Attempting to calculate this expression:
0.43mJ / 25fs / ((0.0109mm)/2)^2
crashes the whole program on my 64 bit Windows 10. I couldn't locate the log files, if they exists.

Hi

I'm experiencing glitching in qalculate-gtk 2.2.0, but I also had the problem with 2.0.0.

Since video explains it much better than words, I've attached a small clip.
qalc.zip

I've just compiled libqalculate and gtk from the website, so it is the newest.

Let me know how I can help solve it.

Regards,
Claus

error messgae:

interface.cc:812:103: error: cannot pass objects of non-trivially-copyable type ‘class std::basic_string<char>’ through ‘...’
     gtk_list_store_insert_with_values(historystore, &history_iter, -1, 0, inhistory[i], 1, i, 4, 0, -1);

I think it should be changed from

gtk_list_store_insert_with_values(historystore, &history_iter, -1, 0, inhistory[i], 1, i, 4, 0, -1);

to:

gtk_list_store_insert_with_values(historystore, &history_iter, -1, 0, inhistory[i].c_str(), 1, i, 4, 0, -1);

C functions are not acquainted with C++ structures. My gcc version is 4.8.5 (Ubuntu 4.8.5-4ubuntu2)

GDK_AVAILABLE_IN_ALL void gtk_list_store_insert_with_values  (GtkListStore *list_store,
                                                              GtkTreeIter  *iter,
                                                              gint          position,
                                                              ...);

In qalculate's GUI, I'd like to see these suggestions while typing:

1. magnetic_constant: When typing μ (the greek letter) or mu, please suggest completion to magnetic_constant. Also, it would be nice to call it vacuum permeability since that name is more common in some fields.

2. electric_constant: When typing ε (the greek letter) or eps, please suggest completion to electric_constant. Also, it would be nice to call it vacuum permittivity since that name is more common in some fields.

When I try to convert between °C and °F units get wrongly detected as coulomb/farad.

Can this be fixed so that it can also support feet/inch notation like 6' + 1" ?

I have install qalculate by snap and i view only squares.

Thanxs

Quand on fait: Functions → Économie → Finance → ...
La première fois le menu s'affiche bien.
Mais si on sort de ce menu et qu'on y retourne, le menu ne s'affiche plus.
Sorry, I speak (write) french.
Yann48

A really big THANK YOU! for bringing Qalculate to Windows. This has been the single one piece of software I've been missing the most at work, where I have to use Windows. Now I can use the best calculator ever there as well. Great!

Now, with people using smartphones more often than PCs these days, it would be great to have a port of Qalculate on Android.

Hello
I can't compile callbacks.cc in qalculate-gtk-1.0.0. Please see attached files. I didn't have any problem with version 0.9.9
Thanks in advance!
Errors.txt
gcc version.txt

E.g.:

> 0x823445d

  1.3652899E8 = approx. 1.3652899E8

There's no point to use scientific notation in the example, because the answer is 136528989.

Here's a general rule: if scientific notation makes amount of symbols in the result bigger than the usual print, it should not be used.

E.g. 1000 vs 1e3 — it is justified, 4 vs 3 symbols. 10 vs 1e1 — it is not justified, scientific notation makes it longer.

And why the result is even approximated, I can't get an approximate file size!

thanks much, just discovered this gem, seems to have everything that i need, one thing that bugs me, is there any way to display result with thousand separators? millions and above just makes me cross-eyed, anyway to remedy this basic ui requirement?

I prefer RPN mode and started using Qalculate on my PC where it works fine. Installed it on my laptop but I can't use it because I don't have a numeric keypad so when you press the operator, e.g. +, it just echos the character behind the number.

On Arch Linux with libqalculate 2.0.0-1 and qalculate-gtk 2.0.0-1.

