Often questions come up about the meaning of FileIDs and ETags. Both values are metadata that the ownCloud Server stores for each of the files and directories in the server database. These values are fundamentally important for the integrity of data in the overall system. Here are some thoughts about what they are why these are so important.This is mainly from a clients point of view, but there are other use cases as well.

ETags

ETags are strings that describe exactly one specific version of a file (example: 71a89a94b0846d53c17905a940b1581e).

Whenever the file changes, the ownCloud server will make sure that the ETag of the specific file changes as well. It is not important in which way the ETag changes, it also does not have to be strictly unique, it’s just important that it changes reliably if the file changes for whatever reason. However, ETags should not change if the file was not changed, otherwise the client will download that file again.

In addition to that, The ETags of the parent directories of the file have to change as well, up to the root directory. That way client systems can detect changes that happen somewhere in the file tree. This is in contrast to normal computer file systems where only the modification time of the direct parent of a file is changing.

File IDs

FileIDs are also strings that are created once at the creation time of the file (example: 00003867ocobzus5kn6s).

But contrary to the ETags, the file IDs should never ever change over the files lifetime. Not on an edit of the file, and also not if the file is renamed or moved. One of the important usages of the FileID is to detect renames and moves of a file on the server.

The FileID is used as an unique key to identify a file. FileIDs need to be unique within one ownCloud, and in inter-owncloud connections, they must be compared together with the ownCloud server instance id.

Also, the FileIDs must never be recycled or reused.

Checksums?

Often ETags and FileIDs are confused with checksums such as MD5 or SHA1 sums over the file content.

Neither ETags nor FileIDs are, even if there are similarities: Especially the ETag can be seen as a checksum over the file content. However, file checksums are way more costly to compute than just a value that only needs to change somehow.

What happens if…?

Let’s make a thought experiment and consider what it would mean especially for sync clients if either fileID or ETag gets lost from the servers database.

If ETags are lost, clients loose the ability to decide if files have changed since the last time that was checked by the clients. So what happens is that the client will download the files again, byte-wise compare them to the local file and use the server file if the files differ. A conflict file will be created. Because the ETag was lost, the server will create new ETags on download. This could be improved by the server creating more predictable ETags based on the storage backends capabilities.

If the ETags are changed without reason, for example because a backup was played back on the server, the clients will consider the ones with changed ETags as changed and redownload them. Conflict handling will happen as described if there was a local change as well.

For the user, this means a lot of unnecessary downloads as well as potential conflicts. However, there will not be data loss.

If FileIDs got lost or changed, the problem is that renames or moves on server side can no longer be detected. That would result in a new download of files in the good case. If a fileID however changes to something that was used before, that can result in a rename that overwrites an unrelated file. That is because clients might still have the FileID associated with another file.

Hopefully this little post explains the importance of the additional metadata that we maintain in ownCloud.

I’ve been looking for a good way to access my corporate email on my linux (opensuse) laptop. Evolution is pretty good, but the evolution-ews plugin is REALLY buggy. The connection was dropping every few minutes for me. The user interface also feels too cluttered for what I’m trying to do. So, I decided to try a different approach.

I wrote a simple python based webkit browser to modify the look and feel of owa to make it more like a desktop app.

You can access the source code here:

http://github.com/DavidMulder/simple_browse

Or you can install the opensuse package here:

http://software.opensuse.org/package/simple_browse?search_term=simple_browse

I also plan to use it to play Netflix on my mythtv box. It seems like I should be able to modify the css styling to make the summaries bigger on the page, etc.

Recently we saw some lively discussions about support of Half within the Tiff image format on the OpenEXR mailing list. That made me aware of the according oyHALF code paths inside Oyranos. In order to test easily, Oyranos uses the KISS format PPM. That comes with a three ascii lines header and then the uncompressed pixel data. I wanted to create some RGB images containing 16-bit floating point half channels, but that PFM format variant is not yet defined. So here comes a RFC.

