > and what manual testing was done on the new version?
> > 
> > Only very lightweight testing that is can be installed, upgraded,
> > and doesn't seem to be completely broken after startup.
> > 
> So the other option here would be to revert the libdbi change.  As far
> as I can tell from #880896 it wasn't prompted by a specific problem, so
> a revert there might be the safest course of action and sidesteps the
> gnucash issue.  László, any thoughts?
> 
> Cheers,
> Julien

Bug#893439: stretch-pu: package gnucash/1:2.6.15-1+deb9u1

[Julien Cristau]

On Mon, Apr  2, 2018 at 14:51:54 +0300, Adrian Bunk wrote:

> Control: tags -1 -moreinfo
> 
> On Mon, Apr 02, 2018 at 01:05:39PM +0200, Julien Cristau wrote:
> > Control: tag -1 moreinfo
> > 
> > On Sun, Mar 18, 2018 at 22:07:25 +0200, Adrian Bunk wrote:
> > 
> > > Package: release.debian.org
> > > Severity: normal
> > > Tags: stretch
> > > User: release.debian@packages.debian.org
> > > Usertags: pu
> > > 
> > > libdbi 0.9.0-4+deb9u1 broke gnucash tests, runtime issues
> > > with this backend were so far not reported but are not unlikely.
> > 
> > How comprehensive are automated tests for this package,
> 
> Comprehensive enough that gnucash in stretch does FTBFS since the 
> stretch-pu update of libdbi.
> 
> > and what manual testing was done on the new version?
> 
> Only very lightweight testing that is can be installed, upgraded,
> and doesn't seem to be completely broken after startup.
> 
So the other option here would be to revert the libdbi change.  As far
as I can tell from #880896 it wasn't prompted by a specific problem, so
a revert there might be the safest course of action and sidesteps the
gnucash issue.  László, any thoughts?

Cheers,
Julien

Bug#893439: stretch-pu: package gnucash/1:2.6.15-1+deb9u1

[Adrian Bunk]

Control: tags -1 -moreinfo

On Mon, Apr 02, 2018 at 01:05:39PM +0200, Julien Cristau wrote:
> Control: tag -1 moreinfo
> 
> On Sun, Mar 18, 2018 at 22:07:25 +0200, Adrian Bunk wrote:
> 
> > Package: release.debian.org
> > Severity: normal
> > Tags: stretch
> > User: release.debian@packages.debian.org
> > Usertags: pu
> > 
> > libdbi 0.9.0-4+deb9u1 broke gnucash tests, runtime issues
> > with this backend were so far not reported but are not unlikely.
> 
> How comprehensive are automated tests for this package,

Comprehensive enough that gnucash in stretch does FTBFS since the 
stretch-pu update of libdbi.

> and what manual testing was done on the new version?

Only very lightweight testing that is can be installed, upgraded,
and doesn't seem to be completely broken after startup.

> Cheers,
> Julien

cu
Adrian

-- 

   "Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
   "Only a promise," Lao Er said.
   Pearl S. Buck - Dragon Seed

Bug#893439: stretch-pu: package gnucash/1:2.6.15-1+deb9u1

[Julien Cristau]

Control: tag -1 moreinfo

On Sun, Mar 18, 2018 at 22:07:25 +0200, Adrian Bunk wrote:

> Package: release.debian.org
> Severity: normal
> Tags: stretch
> User: release.debian@packages.debian.org
> Usertags: pu
> 
> libdbi 0.9.0-4+deb9u1 broke gnucash tests, runtime issues
> with this backend were so far not reported but are not unlikely.

How comprehensive are automated tests for this package, and what manual
testing was done on the new version?

Cheers,
Julien

Bug#893439: stretch-pu: package gnucash/1:2.6.15-1+deb9u1

[Adrian Bunk]

Package: release.debian.org
Severity: normal
Tags: stretch
User: release.debian@packages.debian.org
Usertags: pu

libdbi 0.9.0-4+deb9u1 broke gnucash tests, runtime issues
with this backend were so far not reported but are not unlikely.
diff -Nru gnucash-2.6.15/debian/changelog gnucash-2.6.15/debian/changelog
--- gnucash-2.6.15/debian/changelog 2016-12-21 23:24:13.0 +0200
+++ gnucash-2.6.15/debian/changelog 2018-03-18 21:22:16.0 +0200
@@ -1,3 +1,11 @@
+gnucash (1:2.6.15-1+deb9u1) stretch; urgency=medium
+
+  * Non-maintainer upload.
+  * Add upstream fix for building with libdbi 0.9.0-4+deb9u1,
+thanks to Morham Malpense for the bug report. (Closes: #893077)
+
+ -- Adrian Bunk   Sun, 18 Mar 2018 21:22:16 +0200
+
 gnucash (1:2.6.15-1) unstable; urgency=medium
 
