The following crash happened recently several times without noticeable reason:
It's Geany 1.25 on Windows, Java 1.7
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795
A few questions:
- What version of Windows is that? - Is it the Geany installer that comes with GTK+ bundled or did you provide own GTK+ (if so, what version)? - What plugins are activated, if any, when the abort/exception occurs? - What significance is the Java version (as Geany doesn't use Java at all)?
You might try to run Geany from command prompt[0], passing the `-v` option, which will make it write out debug messages to the command prompt. I'm guessing the messages will contain a bunch of warnings/error messages leading up to the abort/unhandled exception. If so, and you could paste the output here, that would be really useful to troubleshoot the issue.
[0]: from Command Prompt window, type similar to:
``` C:> C:\Program Files (x86)\Geany\bin\Geany.exe -v ```
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-163652435
The same question as I asked on the ML, what were you doing when Geany crashed?
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-163764001
On Thu, Dec 10, 2015, at 16:06, Matthew Brush wrote:
A few questions:
- What version of Windows is that?
Windows 7 (64 Bit)
- Is it the Geany installer that comes with GTK+ bundled or did you
provide own GTK+ (if so, what version)?
I don't remember exactly, but I think it is the version with GTK+ included. In any case, I use the file geany-1.25_setup.exe with size 9770008.
- What plugins are activated, if any, when the abort/exception occurs?
According to the Plugin Manager, I have currently plugins activated which identify themselves with the following names:
"Fenster teilen" "Speicheraktionen" "Zeichen für HTML"
In at least one case, I had this type of crash also with the plugins deactivated.
- What significance is the Java version (as Geany doesn't use Java at
all)?
My mistake! I somehow thought to remember, that Geany was implemented in Java (for portability), but I see that this is not the case.
You might try to run Geany from command prompt[0], passing the `-v` option, which will make it write out debug messages to the command prompt. I'm guessing the messages will contain a bunch of warnings/error messages leading up to the abort/unhandled exception. If so, and you could paste the output here, that would be really useful to troubleshoot the issue.
Here it comes:
Geany: Geany 1.25, German_Germany.1252 Geany: GTK 2.24.10, GLib 2.28.8 Geany: System data dir: C:\p\geany\data Geany: User config dir: C:\Users\fisrona\AppData\Roaming\geany Geany: System plugin path: C:\p\geany\lib\geany Geany: Added filetype Clojure (61). Geany: Added filetype CUDA (62). Geany: Added filetype Cython (63). Geany: Added filetype Genie (64). Geany: Added filetype Graphviz (65). Geany: Added filetype JSON (66). Geany: Added filetype Scala (67). Geany: Loaded: C:\p\geany\lib\geany\splitwindow.dll (Fenster teilen) Geany: Loaded: C:\p\geany\lib\geany\saveactions.dll (Speicheraktionen) Geany: Loaded: C:\p\geany\lib\geany\htmlchars.dll (Zeichen für HTML) Geany: C:\cygwin64\home\fisrona\exp\jr6\rbsrc\vpruby.rb : Ruby (UTF-8)
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-163864423
On Thu, Dec 10, 2015, at 23:10, elextr wrote:
The same question as I asked on the ML, what were you doing when Geany crashed?
In one case, I was switching to a different tab. After the tab switch, the error popup occured.
In the other cases, I was typing some text, and out of a sudden it crashed.
I didn't see a pattern yet. Also, the crashes are not really frequent but since I don't use Geany everyday, it's difficult to estimate the frequency.
My impression was that the crashes came more often recently, but this could also be just by accident (I had two crashes within one day).
One odd observation (but I don't know whether this could be related): I have all my document encodings set to UTF-8, and one of the files I edited, has Japanese characters in it. While I can see them well with other text editors, Geany displays each of them as a tiny square with the digits 0 0 0 0 inscribed. I guess this is a replacement character, which Geany is using when it somehow thinks that it can't display them. However, the characters comprise just normal Japanese text. Could it be that an issue related to the encoding is causing the crash?
Ronald
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-163866333
Geany displays each of them as a tiny square with the digits 0 0 0 0 inscribed.
That is a character with the value 0, which should never occur in the buffer because it will cause problems. It may even cause your crashes. Geany should not load files with that character in it, but maybe its encoded on disk as an overlong UTF8 so it loads but then wreaks its havoc.