A portable float map (PFM) starts with the first line identifier “Pf” or “PF” and contains 32-bit IEEE floating point data. The 16-bit IEEE/Nvidia/OpenEXR floating point data variant starts with a first line “Ph” or “PH” magic similar to PFM. “Ph” stands for grayscale with one sample. The “PH” identifier is used for RGB with three samples.

That’s it. Oyranos supports the format in git and maybe in the next 0.9.6 release.

Last time I’ve took look at Simple Direct Layer and how to get audio out of it. If SDL still feels little bit too hard to cope with I think I have solutions for you: libAO. Besides being no brainier with API libAO provides huge list of supported operating systems.
There is so much audio systems supported that you won’t be dissapointed but as much as I like everyone use Roaraudio. I don’t see it’s happening really soon (sorry roar you had your time in fame) but supporting Roaraudio  doesn’t mean that libAO is obsolete. It’s far from being obsolete. Libao supports OSS, ALSA and Pulseaudio out of the box and only problem is license is GPL 2.0+ so it’s no-go for proprietary development.

History

LibAO is developed under Xiph umbrella. Xiph is the organization who brought you Ogg/Vorbis, FLAC, Theora and currently they are hammering together next generation video codec Daala. Opus-audio codec standard is also Xiph project. LibAO rised from Xiph’s need multi-platform audio output library for Vorbis-audio codec. In this point if you don’t have any glue what I just said in last sentences I think you should take your spoon and start shovelling about Open Source audio codecs.
Becaus of the history libAO only has output mode and doesn’t use any callbacks. It doesn’t have fancy Float32 mode (as much as I understood) but that doesn’t say it’s bad thing. It works as expected you just feed bytes and after while you hear them from your speakers.

What about API

Supported outputs: Alsa, Oss, Jack, Mac OS X, Windows
License: GNU General Public license 2.0+

As said libAO API is difficult to describe since there almost ain’t NAN of it. You initialize, ask output device, put in your mode settings and start feeding data. Pulseaudio simple is almost easy as this but it’s still more difficult if you compare it to libAO. LibAO doesn’t support recording so only output and there must be a way to use another device than default but it’s not very easy to find or I was too lazy to dig it out.

So who wants to use libAO? People in hurry and don’t mind GPL-license, someone with very very tiny need of just getting audio out and people how hate bloat.

Summary: So if you hate bloat and again license doesn’t make you unhappy please use this library. Still libAO has kind of same problem that ALSA has. It’s mature, usable and ready for hardcore torturing but is it sexy? No! Is fresh? No, No! Is something that will change it API every week or hour?

After this I have to choose what to bring next. I have FFmpeg, Gstreamer and VLC in row. If you have opinion about next framework let me know.

As this is hosted on my personal blog, this should probably go without saying, but given the topics covered in this post I wish to state for absolute clarity that the content of this post does not necessarily reflect the official opinion of the openSUSE Project or my employer, SUSE Linux GmbH. This is my personal opinion and should be treated as such

One thing I get asked, time and time again, is "Why openSUSE?". The context isn't always the same "Why should I use it?", "Why should I contribute to openSUSE?", "Why do you use it?", "Why not use [some other distribution]?" but the question always boils down to "What's special about the openSUSE Project?"

It's a good question, and in the past, I think it's one which both the openSUSE Project as a whole and me as an individual contributor have struggled to satisfactorily answer. But I don't believe that difficulty in answering is due to a lack of good reasons, but an abundance of them, mixed in with a general tendency within our community to be very modest.

This post is my effort to address that, and highlight a few reasons why you, the reader, should contribute to the openSUSE Project and use the software we're developing

Reason 1: We're not (just) a Linux Distribution

Normally, the first thing people think of when they hear openSUSE is our 'Regular Release Distribution', such as openSUSE 13.2 which we released last year (Download it HERE)

The openSUSE Project however produces a whole bunch more.