Mac OSX - High Sierra, calculate-gtk is a MacPort installed in /opt/local/bin. When I try to launch qalcuate-gtk, this is what happens...

bin $ sudo ./qalculate-gtk
Password:

(qalculate-gtk:46686): Gtk-WARNING **: cannot open display:

[1]+ Stopped sudo env PATH=$PATH ./qalculate-gtk

Ok, so is it working? The answer is yes, since I can get help.

bin $ sudo ./qalculate-gtk --help
Usage:
qalculate-gtk [OPTION…] EXPRESSION

Help Options:
-h, --help Show help options
--help-all Show all help options
--help-gapplication Show GApplication options
--help-gtk Show GTK+ Options

Application Options:
-n, --new-instance Start a new instance of the application
--display=DISPLAY X display to use

But, how do I specify the display or even list what displays are available?

When going to File->New->Matrix, selecting a table cell in "Elements" and using tab to cycle through them, when using tab on a cell in the last row, cursor jumps to first row and quickly cycles through them, rendering the GUI unusable.
Same goes for the File->New->Matrix menu

Hi.
Thanks for updating my favorite calculator :)

However, after compiling and installing on Opensuse 42.1, qalculate-gtk produces the error in the headline here and halts.

I have attached a 'strace' of launching qalculate-gtk (not the full strace thought). I hope this helps.
Let me know if I can provide anything else.

/C

mprotect(0x7f0f8a123000, 4096, PROT_READ) = 0 lseek(10, 0, SEEK_SET) = 0 read(10, "\211PNG\r\n\32\n\0\0\0\rIHDR\0\0\0000\0\0\0000\10\6\0\0\0W\2\371"..., 4096) = 3874 futex(0x7f0f937924a8, FUTEX_WAKE_PRIVATE, 2147483647) = 0 futex(0x7f0f937924a8, FUTEX_WAKE_PRIVATE, 2147483647) = 0 futex(0x7f0f937924a8, FUTEX_WAKE_PRIVATE, 2147483647) = 0 close(10) = 0 munmap(0x7f0f95d0a000, 4096) = 0 open("/usr/share/locale-langpack/en_US.UTF-8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en_US.UTF-8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-bundle/en_US.UTF-8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-langpack/en_US.utf8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en_US.utf8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-bundle/en_US.utf8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-langpack/en_US/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en_US/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-bundle/en_US/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-langpack/en.UTF-8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en.UTF-8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-bundle/en.UTF-8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-langpack/en.utf8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en.utf8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-bundle/en.utf8/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-langpack/en/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale/en/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/usr/share/locale-bundle/en/LC_MESSAGES/libqalculate.mo", O_RDONLY) = -1 ENOENT (No such file or directory) open("/dev/tty", O_RDWR|O_NOCTTY|O_NONBLOCK) = 10 writev(10, [{"*** Error in", 14}, {"qalculate-gtk", 13}, {"': ", 3}, {"corrupted double-linked list", 28}, {": 0x", 4}, {"0000000001fca570", 16}, {" _\n", 5}], 7_ Error in qalculate-gtk': corrupted double-linked list: 0x0000000001fca570 *** ) = 83 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f0f95d0a000 futex(0x7f0f91c9b620, FUTEX_WAIT_PRIVATE, 2, NULL

Hello, i'm enjoying the vast amount of features in Qalculate! what a great program.

i have a request to make, can you modify the currency list so that all countries have their currency units ? as it is only a few have their full definition(Indian Rupee, Danish Kroner, Swedish Krona, etc) while the majority of currencies are listed by the countries they are used in(Algeria, Benin, Cuba, etc) or alternately could you show me how to modify this information myself ?

Thank you and keep up the great work on Qalculate.

Hi, Long time user of Qalculate here.

A recent change in Qalculate makes launching only a single instance of the GUI by default.

However, in my usage, I find that some times there is a need to launch multiple Qalculate instances.

Though there is a switch to bring back the old behaviour, Qalculate is almost always launched using 'Application Shortcuts', 'Alt+F2', 'Application Menus', etc. making it difficult to provide that additional switch.

