some changes 2

[ghost]

Hello William,

I had more free time these days and did some changes to screenfetch-c, will list the changes again. I don't have internet at home, so execuse me if this pull request overrides any previous changes.

colors.h - edited the manifest constant TNRM in order to prevent any issues like the one with the arch logo being showed with bright/bold colour when darker one was substituted.

detect.h - the detect_distro prototype/declaration now takes one formal parameter instead two, the second one which we was using while ago for the "user@hostname" colour is now passed as argument to the function main_ascii_output located in disp.c . This way the user can pass the -D flag/option/command line argument and print different logo, and the "user@hostname" will be showed with the same colour as the passed distro string or detected distro one.

plat/linux/detect.c - The second formal parameter/argument to the detect_distro function header was removed, same goes for all sprintf(str2, "%s", MANIFEST_CONSTANT_COLOUR); function calls. detect_host now fills the passed char arrays separately. detect_uptime variables secs, mins, hrs, days was changed to unsigned int instead signed int types, same goes for their formatting specifiers used by snprintf(); function calls. detect_pkgs is using unsigned short int instead signed int type. detect_cpu, the 'end' variable is now of type size_t instead int, because strlen() returns the string length of size_t type, the 'for' loop variable 'i' declaration and initialization was declared right below size_t end; also it's type was changed to unsigned short int instead signed int. detect_disk, the function calls to snprintf() was using wrong formatting specifiers of signed long int while the passed variables was declared as unsigned long int, so had to be corrected with "%lu" instead "%ld". detect_mem total_mem, free_mem, used_mem variables are now of type unsigned long int instead long long int, thus making the program 32-bit compatible, their cast type was also exchanged to accomodate the new int type, used_mem now substracts the 'bufferram' and 'sharedram' thus displaying smaller ram usage, the function call to snprintf() is now using newer formatting specifiers to accomadate the new variables int type. detect_res, the variables 'width' and 'height' are now of type unsigned short int instead signed int, the snprintf() formatting specifiers was also exchanged.

util.c - the split_uptime function header was edited, so the passed arguments/parameters will be of type unsigned int instead signed int, same goes for the cast types in it's function body for the pointer variables. Added header file, and edited the take_screenshot function body so it can execute the external screenshot utility either for Macintosh or GNU/Linux with the following formatted time: "screenfetch-%Y-%m-%d-%H%M%S.png", the strftime function handles the date formating before the whole char array is passed to the sysmte function.

util.h - the split_uptime function prototype was edited to accomodate the newer int type.

Wrote header file "logos_length.h" which contains manifest constants/macros used to control the 'for' loop iteration cycles in some of the disp.c functions, the dragonfly bsd logo was with wrong length causing the program to terminate with dumped core.

disp.c - comment in display_verbose function because it was merged into main_text_output function. process_logo_only and process_data are using the manifest constants from logos_length.h header file, the function header process_data was edited to accomodate the newer changes which include: passed colour from the detected distro or the one set with -D flag/option/command line arg, detected user and host from the plat/linux/detected.c module, user_host character array that will be colorized with the passed colour and it will combine 'user' and 'host' arrays into it. main_text_output function header now takes three additional formal parameters/arguments which we will be used to fill the user_host char array, but this time it won't be colorized.

disp.h - same description as the edited function headers in disp.c, but this time the function prototypes/declarations was edited to accomodate the newer changes.

logos.c - fixed some of the char array logos to be with the real char array size as the ones in logos_length.h, later on saw that some of the printed logos indentation was more or less and fixed those logos. DragonFly BSD is still printed with abnormal indentation.

main.c - replaced host_colour char array with two other char arrays - 'UseR' and 'HosT' which will be filled by detect.c and more especially the detect_host function. detected_arr_names will print the semi-colin with white colour thus making it distinctive from the previous colourized text. Removed the actual argument host_colour that was passed to some of the function calls, and/or replaced it with the UseR, HosT char arrays. display_verbose was renamed to main_text_output which now takes bool type argument/parameter to check whether we are in verbose mode.

This pull request is result of 3 days work, and I described only the first one here. On the second day decided that the program will use only structures but on third day saw that it would break the compatibility with the others plat/*/detect.c modules, nor I have to mention that it would require a lot time to accomodate the structure syntax in those modules, and on third day took the copy of my first day work and did other changes, so my second day work was pretty much wasted time. The most notable change is in disp.c and I'm sure you'll like it, and probably inspire you to add more distros.