For starters, we have not one, but technically TWO other Linux Distributions which we release and maintain

openSUSE Tumbleweed - our amazing rolling release which manages to give it's users stable, usable, Linux operating system, with constantly updating software.
Perfect not only for Developers, but anyone who wants the latest and greatest software, leading Linux experts like Greg Kroah-Hartman have stated rolling releases like Tumbleweed represent the 'future' of Linux distributions. I agree, and I think openSUSE's is the best (for reasons that will be made clear later in this post).
In my case it's replaced all but one of my Linux installations, which previously included non-openSUSE distributions. The last (my openSUSE 13.1 server) is likely to end up running Tumbleweed as soon as I have a weekend to move it over. I started small with Tumbleweed, having it on just one machine, but after a year, I just don't see the need for the 'old fashioned' releases any more in any of my use cases.
If you want to see it for yourself, Download it HERE

openSUSE Evergreen - this is a great community driven project to extend the lifecycle of select versions of our Regular Release beyond its usual "2 Releases + 2 Months" lifespan.
While technically not a 'true' separate Distribution in the purest sense of the word, it takes a herculean effort from our community to keep on patching the distribution as it gets older and older, but it allows us to offer a 'Long Duration Support' for a significant period after the regular lifespan.
The current openSUSE Evergreen release is openSUSE 13.1, which will be supported until at least November 2016 and can be Downloaded HERE

So, we're not just *a* Linux distribution but several, but wait, there's more!

The openSUSE Project hosts a whole pile of other software Projects, always Open Source, which we both use ourselves and encourage others to use and contribute to. These include:

Open Build Service - our build tool, building all of our packages as well as ones for SUSE Linux Enterprise, Arch, Debian, Fedora, Scientific Linux, RHEL, CentOS, Ubuntu, and more.

openQA - automated testing for *any* operating system, that can read the screen and control the test host the same way a user does.

YaST - The best/only comprehensive Linux system configuration & installation tool.

KIWI - Create Linux images for deployment on real hardware, virtualisation, and now even container systems like Docker. Kiwi is the engine that powers SUSE Studio

Snapper - Create, manage, and revert system snapshots. Roll-back without hassle when your system gets messed up (p.s. It's probably your fault it messed up, so Snapper can show you what changed between snapshots too)

OSEM - Open Source Event Management. Lots of Open Source projects have their own conferences and other events. We have our own tool for managing proposals, organising schedules, taking registrations, etc.

Trollolo - Command line tool for Trello with a brilliant name ;)

And many more, many of which can be found on our GitHub project page

Reason 2: We build and use the best tools

You might have noticed a common theme among many of the software projects listed above. We're a Project that cares a lot about using the right tool for the job.
In many cases that means using many of the same tools that other projects use, IRC, Git & GitHub, Forums, Mailing Lists, etc. We're not the kind of project that wants to invent new types of wheels just because we didn't invent the first one, but when the tools that are out there don't do the job well enough, we roll up our sleeves and try to produce the *best* tool for that job.

Obviously, we do this motivated by the need to scratch our own itches, but we also work very hard to produce tools which can, and should, be adopted by a much wider audience than just 'users and contributors to the openSUSE Distributions'.
This is probably best highlighted by talking about two of our best tools, the Open Build Service, and openQA.

Open Build Service - As already mentioned, our build tool builds all of the openSUSE Distributions packages, ISOs, Images, etc.
It also generates and hosts repositories for these packages.
It has always been built as a 'cross-distribution' tool, currently able to build packages for SUSE Linux Enterprise, Arch, Debian, Fedora, Scientific Linux, RHEL, CentOS, Ubuntu and more.
With both a web-based UI (you can see and use our installation HERE) and the osc command-line tool (Documentation HERE) the barrier to entry is pretty low. Even someone with no packaging experience can easily see the contents of any package in OBS, understand how it's put together, branch it (aka fork it to use the github equivalent term) into their own Project and get hacking away.
We also allow our openSUSE instance of OBS to be used by *anyone* without charge, so they can build and host their packages on our infrastructure for free (though at the very least, we'd like it very much if they made sure their openSUSE packages work really nice in return ;) )
OBS is already used by other companies and projects like VideoLAN, Dell, HipChat and ownCloud, but if you're in the business of making software for Linux, you really need to ask yourself the question "Why don't I use OBS to build my packages?"

