First year of monthly news posts completed!
Some will tell me now that I already started in September 2019, but the September 2019 news post was only a test for the new platform and it was already October when I wrote it as an exact copy of the original mailing list news post.
Google Summer of Code 2020 Mentor Summit
Here also the GSoC 2021 and its new mode was announced.
Google Summer of Code 2021
There will be a Google Summer of Code next year again, but it will be different. The 3-months student projects will be replaced by part-time projects, 6-weeks full-time-equivalent, to be done in a 10-week time-window. Stipends are appropriately reduced to the half amount, leading to the smae per-hour value.
Google hopes to attract more students by this mode. For us as org admins and mentors it means more work, as we have to select more students and we have to work with more students when mentoring. Also we cannot post bigger projects. So let us hope that there will also be a Linux Foundation Mentoring Program for the larger projects.
Google Season of Docs 2020
Our OpenPrinting project
Tutorial and Design Guidelines for Printer/Scanner drivers in Printer Applications
is going well. Piyush Goyal has already written a lot and will continue until December 5.
He gets currently a lot of input, from me with my PostScript Printer Application as example for a non-Raster and CUPS-driver-retro-fit Printer Application and from Michael Sweet by adding documentation to his PAPPL project.
Linux Foundation Mentorship Program
In our first LFMP project also Nidhi Jain has withdrawn, also due to suddenly re-started classes in her college and a COVID-19 case in her family.
We had some discussion between Alexander Pevzner, Michael Sweet, and me whether we should actually use IPP Scan as network protocol for sandboxed packaging of scanner drivers/Scanner Applications or better eSCL instead, as eSCL is already implemented as client and as server, but we came to the final conclusion of continuing with IPP Scan.
We only will move the assignment of the remaining work somewhat. Alexander will complete the IPP Scan support in sane-airscan by himself and both students will work on the server side.
After the Pull Request for the implementation of the interfaces got merged we are still waiting for the completion of the pull request for adding the API functionality so that a Snap can check whether a client Snap plugs the needed interfaces is still in the works.
PostScript Printer Application
After having all needed filter functions and PPD file collection handling implemented in cups-filters and having the HP PCL Printer Application as example I have created a Printer Application for PostScript printers.
It is now available as the ps-printer-app repository on the OpenPrinting GitHub.
Currently it is a first working model, it will get much more functionality, especially for best user experience, added in the near future. See all the details of properties, planned features, known issues, and how to use in the README.md file.
This Printer Application is especially a working model for
A non-raster Printer Application: Destination format is PostScript, a high-level/vector format. Input data in PostScript or PDF is accepted and needed conversion is not done through an inbetween raster step.
A Printer Application which uses the new filter functions of cups-filters 2.x. Filter functions are library functions derived from cups-filters and contain decades of development and refinement starting from the introduction of CUPS in 2000.
A retro-fit Printer Application for classic CUPS drivers, in this case the simplest form of only PPD files for PostScript printers. It lists PPD files from repositories included in the Snap, loads the PPD needed for the actual printer, extracts options from the PPD to display them in the web interface, accepts job settings as IPP attributes, and inserts the PostScript code provided by the PPD correctly into the output data stream.
A Printer Application which does not pass through raw (input format is printer’s native format) jobs. To assure that always the PostScript code of the PPD file is inserted into the output stream, we call the printer’s native format “application/vnd.printer-specific” which does not exist as input format, so “application/postscript” input is forced through the pstops() filter function.
To do not need to re-invent the code for forking into sub-processes so that we can pass data through a sequence of filters, we create a filter function to send the data off to the printer and form a chain of the actually converting filter function (one of pstops(), pdftops(), imagetops(), rastertops()) and the with this filter function using the filterChain() filter function.
The PostScript Printer Application has all PostScript PPD files of the foomatic-db and HPLIP projects built-in, so most PostScript printer PPDs which usually come with Linux Distributions. Note that some PPDs use certain CUPS filters for extra functionality. These filters are not included. The user can add additional PPDs without needing to rebuild the Snap.
Driverless Scanning and IPP-over-USB
The Main Inclusion Requests (MIR) to get ipp-usb passed the Ubuntu security team and so made it to be the default IPP-over-USB implementation in the Groovy Gorilla, Ubuntu 20.10, released October 22, some days ago. So we have reliable IPP-over-USB in Ubuntu now.
