Your suggested changes sound fine to me. allan
On Sun, Aug 25, 2013 at 1:09 PM, Chris Bagwell <chris at cnpbagwell.com> wrote: > Hi all, > > This is probably directly more towards St?phane based on recent USB work but > of interest to all. > > The sanie_usb_test.c test is failing on one of my Linux boxes in the > test_open_all() test... I'd like to get some input on best way to fix this > failure or if we should fix it at all. > > The reason it is failing for me is because this box has a USB wifi module > that has a report with Vendor Specific Class/Subclass/Protocol and so sane > adds it to the list as possible scanner. The open itself fails when ran by > normal user because its accessable only by root (EACCES). The open fails > when ran as root because its already opened by networking stuff (EBUSY). > > Because these two cases are kinda soft failures, I was thinking about > changing the test to simply skip over these opens instead of failing the > whole test case. What do you think about this change? > > If you agree with that change, the second issue I'll have is that > sanei_usb_open() seems to treat EPERM as permission denied > (SANE_STATUS_ACCESS_DENIED) but not EACCES. Do you see any issues with me > changing it so both EPERM and EACCES map to SANE_STATUS_ACCESS_DENIED? > > If I don't make that change then the EACCES error code is mapped to generic > SANE_STATUS_INVAL and I can't tell difference between real failures and > permission failures. > > Chris > > -- > sane-devel mailing list: sane-devel at lists.alioth.debian.org > http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/sane-devel > Unsubscribe: Send mail with subject "unsubscribe your_password" > to sane-devel-request at lists.alioth.debian.org -- "The truth is an offense, but not a sin"