openQA - Another jewel in our crown and somewhere I've been spending a lot of my time lately. openQA is a testing tool that lets you build tests for any operating system or application. Unlike practically every other testing tool I've looked at, openQA actually tests software the same way *users* do. It looks at the screen, and makes sure the user will see what you expect them to see. Then press the same keyboard and mouse buttons that a user would to work through the application. You can see it in action HERE
openSUSE use openQA extensively to build Tumbleweed. Proposed code for Tumbleweed is tested by openQA before it is accepted, and tested again AFTER it's accepted (to make sure it integrates well with everything else that has been recently merged).
This is how we're able to offer Tumbleweed as a 'stable' Rolling Release, because thanks to openQA we *know* the Distribution works *before* we put any updates in the hands of any users. If something does break, openQA stops the release, and users don't notice any problems, besides perhaps a slight delay in receiving new updates. On the (exceptionally rare) chance that something slips by openQA, that becomes a new test case, so we never make the same mistake twice.
openQA is also being used by SUSE for the testing of SUSE Linux Enterprise, and the Fedora Project have recently started using it. We'd love to see more distributions using it, and while playing around we've already managed to get openQA testing different operating systems like Android and even Windows :)

Reason 3: Engineering matters

Why all this focus on tools? Well for our community, engineering matters. Beyond the tools that we use to make this happen, we spend a lot of time, braincells, and discussions, getting under the hood and trying to find the best way to build our software.

This has led to openSUSE leading the way with lots of innovations in areas like package management. zypper and libsolv mean that, unlike other RPM-based Linux distributions 'dependency-hell' is a thing of the past for us, and we manage to stand toe to toe with Debian and Ubuntu, look at their package managers and go 'yeah, we can do that, and more'

We're also not interested in being another Open Source project taking everything from dozens of upstream projects and either stamping our logo on it and calling it our own, or changing a whole bunch of stuff and not giving stuff back.
We work hard with our upstream projects, be that the Kernel, GNOME, KDE, whatever, to give our feedback, get our patches merged, not just for their benefit, but our own - the more they accept, the less work we'll have to do maintain long winded openSUSE specific patch sets in the Future.

Reason 4: Community, Community, Community

The openSUSE Community is great, so great they deserve a blog post of this length on their own. But for the purposes of this post I want to summarise a few things.

First, a little bit about SUSE. We're sponsored by SUSE, and they make up an important part of the community, employing a good number of openSUSE contributors, and regularly contributing to the openSUSE Project in terms of code, staff, money.
In many respects their the 'Patron' of the openSUSE Project, having our back when we need it, but not overarching overlords controlling the project. The openSUSE Project is independent, free to set it's own direction.
SUSE sees themselves as peers in the community, taking part in the same way as everyone else, encouraging its engineers to be involved in discussions, and help set the direction of the openSUSE Project via code submissions, just like everyone else.
Open Source works best when there is freedom to innovate, that's the business openSUSE is in, and SUSE is a company that understands at it's Core - as their current recruitment campaign says "Open Source is in their Genes"  (P.S. They're Hiring)

So if SUSE aren't 'in charge', who is? Well, the community is. Those who do, decide, and we want more people to help join us doing stuff! We're the kind of project that is very open to new contributors, and we try hard to maintain a very low barrier to entry.
We're not the sort of project with lots of hoops you need to jump through before you're allowed to do anything.
If you see something that needs fixing, talk to the people also working on it, fix it for/with them, and submit your contributions. This is true for code, the website, wiki, marketing stuff, anything, anywhere. We're not a project where you need permission before you can get involved.
We try and keep our processes lean but the ones we do have are on our Wiki, and in terms of getting together with existing contributors, they can most easily be found in our IRC Channels and on our Mailing Lists

If you're a user and need help actually using openSUSE but aren't ready to get involved with contributing to the project yet, then we have a great community of contributors helping support people via IRC and our Forums

But doesn't everyone doing everything everywhere lead to chaos? No, not really. The Open Source model works, with the few processes we have in place providing enough structure and safety that problems get caught and dealt with as part of a natural flow and learning process.
But, it's not utopia, from time to time of course stuff pops up, and that's why the openSUSE project has the Board.
Our job is to help keep this Project on the right track, keep people talking to each other, resolving conflicts, keeping everyone on the same page, and being the 'decision makers of last resort' if there is no one else able/suitable to make a decision that affects the Project.