Can you please switch back the DEFAULT behaviour as it was earlier and make the additional switch necessary only in case a single instance is required.

Thanks.

1/2.5e-6 which should give 400k. it interprets it as 1/25x10-6 yielding 40k.

see attached photo

I'm using the sudo apt install version that goes with ubuntu 16.04

Hey, I think it would be convenient if one could bring up old inputs from the history in the input box by pressing arrow up/down, the same way it works in most shells :)

Cheers!

If the angle is not in radians, the result is always just ≈ sin(<angle in radians>) (the same is true for cos() and tan()). To get the actual result, I have to specify the angle in radians.

Is there a setting to turn this off, or is this a bug? It used to work just fine in older (< 1.0.0) versions.

In spite of having installed gnuplot 5.2.2 64-bit in Windows 8, the option Plot Functions/Data appears greyed out. How can I activate this option?
Any help will be greatly appreciated
Best regards

When I type "42 USD to GBP", there's a blocking popup saying Fetching exchange rates.

It'd be really nice if, instead, we just got the result right away using cached/possibly stale exchange rates, but with a notice that Exchange rates are being updated in the background, sit tight and your result will be updated once they're fetched.

When compiling qalculate-gtk 0.9.11 with g++-4.8.5 (openSUSE Leap 42.2), I get the following error message:

interface.cc: In function ‘void create_main_window()’:
interface.cc:812:103: error: cannot pass objects of non-trivially-copyable type ‘class std::basic_string<char>’ through ‘...’
gtk_list_store_insert_with_values(historystore, &history_iter, -1, 0, inhistory[i], 1, i, 4, 0, -1);
^
Makefile:482: recipe for target 'interface.o' failed
make[2]: *** [interface.o] Error 1

I'm no programmer at all, so I googled around and got the code compiling by changing interface.cc line 812:
gtk_list_store_insert_with_values(historystore, &history_iter, -1, 0, inhistory[i], 1, i, 4, 0, -1);
to
gtk_list_store_insert_with_values(historystore, &history_iter, -1, 0, inhistory[i].c_str(), 1, i, 4, 0, -1);

It compiles with g++-4.8.5 and g++-6.2.1

I noticed issue, which may be either pros or cons, depending on what you want. I have qalculate 2.2.0 with gnuplot. When it renders specified x range, it seems to warp image to fit into square, and this leads to different units for axis. To better understand, what is it about, try to render something like 1/x - or anything with infinity 'y' in selected range. Image gets stretched vertically so, that while x has units like 1,2,3, 'y' will have like '0,50,100, for some functions even 0,5000,10000 and more.

It also seems to depend on selected points density - the more points you use - the more part of y axis will be occupied and took into account by stretch.

I found this discussion, which seems to be common gnuplot problem:
https://stackoverflow.com/questions/38616925/how-to-determine-axes-aspect-ration-in-gnuplot#38631837

Plotting by stepping only x values has one problem, that fidelity is limited - when i set too high values (like 0,0005 as step size with even -8..8 range), it gives error about too high number of points.

However, it could use somewhat more robust way for drawing:

• instead of watching for only dx, it could watch for sqrt(dx^2+xy^2). There may be two ways: either get coordinates for each next point by this formula, or just step in x, but check distance to new point, and if too high, just insert point between them.

This may have problem due to another issue: y range is unlimited, and when combined with this way, could cause qalculate to just fail on places wiht infinite 'y'.

1. Who, if anyone, is maintaining qalculate-kde
2. In any case, clicking on "Mode" in the title bar creates a pop-up showing various items: "Number Base", "Numerical Display", ... However hovering over or clicking on an item doesn't show the menu of choices for the item, so it can't be set; the pop-up works correctly for other title bar choices. Editing the corresponding item in ~/.qalculate/qalculate-kde.cfg has the desired effect, so I have been able to use qalculate-kde successfully, for at least a few calculations.

