Cups 2.3.1 CUPS_SIDES returns different default value than sides-default

[RolandHughes]

Yes, I know 2.3.1 is old but I'm on Ubuntu 20.04. In case this issue still exists I wish to report it and I also wish to get edification on what method should actually be trusted. My little test program dumps out the following at the start.

`Dest Name: Brother_MFC_L3770CDW_series
uri:
resolution printer-resolution-default: 600dpi
text value: 600 integer value: 600
DPI 600

==========  media-source: media-source  default: auto

 !!!!!!! orientation: 
                      portrait
                      landscape
 ;;;;;;;;;;;;;media source: 0  name: auto
 ;;;;;;;;;;;;;media source: 1  name: by-pass-tray
 ;;;;;;;;;;;;;media source: 2  name: tray-1
 ++++++++++++output bin: 0  name: face-down

``````````  output-bin-default: output-bin-default   value: face-down
 ********side: 0  name: one-sided
 ********side: 1  name: two-sided-long-edge
 ********side: 2  name: two-sided-short-edge
 ********color: 0  name: auto
 ********color: 1  name: color
 ********color: 2  name: monochrome
 ********color: 3  name: auto-monochrome

xxxxxxxxxxxx found CUPS_PRINT_QUALITY_SUPPORTED
jjj: 0 integer: 4 string version: normal

---------- CUPS_PRINT_COLOR_MODE   name: print-color-mode-default  default: color

........... CUPS_SIDES   name: sides-default  default: one-sided`

A bit later down when I'm dumping all of the attributes for same printer.

output-bin-default = face-down tag: 4 output-mode-default = color tag: 4 print-quality-default = normal tag: 4 printer-resolution-default = 600dpi tag: 4 sides-default = two-sided-long-edge tag: 4 print-color-mode-default = color tag: 4 print-content-optimize-default = auto tag: 4 print-scaling-default = auto tag: 4

Shouldn't the output match?

The little test program doesn't do anything kinky.

` if ( ( attr = cupsFindDestDefault( CUPS_HTTP_DEFAULT, dest, info, CUPS_PRINT_COLOR_MODE ) ) != NULL )
{
name = ippGetName( attr );
ippAttributeString( attr, value, sizeof( value ) );
out << endl << "\t---------- CUPS_PRINT_COLOR_MODE name: " << QString::fromUtf8( name ) << " default: "
<< QString::fromUtf8( value ) << endl;
}

        if ( ( attr = cupsFindDestDefault( CUPS_HTTP_DEFAULT, dest, info, CUPS_SIDES ) ) != NULL )
        {
            name = ippGetName( attr );
            ippAttributeString( attr, value, sizeof( value ) );
            out << endl << "\t........... CUPS_SIDES   name: " << QString::fromUtf8( name ) << "  default: "
                << QString::fromUtf8( value ) << endl;
        }

`
value is a generic char [4096] dumping ground that got used for color in the output just prior to this so not a stale data issue.

If this has been fixed in later versions, so be it. I included the printer model here in case this was a known Brother issue for one or more models. Am I correct in assuming CUPS_SIDES should return the same value as the sides-default attribute when it gets dumped just a little later?

[zdohnal]

Please attach a file with the whole code - you can put .txt as suffix and github allows it. I don't see how you get dest, or where do you print all attributes - I suspect you might not be talking to the same object.

[RolandHughes]

I included the full output in the file as well.
cups-experiment.txt

It is a very simple program lifted almost verbatim from one of the examples.

When one physically checks the printer in question, going through the menu on the little display, the default really is 2-sided. That printer/fax doesn't get used a lot so spends most of its life sleeping.

Since you didn't respond with "oh, yeah, that's fixed in . . . " I suspect but have not dug into

if ( ( attr = cupsFindDestDefault( CUPS_HTTP_DEFAULT, dest, info, CUPS_SIDES ) ) != NULL ) { name = ippGetName( attr ); ippAttributeString( attr, value, sizeof( value ) ); out << endl << "\t........... CUPS_SIDES name: " << QString::fromUtf8( name ) << " default: " << QString::fromUtf8( value ) << endl; }
is not waiting for actual response from the printer which has to come partially out of slumber so it uses a stale buffer.
By the time we get done dumping all of the media, the attributes are no longer stale.

That is just a knee-jerk suspicion though.

[RolandHughes]

Additional data point in the output.

You will find that in the output file and it is for the same printer. According to what I've read in the doc the following should first retrieve the default and then retrieve the supported list.

No matter how many passes are made, it does not.

I should get the string
draft,normal,high
and it should comma split into a string list for combobox items.

Header file has correct values

It's almost like the cupsFindDestSupported() call isn't adding "-supported" under the hood like it does for other values.

Do I have to build my own cupsDoRequest() to work around this? I would prefer to use the higher level API if possible, but understand if things aren't quite good.

[RolandHughes]

In case you wish to build and go with example I have uploaded the version of LsCs it is built against. The Debian package is large because compiled with debug information. You can download here. Please let me know when you have it pulled down. That thing is consuming a gig of my 2GB free account.

[RolandHughes]

In case there is any confusion, that cups-experiment.txt was a zip file.

[RolandHughes]

Not an answer, but have identified the bulk of the problem.
cups-experiment2.txt
The attached zip file contains trimmed down output file cups-experiment-output-2.txt as well as the source.

cupsEnumDests() returns 6 printers. 3 configured for only "defaults" and 3 others showing some of the actual capabilities.

cupsGetDests() and cupsGetDests2() return only the 3 configured for defaults.

The Brother printers have 600 default then either 1200 or 2400dpi for their fine/high quality.

There is no real work around here. To build a list of valid dests one would have to have a bogus cupsEnumDests() call each and every time just to build the list cupsGetDests() should be building.

It is also odd that the ancient HP OfficeJet Pro is playing nicest with Cups given the brother printers are much newer and driverless.

I don't think I can provide more information.

Yes, I know Ubuntu 20.04 and the Cups 2.3.x it has is old, but . . .