http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=11188 --- Comment #1 from Magnus Enger <magnus@enger.priv.no> --- Created attachment 23181 --> http://bugs.koha-community.org/bugzilla3/attachment.cgi?id=23181&action=edit Bug 11188 - Make gather_print_notices.pl die on failed open() Problem: If you tell gather_print_notices.pl to write output to a location you do not have write access to, it will silently fail to write the data, but still mark unsent messages as sent. Solution: This patch adds two lines of defense: 1. Check that the location given for the output is writable 2. use "open() or die" instead of just "open()" when writing the output The first measure should catch most of the potential errors, but I guess a directory can be writable, but the open() still can fail because the disk is full or something similar. To test: - Make sure you have some unsent messages in the message_queue table, that do not have an email adress - Apply the patch - Run the script, pointing at a location you do not have access to write to. Check that the script exits with an appropriate error message, and that the unsent messages are still unsent. Do this both with and without the -s option. - To fake passing the first line of defence, comment out line 62 and put this in instead: if ( !$output_directory || !-d $output_directory ) { - Run the script again as above, check you get an appropriate error and that the message queue is not touched - Reset line 62 to how it was - Run the script against a directory you do have access to write to and check that output is produced as expected and that messages are marked as sent - Sign off -- You are receiving this mail because: You are watching all bug changes.

http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=11188 Chris Cormack <chris@bigballofwax.co.nz> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #23181|0 |1 is obsolete| | --- Comment #2 from Chris Cormack <chris@bigballofwax.co.nz> --- Created attachment 23849 --> http://bugs.koha-community.org/bugzilla3/attachment.cgi?id=23849&action=edit Bug 11188 - Make gather_print_notices.pl die on failed open() Problem: If you tell gather_print_notices.pl to write output to a location you do not have write access to, it will silently fail to write the data, but still mark unsent messages as sent. Solution: This patch adds two lines of defense: 1. Check that the location given for the output is writable 2. use "open() or die" instead of just "open()" when writing the output The first measure should catch most of the potential errors, but I guess a directory can be writable, but the open() still can fail because the disk is full or something similar. To test: - Make sure you have some unsent messages in the message_queue table, that do not have an email adress - Apply the patch - Run the script, pointing at a location you do not have access to write to. Check that the script exits with an appropriate error message, and that the unsent messages are still unsent. Do this both with and without the -s option. - To fake passing the first line of defence, comment out line 62 and put this in instead: if ( !$output_directory || !-d $output_directory ) { - Run the script again as above, check you get an appropriate error and that the message queue is not touched - Reset line 62 to how it was - Run the script against a directory you do have access to write to and check that output is produced as expected and that messages are marked as sent - Sign off Signed-off-by: Chris Cormack <chris@bigballofwax.co.nz> -- You are receiving this mail because: You are watching all bug changes.

http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=11188 Katrin Fischer <katrin.fischer@bsz-bw.de> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #23849|0 |1 is obsolete| | --- Comment #3 from Katrin Fischer <katrin.fischer@bsz-bw.de> --- Created attachment 23855 --> http://bugs.koha-community.org/bugzilla3/attachment.cgi?id=23855&action=edit [PASSED QA] Bug 11188 - Make gather_print_notices.pl die on failed open() Problem: If you tell gather_print_notices.pl to write output to a location you do not have write access to, it will silently fail to write the data, but still mark unsent messages as sent. Solution: This patch adds two lines of defense: 1. Check that the location given for the output is writable 2. use "open() or die" instead of just "open()" when writing the output The first measure should catch most of the potential errors, but I guess a directory can be writable, but the open() still can fail because the disk is full or something similar. To test: - Make sure you have some unsent messages in the message_queue table, that do not have an email adress - Apply the patch - Run the script, pointing at a location you do not have access to write to. Check that the script exits with an appropriate error message, and that the unsent messages are still unsent. Do this both with and without the -s option. - To fake passing the first line of defence, comment out line 62 and put this in instead: if ( !$output_directory || !-d $output_directory ) { - Run the script again as above, check you get an appropriate error and that the message queue is not touched - Reset line 62 to how it was - Run the script against a directory you do have access to write to and check that output is produced as expected and that messages are marked as sent - Sign off Signed-off-by: Chris Cormack <chris@bigballofwax.co.nz> Signed-off-by: Katrin Fischer <Katrin.Fischer.83@web.de> Passes all tests and QA script. Works as described. -- You are receiving this mail because: You are watching all bug changes.

http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=11188 Fridolin SOMERS <fridolyn.somers@biblibre.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|Pushed to Master |Pushed to Stable CC| |fridolyn.somers@biblibre.co | |m --- Comment #5 from Fridolin SOMERS <fridolyn.somers@biblibre.com> --- Patch pushed to 3.14.x, will be in 3.14.3. -- You are receiving this mail because: You are watching all bug changes.