Dear upstream,

I was trying/fiddling around with qalculate and came across the config file in ~/.config/qalculate where this is what is says about getting updated exchange rates -

~/.config/qalculate$ cat qalculate-gtk.cfg | grep auto_update_exchange_rates
auto_update_exchange_rates=-1

This is somewhat related to #49 but not really.

I compared lines look in qalculate-gtk plot with raw gnuplot. Difference appears both for wxt and qt.
In raw gnuplot, by default, lines have width=1, and when zoomed out, don't lose quality. But in qalculate-gtk lines are initially thicker - nearly 2 pix. When zoomed out, some aliasing appears, and visually line begin to seem thicker, if not really. It is obvious, that there is different style. But when i tried to tune styles in gnuplot, "$ set style line 1 linewidth 3.0 pointsize 3.0 color '#444444' " did not work, so i don't know, how to reproduce it. However, i can ran "test" to see current style, and i can say, it is better.

As variant, there may be universal way to setup gnuplot not only for look, by running custom gnuplot commands, for experts.

These functions work the opposite way as expected.
From Hex(10) I would expect 0xA to be the result and not 16. Like for sin() where I expect the sine as result I would expect hex() has the hex-digit as result

Currently plotting functions like 1/x, arcsin, arccos or tg causes plot to be stretched vertically , to fit in square window.

One of your screenshots has the calculator in a night theme, i was wondering how to apply that as i can't find the option in the menu.

I was expecting the exchange rates to be put in either ~/.config/qalculate or better yet ~/.cache as can be seen at https://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html . It does seem you do honor the freedesktop specs so would make sense to have that smallish database which would constantly change therein or have a directory if more than one datasets would be volatile

It would be great to have a portable version of Qalculate for Windows:

• No installation, simply unzip and run
• Settings are saved to the folder with executable and neither Windows registry nor system/user account folders are used

Using v2.2.0 snap package in Ubuntu 17.10 with Wayland gives the following error:
(qalculate-gtk:1883): Gtk-WARNING **: cannot open display: :0

qalculate-gtk, v0.9.9 :
The tool tip of the STO key says "Delete", which is probably wrong, should be "Store to variable" or smtg similar.

Whenever Qualculate 0.9.9 gains focus, it seems to put the current input field contents into X11's primary selection (middle click buffer). This makes transferring copy and pasting via middle click pretty much impossible.

When i try to build elliptise as second order curve, supplying it as two functions for lower/upper arc, even with samples 50000 ends appear like straight.

While doing in gnuplot:

gnuplot> set samples 10000
gnuplot> plot -sqrt(2-(x**2)/2)-2, sqrt(2-(x**2)/2)-2

I get clear arc (setting samples to 100000 almost makes invisible gaps in end, though also slowdowns gnuplot itself).

Upd. When building qalcualte data with external gnuplot, it is as amazing, as just building funcs in gnuplot:

Compiling on Mac (Yosemite; likely for other flavors):

In file included from callbacks.cc:27: In file included from ./callbacks.h:27: In file included from ./main.h:33: In file included from /opt/local/include/libqalculate/qalculate.h:15: In file included from /opt/local/include/libqalculate/includes.h:21: In file included from /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/../include/c++/v1/vector:265: In file included from /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/../include/c++/v1/__bit_reference:15: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/../include/c++/v1/algorithm:702:71: error: invalid operands to binary expression ('const tree_struct' and 'const tree_struct') bool operator()(const _T1& __x, const _T1& __y) const {return __x < __y;} ~~~ ^ ~~~

After start and after resize it is always ok. But when mouse hovers history item, each redraw showes area from much lower areas. Thus, to get it more clear, it is better to move pointer down. For example:

• Entering item nearly at middle (from side), it drawes item separator in the middle.
  • while moving towards upper separator, bold result appears under above separator
  • while moving towards lower separator, formula from lower history item appears above this border.