   * New upstream release [December 2016].
diff -Nru gnucash-2.6.15/debian/patches/789928u.patch 
gnucash-2.6.15/debian/patches/789928u.patch
--- gnucash-2.6.15/debian/patches/789928u.patch 1970-01-01 02:00:00.0 
+0200
+++ gnucash-2.6.15/debian/patches/789928u.patch 2018-03-18 21:21:48.0 
+0200
@@ -0,0 +1,75 @@
+Forwarded: not-needed
+Origin: upstream, 
https://github.com/Gnucash/gnucash/commit/45bab93613e6a93b206b74ffc18f63708b07293b.patch
+From 45bab93613e6a93b206b74ffc18f63708b07293b Mon Sep 17 00:00:00 2001
+From: John Ralls 
+Date: Tue, 7 Nov 2017 18:06:04 -0800
+Subject: Bug 789928 - FTBFS with libdbi 0.9.0-5 on Debian
+
+Commit 88b8477 on libdbi calls the error handler if one attempts to run
+off the end of a result set. Since we often loop on
+dbi_result_next_row() returning 0 this breaks our logic in several
+places. This change simply returns from the error handler on a
+DB_ERROR_BADIDX allowing the logic to work as before.
+
+--- a/src/backend/dbi/gnc-backend-dbi.c
 b/src/backend/dbi/gnc-backend-dbi.c
+@@ -395,10 +395,15 @@
+ {
+ const gchar* msg;
+ GncDbiBackend *be = (GncDbiBackend*)user_data;
+ GncDbiSqlConnection *dbi_conn = (GncDbiSqlConnection*)(be->sql_be.conn);
+-
+-(void)dbi_conn_error( conn,  );
++int err_num = dbi_conn_error( conn,  );
++/* BADIDX is raised if we attempt to seek outside of a result. We
++ * handle that possibility after checking the return value of the
++ * seek. Having this raise a critical error breaks looping by
++ * testing for the return value of the seek.
++ */
++if (err_num == DBI_ERROR_BADIDX) return;
+ PERR( "DBI error: %s\n", msg );
+ gnc_dbi_set_error( dbi_conn, ERR_BACKEND_MISC, 0, FALSE );
+ }
+ 
+@@ -610,16 +615,20 @@
+ {
+ GncDbiBackend *be = (GncDbiBackend*)user_data;
+ GncDbiSqlConnection *dbi_conn = (GncDbiSqlConnection*)be->sql_be.conn;
+ const gchar* msg;
+-gint err_num;
+ #ifdef G_OS_WIN32
+ const guint backoff_msecs = 1;
+ #else
+ const guint backoff_usecs = 1000;
+ #endif
+-
+-err_num = dbi_conn_error( conn,  );
++int err_num = dbi_conn_error( conn,  );
++/* BADIDX is raised if we attempt to seek outside of a result. We
++ * handle that possibility after checking the return value of the
++ * seek. Having this raise a critical error breaks looping by
++ * testing for the return value of the seek.
++ */
++if (err_num == DBI_ERROR_BADIDX) return;
+ 
+ /* Note: the sql connection may not have been initialized yet
+  *   so let's be careful with using it
+  */
+@@ -1330,10 +1339,16 @@
+ const guint backoff_msecs = 1;
+ #else
+ const guint backoff_usecs = 1000;
+ #endif
++int err_num = dbi_conn_error( conn,  );
++/* BADIDX is raised if we attempt to seek outside of a result. We
++ * handle that possibility after checking the return value of the
++ * seek. Having this raise a critical error breaks looping by
++ * testing for the return value of the seek.
++ */
++if (err_num == DBI_ERROR_BADIDX) return;
+ 
+-(void)dbi_conn_error( conn,  );
+ if ( g_str_has_prefix( msg, "FATAL:  database" ) &&
+ g_str_has_suffix( msg, "does not exist\n" ) )
+ {
+ PINFO( "DBI error: %s\n", msg );
diff -Nru gnucash-2.6.15/debian/patches/series 
gnucash-2.6.15/debian/patches/series
--- gnucash-2.6.15/debian/patches/series2016-12-21 23:24:13.0 
+0200
+++ gnucash-2.6.15/debian/patches/series2018-03-18 21:22:16.0 
+0200
@@ -1 +1,2 @@
+789928u.patch
 hardening-fortify.patch