This basically boils down to being the 'Universal Escalation Point' - If you have a problem, if no one else can help, and if you can find them (and that should be easy), maybe you should contact board@opensuse.org (LINK - ahh, nostalgia)

The structure of the Board is also a nice reflection of how the rest of our community is structured - of the 5 elected Board seats, no more than 2 can be occupied by people employed/controlled by the same organisation. This means no company (not even SUSE) can have a majority of the seats on the Board, ensuring it's healthy and balanced just like the rest of our community.

And finally, the openSUSE community is just a great bunch of people. I've been contributing to this project since it started, and it's dragged me in and is as much fun as ever. We have great, smart, funny, dedicated people, and it's a pleasure working with them both as part of my day job and hobby.

A big part of that is our regular openSUSE Conferences. 2011 was my first, and the wonderful combination of meeting people, great talks, and (lots of) openSUSE Beer, enabled me to jump from being a 'casual contributor' on the fringes to someone who really felt part of this great community.

We have another openSUSE Conference coming up in May, in Den Haag/Netherlands. There's still time to Submit talks (we'd love to see more from other Projects) or Register. If this post has got you interested in openSUSE, come along, it's a great opportunity to meet us :)

And remember our motto and default /etc/motd.. Have a Lot of Fun!

Qactus, a Qt-based OBS notifier, is out in the wild. Version 0.4.0 is the first release.
I started it a long time ago together with Sivan Greenberg as a personal project for learning Qt. And now it’s back into life

It features
– Build status viewer
– Row editor with autocompleter for project, package, repository and arch
– Submit requests viewer
– Bugs

This application is possible thanks to Marcus ‘darix’ Rueckert. He has helped me getting further knowledge of the OBS API.

I think this version is usable. So, why don’t you give it a try?
The source code is hosted on GitHub.

There’s a new build published today for the AMD FGLRX drivers.

It include the new patch made by Sebastian Siebert supporting Kernel 3.19x series you could have on 13.1, 13.2 and Tumbleweed openSUSE distribution.

The server just got the new rpms, so you should be able to update with zypper ref -f && zypper up

Have fun.

In this article I would like to describe how pass variables(configuration) to application and core of libgreattao. Writing about core I have in mind the main libgreattao librarary – without backends. Core configuration can be separated into shell configuration and window classes configuration.

Each sub chapter I begun with describing how invoke application, giving t o it application’s configuration, shell’s configuration and window classes configuration.

Configuration can went from files or command line. Additionally, application’s configuration can went from arguments(command line) after three additional decorators and additionally from configuration window. Configuration window is automatically generatedby libgreattao. In future I would like to add configuration coming from network. I would like to add to libgreattao web serwer and wrote client. Of course, network source will be less important, so there’s no additional risk to application.

Application domain

Configuration from file


./our_program --tao-app-config-from-file /path/to/first/file,/path/to/second/file

Libgreattao will read configuration from files in giving order. Files read in future will override file read in past. There’s the same rules than in user’s file for specified application for selecting user’s application directory(~/.tao/apps), so options are assembled with spaces, name, equal signs and value. Name and value must be divided by equal sign not preceded with backslash character. Using of escape characters are allowed, like ‘\ ‘, ‘\n’, ‘\=’. Using sentences are allowed too.

Configuration from command line – line


./our_program --tao-app-config-from-line first_variable=first_value,second_variable=second_val

In this case application’s parameters are separated by commas. Each parameter are assembled by variable(variable’s name) and value.

Configuration from command line – tao-switch


./our_program --tao-app-config-tao-switch-first-option first_value --tao-app-config-tao-switch-second-option second_value

In this case you should use for all option –tao-app-config-tao-switch- and name of option. You should put value as next argument.