It seems, that redraw is broken rather on resize, creation, or other cases, requiring complete redraw, while on hover it is ok. Because when i move pointer to one text, and another one appears under pointer, then if i paste it by double click, appeared text gets into entry, instead of what was before.

Last successful build was from 94f91d5.
When i detected issue, i tried to rebuild latest code, but it fails, probably after commit from yesterday:
with errors: error: ‘recursiveMakeDir’ was not declared in this scope
at main.cc:347:34, callbacks.cc:11687:34, callbacks.cc:11710:26, callbacks.cc:12970:32

Hi, I experience the above issue when I use any calculation x V/y Ai get W/A² which is not wrong (I assume haven't actually checked) but not common and rather confusing since ohm is the common unit, but this is only shown when I force it with to ohm.
However, I'm not shure if this bug belongs here or into the lib repo. I'm using Qalculate V0.9.7 on Ubuntu 16.10.

When defining a new variable (Edit menu -> Manage Variables option -> New button -OR- File menu -> New submenu -> Variable option), if you click the Names Details button, the program crashes with a segmentation fault.

I'm using the v2.1.0 tagged branches of both libqalculate and qalculate-gtk on Ubuntu 16.04.

Most (all?) RPN calculators allow you to press enter/return to push the current value to the RPN stack
which duplicates the value in top 2 stack positions. Qalculate instead pushes a 0 to the top of the stack which precludes many of the normal use cases of an RPN calculator. You can use the Copy function but that is inconvenient and awkward.

Please consider changing this so that Qalculate works like other RPN calculators.

Sometimes, when dealing with complex matrices, there is a need of a lot of vertical space in the output. In that moment I don't need the keypad at all, so it would be nice to be able to give more vertical space to the output, removing it from the keypad area.
Thank you.

I prefer create separate issues, so that if some is done, it may be closed without waiting others.

There i propose to utilize x11 terminal, which is default, when gnuplot is started by command line. It is not smoothed, but lines are printed properly, to not look like with unexpectedly failed smoothing.
And closely reason, to proposed it, is that it allowes to see, what is out of initially specified window scope.
Also, when i tried in this mode, combining of range -10..10 with sampe rate 1000 did not even make it slow, unlike current variant, used in qalculate.

Thus, there may be choice between available backends, each with its trong sides (really infinite range for x11 and antialiasing/alpha for pngcairo). This of course, should be disabled with non-x11 systems, like win, mac, or simply wayland env.

The tooltips in fields of the Percentage Calculator show "Binary value", "Octal value" and such. This seems to be a leftover from the "Number Bases" Tool.

Dear maintainer

Something has been bugging me about Qalculate(which is amazing on ALL other aspects, much great, such wow! many congrats!). It always prefers s^-1 in stead of Hz, as a Telecommunication Engineer, this really is a pet peeve for me. Also, the UI used to have a button next to the result where you could choose your unit, but that's gone and I really miss it, could you please give it back or set some option in the preferences for this? That and/or adding support for Hz if at all possible?

Thanks!

During play with raw gnuplot i tried pngcairo terminal type.
With command "set term pngcairo transparent" - it is possible to get fully antialiased, yet transparent image, which then may be placed in plot window over default gtk background, specified by theme.

Currently rendering result is not smoothed, looking weird.

upd, p.s.: What is current term mode used? Comparing with x11, it makes me to thing, that most trouble in current backend is in drawing. I suppose, pngcairo should be much faster. Though i'm not sure about non-linux platforms. Need to test pngcairo speed eventually :) .

I got gnuplot crash due to double free, which happened, when i tried to get normal 1x1 aspect with usable scale, after entering cos(x)^x, without setting y limits. During zoom window disappeared, and did not appear on following update requests, until qalculate-gtk restarted.

Right now, when I try to open a second qalculate it will just give focus to the first, but having multiple independent instances would be much more useful to me instead of forcing a global one.