I've wrote nice shell version processing function, so take advantage of it and add support for moar shells.

I really want to contribute more often but I don't have internet nor I could test the other plat/*/detect.c modules, as I have a lot experience with OpenBSD

In *BSD the header file sys/sysctl.h includes few sysctl functions, replace all popen calls.

detect_mem (BSD sysctl): total - vm.stats.vm.v_page_count free - vm.stats.vm.v_free_count inactive - vm.stats.vm.v_inactive_count

If you want to add support for systems average load:

include

double x[3];

getloadavg(x, 3);

string printf "%.2f %.2f %.2f" x[0], x[1], x[2]

Best regards, Aaron Caffrey

[woodruffw]

Thanks again!

There are a few conflicts/style changes in this one, so I'll have it merged it a few days.

Best, William

[woodruffw]

Also, as a side note: You might want to try sending your changes as individual commits, instead of as one giant commit. It makes it easier for me to track changes, which allows me to merge faster.

Best, William

[ghost]

I wish I could this but the git client on my tablet lacks a lot essential options.

All development is done at home and all changes are pushed via my tablet at the nearest hotspot which is 20 miles away from home.

Edit: for some reason github attached my latest commit to this pull request, so you dont have to copy my screenfetch-c fork files manually.

[woodruffw]

@wifiextender:

I'm not going to be able to merge your last two commits into the master branch because of massive changes to structure and style you've made. The changes have introduced too many differences for me to keep track of in a timely manner, and I don't have much free time on my hands anymore. I'm going to try to include your changes through a series of commits (you'll be credited in each of them), but I can't guarantee that I'll include everything.

Best, William

[ghost]

The program is 99% ready, all strings that will be translated was added in en_gb.po, all detect.c modules are now up-to-date, merged logos.c in arrays.c, added screenshot code based on jpeglib, so take_screenshot doesn't have to depend on 3rd party programs, the bootstrap script was improved so it does moves the bash scripts when the user installs the program, also mac and solaris tests was added in it.

Best regards, Aaron Caffrey

[ghost]

Hello William,

The fifth commit brings fallback mechanisms regarding Solaris and GNU/Linux that allows them to compile the program with and without X11 and OpenGL headers, so those running servers can enjoy the program without the need to install those libraries in order to compile it flawlessly.

The bash script detectwm was using repeateable code and some of the case labels was trying to match WMs that was not present in the 'wmnames' array. Removed the repeatable code and added more WMs.

Edited the README file, so it includes infor mation regarding the jpeglib.h library and how to install it, so the users can benefit from the custom screenshot creation and sav ing code in 'take_screenshot()' function.

x11.m4 now have fallback test if X11 headers are not installed and if the user runs GNU/Linux or Solaris the test will check for the presence of lspci which is used by detectgpu script. If it's missing too 'configure' will terminate with pre-defined error message letting the user know that they have to install X11 and/or lspci in order to compile the program. xdpyinfo is checked for presence too but it is not mandatory as lspci. Same goes for their detect.c modules, they will fallback to the bash scripts if the program is compiled without X11 headers. GNU/Linux will fallback to detectgpu if OpenGL headers are not installed too.

The detection functions headers and their prototypes was deparameterized, those credits goes to you.

Will wait till you push all changes from my screenfetch-c fork, as I have wrote some huge functions that can go in util.c, for example is the hybrid between xml and config parser that depends entirely on the standard C libraries, I'm talking about the manual mode and the configu ration file that is created and read afterwards. There is lot more code to be revealed and pushed, but I don't want to put you off with my huge commits as the master branch have to catch up nearly 4000+ changes.

Best regards, Aaron Caffrey

[woodruffw]

Hey Aaron,

I'm working on merging your new build system into my (current) source tree. After that, I'll be merging your changes to take_screenshot and then the internationalization.

By the way, you don't need to take screenshots if it's an inconvenience. Your descriptions are fine, but individual commits are nicer (I understand if your client can't do this, though).

Best, William

[woodruffw]

Also, for future reference, please use tabs (not spaces) for indentation.

[woodruffw]

Closing this due to inactivity; it's significantly out-of-date with the current tree and manually resolving it would produce relatively little benefit at this point.