Long option – equivalent to getopts syntax


./our_program --first-option first_value --second-option second_value

It’s equivalent to long option getopts syntax. We here put two dash and next name of option.You should put option’s value as next argument. Difference between this and previous method is in handling way. In this case option with value are not removed from arguments list(char **argv, char *argv[]). As you can guess, in previous case, we remove option decorator and value from arguments list. It’s indented, because this syntax is also used by getopts and application can support custom options, so we cannot remove it on libgreattao_initialize.

Character-switch – (not) equivalent to getopts syntax


./our_program -a first_value -b second_value

In this case it looks like previous way. Decorator and value are not removed from list. As getopts short arguments, you can puts dash and letter as decorator. In difference to getopts you cannot put this characters together and after each decorator you must put value.

Reading order

In first position libgreattao tries to read files. As I explain above, files are read in order as given, so values in files read in future will override, what was read in past. In second step command line are read. In last step window with options are generated. For each command-line parameter application can select decorators for each option, so libgreattao can search for tao-switches, long switches, short switches. Tao command line switch(there, where you put options separated by commas) are always enabled to command-line source. Order for command line switches are described below:

• tao-switches option
• Long decorators
• Characters(short) decorators
• Tao line

For each option used decorators are determined by used flags for option. Values read in future overlapped values read in past.

Arguments querying


#include <libgreattao/tao.h>
#include <libgreattao/config.h>

#include <stdlib.h>
#include <string.h>

const struct tao_app_config_option options[4] =
{
{"a", "a", "a", TAO_TYPE_STRING, 'a', TAO_CONFIG_OPTION_ARG_TAO_DECORATOR | TAO_CONFIG_OPTION_ARG_LONG_DECORATOR | TAO_CONFIG_OPTION_ARG_CHARACTER_DECORATOR},
{"b", "b", "b", TAO_TYPE_INT, 'b', TAO_CONFIG_OPTION_ARG_TAO_DECORATOR | TAO_CONFIG_OPTION_ARG_LONG_DECORATOR | TAO_CONFIG_OPTION_ARG_CHARACTER_DECORATOR},
{"c", "c", "c", TAO_TYPE_UINT, 'c', TAO_CONFIG_OPTION_ARG_TAO_DECORATOR | TAO_CONFIG_OPTION_ARG_LONG_DECORATOR | TAO_CONFIG_OPTION_ARG_CHARACTER_DECORATOR},
{"file", "file", "file", TAO_TYPE_FILE_PATH, 0, 0}
};

struct tao_app_config_result results[4];

void close_window(void *mesh, void *mesh2)
{
tao_close();
}

int main(int argc, char **argv)
{
char *buffer;
void *message_window;
int length;
int number;

tao_initialize("tao options demonstration", "", &argc, argv);

tao_get_app_configuration(options, results, 4, TAO_APP_CONFIG_FROM_LINE | TAO_APP_CONFIG_FROM_FILE) ;

buffer = (char *) malloc(sizeof("Error getting string\n") - 1
+ sizeof("Error getting integer\n") - 1
+ sizeof("Error getting unsigned\n") - 1
+ sizeof("Error getting file path"));

buffer[0] = '';
if (results[0].error) {
strcat(buffer, "Error getting string\n");
}
if (results[1].error) {
strcat(buffer, "Error getting integer\n");
}
if (results[2].error) {
strcat(buffer, "Error getting unsigned\n");
}
if (results[3].error) {
strcat(buffer, "Error getting file path");
}

if (buffer[0] != '') {

message_window = tao_new_window("/desktop/dialogs/message_dialog");
tao_add_handler(message_window, "/message", NULL, 0);
tao_add_handler(message_window, "/:abort", close_window, 0);
tao_set_hint(message_window, "/message", HINT_NAME, buffer);
tao_handle_events();
tao_release_window(message_window);
}

length = sizeof("\"\" \"\"\n");
length += 2; /* Two numbers */
if (!results[0].error) {

length += strlen(results[0].result);
}
else {
length += sizeof("(null)") - 1;
}
if (!results[1].error) {

number = (int) results[1].result;
while (number /= 10) {
++length;
};
}
if (!results[2].error) {

number = (int) results[2].result;
while (number /= 10) {
++length;
}
}
if (!results[3].error) {

