by EmuandCo | May 24, 2017
Meeting started at 20:31 by Aleksey Bragin
- Point 1: Status Reports
- Point 2: Hiring developers
- Point 3: GSoC merging
Point 1: Status Reports
Aleksey Bragin: He sold his startup, got a CTO position in a company which bought his startup, and ideally he should have more time, but it's not the case right now. So basically he was busy whole October and he started doing some research into software image processing for optical tech (his students, not him)
Aleksandar Andrejevic: He hasn't had much time to do anything yet, but he's still planning on writing the Sound Blaster driver and finishing NTVDM VESA support. And of course he will continue fixing NTVDM bugs, as he's always been doing.
Colin Finck: As you have seen, he finally got his printing branch merged (yeeey!!) after more than a year of apparently no progress. This only happened after implementing that missing RPC function in Wine's rpcrt4.dll. Christoph really helped on that one! And by the way, rpcrt4.dll really looks like a component we need to fork at some point being totally ANSI-only in many places and only implementing RPC up to the point WINE needs. At least, it looks much friendlier now after spending some time with it. Apart from that, you can expect his mkisofs patch to merge that evening (CORE-11988) giving us proper ISO9660:1999 images, build support for OS X as host again and soon a feature for storing duplicate files only once. The rest has been server stuff behind the scenes. He doesn't have good news to report regarding our previous server farm, he really wishes he had. Aleksey promised to visit Jan the other week and finally bring those to his hosting place at home, but he got no update from either one. Many rather boring upgrade tasks are still outstanding, but he cannot say if he will have them done by the next meeting. All in all, just follow his Sprint at JIRA. He tries to update it rather regularly with what he's working on.
Kamil Hornicek: He just got back after almost a year of inactivity so there's not much from his side.. he's been working on Testman and intend to fix the issues Victor brought up. Also he started tackling failed winetests. He's currently working on bitmaps and co and has uncommited changes that fix the majority of gdi32:bitmap winetests but these will need to be reviewed first because the DIB code is kinda fragile.
Johannes Anderwald: He has been working on usb audio - it works now already in win2k3 for usb microphones, next step is to get usb speakers working. Though it will be long road until it will work due to missing functionality in ks and usb host drivers. He's implementing it right now for usb audio 1.0 spec - though it is not too much work to get it done for 2.0 of spec.
Mark Jansen: He got a big part of the application compatibility framework related to the launching of processes working (and testable!), which he hopes to land somewhere in the upcoming weeks. After this, he should tackle the initialization (in-process), and at that point it should be functional enough for basic fixes. he also started on rapps_new, familiarizing himself with the code and starting on the visual overhaul.
Pierre Schweitzer: Slowly trying to get back to ROS dev, as much as his daily work lets him doing. He's managed to get his last MPR patch into Wine. For the rest, he'll have to make a diff against Wine (saving/restoring connections). He's also working on RDBSS/RXCE as these reading #reactos-dev may have seen. And he's updating btrfs driver.
Sylvain Petreolle: He completed the work on his AHK bot. It is ready to test and accepts patches. The ahk tests can be patched, too. They are built on the fly.
Hermes Belusca Maito: He's worked on reducing code redundancy for our console apps by putting common code in a library. This relates to code that correctly displays unicode strings on the console, as well as some code (mainly coming from cmd.exe so far) that can be found in some tools. What he's going to do next, he doesn't know, it depends on his time (as he's continuing searching for postdocs until he gets a job).
Thomas Faber: He's been looking and Chrome and Mm a little bit, but not spending a whole ton of time on ROS atm. Hopes to get CORE-12132 done in time for the release. That's about it. He's looking at things one problem at a time. What's blocking the installer right now is a fastfat bug. He'll see what comes after.
Victor Martinez: During these last 3 weeks he has been really out of testing and in IRC as you may have noticed, basically because he was working on a small surprise for ReactOS. Our current way to manage Rapps is messy, basically we are using Jira to report Dreimer as that should be included, and since rapps_new is coming...we need a better approach. He presents you a small idea, a web based tool to update the database. Basically this tool lets us to have 2 kind of users: admin and users. Users can send apps suggestions and apps translations. The admin is notified by a big number and he can there accept new apps and new translations. You can make users as admins or make them inactive.
1) Multiple admins.
2) Automatically lets the user to send translations on its own language.
3) Accept/Reject apps/translations.
4) Autocheck for links.
Right now it is using it's own users database but we can use drupal, should be no issue. CSS styling and much more testing is needed, but the basis is there. He can have this probably rolling before 0.4.3 is out and from here we can think in tons of other ideas he has in mind. All the info can be automatically zipped and automatically downloaded by our rapps. So all the translations and apps are managed through this platform but could be used by rapps to populate the apps inside it. This way, the rapps db upgrade can be decoupled from rapps upgrades.
Point 2: Hiring developers
Due to the fact that money doesn't code Victor Martinez, Colin Finck and Aleksey Bragin discussed the current situaion. Colin presented the current financial situation we have and how many donation come in every month. According to his statistics the new website did improve this a lot. Victor asked for some numbers regarding number of recurring donations and number of donations done for his portfolio and knowledge. Colin guided him to the publicly available statistics, because no other statistics exist: https://reactos.org/donors. No developer is funded right now. No contract, no scholarship. Here Hermes Belusca Maito asked what happened to his papers to Amine for candidating. Colin already dicovered that putting the burden regarding the contract paperwork on single persons is a flaw. One single person is a bottleneck by definition and thus he asked to him on CC in the mails in the future, whenever ppl apply for a contract, negotiate a contract, etc. He doesn't want us to lose important chances here just because a single person lacks time. All in all the current situation allows us to look for full time developers. Aleksey suggested to check Russia for developers, because they are cheaper there. More will follow.
Point 3: GSoC merging
Thomas Faber reviewed quite some code from our GSoC projects.
USB ist just partially reviewed and should not be merged before he finished it.
AHCI can simply be merged... our headers need fixing for it to be able to build. It is gonna need a working storport.h, but that can be a separate step. Someone should make a ticket regarding that storport.h, mention that it blocks merging the AHCI branch back and it will be done
TCP is an alternative to our current implementation, so it could be added as a _new.
Colin Finck's printing branch was just replacing our ancient winspool.h with mingw-w64's version to work.
Meeting was closed at 21:08 by Colin Finck
Meeting minutes prepared by Daniel Reimer