Unfortunately, the security team was swmped with requests and so did not come to processing the MIR for sane-airscan. So it will only get into the Main repository in Ubuntu 21.04, to be released in April 2021. As it is already in the Universe repository, it can still get easily installed in 20.10. Also upstream is always supplying the newest version in packages for common Ubuntu releases.
Alexander Pevzner is continuing to take user reports and make the packages working with as many devices as possible.
IPP Scan support is currently still worked on in sane-airscan.
Currently released is 2.3.3.
Due to dormant upstream development we have created a fork an OpenPrinting which will replace the original repository if Apple does not come back with the development. So for the time being CUPS is maintained by Michael Sweet and me.
Michael Sweet has already committed a lot of changes and fixes and Debian has put the first packages of the fork into Experimental.
Apple having stopped CUPS development got also talked about at Phoronix, at the German Linux Magazin, and on the Register. I have answered on the former two, telling that we have started the fork and that CUPS will live on.
Currently released is 1.28.5.
On the way towards 2.0.0 we have converted many filters to filter functions, having now pstops, pdftops, pdftopdf, ghostscript (replaces gstoraster, gstopdf, gstopxl), pclmtoraster, rastertopdf (replaces rastertopdf amd rastertopclm), rastertops, imagetopdf, imagetops, imagetoraster, and filterChain.
With the PostScript Printer Application the filter function concept got its first real life use. Due to this I found and fixed many bugs, especially filterChain() got actually tested for the first time and so found some important bugs and fixed them so that filterChain() actually works now.
Also the functions for handling PPD file collections in libppd got actually used for the first time and I spotted and fixed some bugs in them.
The release of Ubuntu 20.10 (Groovy Gorilla) also made me fix some bugs in the 1.x series of cups-filters.
Bug fix release for a quick, potential crasher correction in cups-browsed
Bug fix release, mainly to solve CUPS performance problems caused by the IPP fax support of the “driverless” utility
Bug fix release to solve problems caused by an inconsistency between the resolvers for DNS-SD-based URIs
CHANGES IN V2.0.0 - libppd: Added PPD collection handling functions (based on the code of cups-driverd in CUPS), for listing collections of PPD files and picking a PPD out of the collection. PPDs can be in arbitrary subdirectory structures, .tar and .tar.gz files, and also be generated by executables, as one uses them with CUPS in /usr/lib/cups/driver. This is for Printer Applications which retro-fit classic CUPS drivers. - libppd, libcupsfilters: Share the definition of logging functions between the two libraries, so that the same logging functions can be used with both. - Sample PPDs: Corrected manufacturer name in Fuji_Xerox-DocuPrint_CM305_df-PDF.ppd. - texttopdf: Added support for PPD-less use. - libcupsfilters, imagetops: Introduced imagetops() filter function for Printer Application to print on PostScript printers. It is used to print images (IPP standard requires at least JPEG to be accepted) without need of a PDF renderer (imagetopdf -> pdftops) and without need to convert to Raster (imagetoraster -> rastertops). The CUPS filter imagetops uses this filter function now, too. - libcupsfilters, imagetopdf: Moved core functionality of imagetopdf into the imagetopdf() filter function. - libcupsfilters, imagetoraster: Moved core functionality of imagetoraster into the imagetoraster() filter function. - sys5ippprinter, cups-browsed: Removed sys5ippprinter, as CUPS does not support System V interface scripts any more. This first approach of PPD-less printing was also not actually made use of.
CHANGES IN V1.28.5 - cups-browsed: UUID from IPP response was used after its pointer was freed by ippDelete() (Pull request #311). CHANGES IN V1.28.4 - driverless: Avoid duplicate PPD list entries from the same device via UUID - driverless: Reduce ippfind calls by "driverless" and "driverless-fax"called by CUPS. Let "driverless list" list both print and fax PPDs and "driverless-fax list" do nothing. - driverless: Avoid duplicate listings in printer discovery, by "driverless-fax" not listing any URI as "driverless" lists them all already. - driverless: Vastly improve performance by doing only one ippfind call instead of two (IPP, IPPS) as ippfind accepts more than one reg type on the command line. - Sample PPDs: Corrected manufacturer name in Fuji_Xerox-DocuPrint_CM305_df-PDF.ppd. CHANGES IN V1.28.3 - libcupsfilters, cups-browsed: Fixed inconsistency between resolvers for DNS-SD-based URIs, resolve_uri() and ippfind_based_uri_converter(). Now both return a freeable string. - libcupsfilters: Fix uninitialized buffer and parsing ippfind output in ippfind_based_uri_converter() function (Issue #308, Pull request #309).