length += strlen(results[3].result);
}
else {
length += sizeof("(null)") - 1;
}

if (length > (sizeof("Error getting string\n") - 1
+ sizeof("Error getting integer\n") - 1
+ sizeof("Error getting unsigned\n") - 1
+ sizeof("Error getting file path"))) {

free(buffer);
buffer = (char *)malloc(sizeof(*buffer) * (length+1));
}
snprintf(buffer, length+1, "\"%s\" %d %d \"%s\"\n",results[0].result,results[1].result,results[2].result,results[3].result);
if (buffer[0] != '') {

message_window = tao_new_window("/desktop/dialogs/message_dialog");
tao_add_handler(message_window, "/message", NULL, 0);
tao_add_handler(message_window, "/:abort", close_window, 0);
tao_set_hint(message_window, "/message", HINT_NAME, buffer);
tao_handle_events();
tao_release_window(message_window);
}
free(buffer);
}

This is simple program, which obtain options and next displaying, which options are not passed and next displaying all option’s values. Most important is tao_get_app_configuration. It have pointer for option description as first argument. This structures are not modified by function. As next argument, there’s pointer for structure of results, which will be modified by this function. Program must result structures by 0 before calling this function. As next argument we have count of parameters and results structures. The numbers of parameter structures and result structures must match. As next argument we have flags. We can use these flags:

• TAO_APP_CONFIG_FROM_FILE
• TAO_APP_CONFIG_FROM_LINE
• TAO_APP_CONFIG_INTERACTIVE
• TAO_APP_CONFIG_INTERACTIVE_FORCE

First flag makes reading from files. Configuration came from files are read in libgrattao_initialize method – not in tao_get_app_config. Next flag makes reading form command line. Last two flags makes libgrattao display configuration window. The difference is flag with suffix _FORCE causes displaying all options – filled options too. Additional difference is flag without this suffix don’t display configuration window in case, when all options are filled.

In this paragraph I will describe option description structure. First field of this structure is name of option. It is used to add suffix of switches in command line and it’s also used to searching in command line. Second field is name displayed to user in configuration window. Third field is description displayed to user in configuration window. Fourth filed is option type. In example I uses all allowed types, like string, natural number, integer and file’s path. File path are handled in the same way as string, but for file path(in config window) there’s displayed additional button, which allows to select file. Next filed is character, which are used to short-switches(character decorators). Last argument are flags, which allows to select decorators allowed for current option.

Explained program don’t uses configuration window. To change this you should add flag to last argument of tao_get_app_configuration. It makes this function not returning, while passed data are incorrect. Configuration window guards to all option are correct and filled. Only in case, when user close configuration window, this function assign error to all options.

Shell configuration


./our_program --tao-shell-config-from-file /path/for/configuration/file/1,/path/for/configuration/file/2

It’s similar to application configuration. Files are parsed in giving order and have similar syntax as application configuration files.
Second case:

./our_program --tao-shell-config-from-line first_option=first_value,second_option=second_value

It’s similar to application configuration too

Of course, options given in command line are more important from that reading from files.

I will now show, how uses this data.

=fromconfig variable_name configuration_option_name

This command writes to variable with name given as first argument configuration option with name given as second argument.

Windows classes configuration


./our_program --tao-design-config-from-file /path/for/configuration/file/1,/path/for/configuration/file/2

It’s similar to application configuration. Files are parsed in giving order and have similar syntax as application configuration files.
Second case:

./our_program --tao-design-config-from-line first_option=first_value,second_option=second_value

This is similar as configuration of application.
Now, I will show how use this:

<root>

<fromconfig name="instance_id" configuration="instance_id" />
<label>
<attr-connect name="label" variable="instance_id" />
</label>
<label dir-for="label" path="/message" />
<label dir-for="description" path="/message" />

<hbox>
<template path="/actions/*">
<button label="ok" dir-for="event,label">
<attr-connect name="label" function="last-dir" />
</button>
</template>
</hbox>

</root>

We are using here fromconfig element. Name means variable name, which will be created/overwritten. Configuration is configuration option’s name of window classes. In next step we assign variable’s value to label element.

That’s all!