Were you entering characters using some entry method, Scintilla (the project that makes the editing widget Geany uses) has had a lot of changes in that area recently and maybe it leaks 0 characters into the buffer. Unfortunately none of the Geany devs does much with texts beyond Latin so we probably have never even tried Scintilla's entry methods in Geany.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-163876694
On Fri, Dec 11, 2015, at 09:46, elextr wrote:
Geany displays each of them as a tiny square with the digits 0 0 0 0 inscribed.
That is a character with the value 0, which should never occur in the buffer because it will cause problems. It may even cause your crashes.
If it just were so simple! I checked the source code with a hex editor, and there are no 0-bytes in the *file*. If Geany has Nullbytes in the *buffer*, it means that it does something weird when loading the file.
Two interesting observations here:
- It could well be that the crashes occured only when one of the tabs had a file with Japanese text and I had the files created with a different editor. - After editing the file with Geany (not touching the Japanese part) and saving, the saved file contains the correct text and the program produces the correct output.
Would it help if I send this file to the developers for inspection? I can send the file as it is - it doesn't contain any company secrets. Even if the crashes can't be reproduced, it should be easy to verify the strange display of the characters. Or, I could try to reproduce the problem with a smaller file, containing just a few characters (Western and Japanese mixed). For comparision, it should be sufficient if the developer opens this file with Geany, and with an editor which displays the text correctly (I'm using Code::Blocks as my other editor, but most UTF-8 enabled editors should work well).
BTW: The file has Unix-Lineendings (0A) ... just in case that this fact matters.
Ronald
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-163882173
Would it help if I send this file to the developers for inspection? I can send the file as it is - it doesn't contain any company secrets.
Can't hurt, can you put it in a Gist?
I might add that (on Linux) I have happily edited files containing mixed Latin and Japanese characters with no problems, but thats on a completely UTF-8 system.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-163883591
Also after you open the file, look in the menu->help->debug messages and see what encoding Geany thought the file was (it is parens at the end of the open message).
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-163885336
[...] I use the file geany-1.25_setup.exe [...]
I believe this is the one without a bundled GTK+ so you must have GTK+ 2.24 in your `%PATH%` from somewhere else. It should still work though, assuming no weird build options were enabled for that GTK+ build.
I'm kind of curious if the latest release (1.26) fixes the issue somehow. You might like to try it, and using the one with a GTK+ build bundled in it.
Here it comes: [...]
Is that the output when it crashes or just during a normal run? I was hoping before it terminated abnormally, it would print some assertion failures or such messages on its way down.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-163893048
[...] I use the file geany-1.25_setup.exe [...]
I believe this is the one without a bundled GTK+ […]
No, AFAIK the ones without GTK+ include "nogtk" in their name, see http://download.geany.org/
I'm kind of curious if the latest Geany release (1.26) fixes the issue somehow. You might like to try it, […]
Agreed. IIRC @zhekov fixed several problems related to incorrect struct size used for `stat()` buffer on 64 bits systems, so it might, perhaps, be related.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-163970262
On Fri, Dec 11, 2015, at 10:19, elextr wrote:
Would it help if I send this file to the developers for inspection? I can send the file as it is - it doesn't contain any company secrets.
Can't hurt, can you put it in a Gist?
Created here:
https://gist.github.com/rovf/d446dc8565ea82b4a710
I might add that (on Linux) I have happily edited files containing mixed Latin and Japanese characters with no problems, but thats on a completely UTF-8 system.
Maybe it's a good idea to double check *my* file with Linux too. Maybe it's something special at this particular file ...
Ronald
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164124146
On Fri, Dec 11, 2015, at 10:30, elextr wrote:
Also after you open the file, look in the menu->help->debug messages and see what encoding Geany thought the file was (it is parens at the end of the open message).
It says "UTF-8".
BTW, this can also be seen in the status line of Geany.
Ronald
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164124422
BTW, here is another Gist on the same issue, possibly easier to deal with:
https://gist.github.com/rovf/e6ec93c27b637bddc4d8
This one consist of only one single line, and it shows the same problem.
I've attached to this mail a screenshot. The number of "Null" displayed by Geany corresponds exactly to the number of Japanese characters in this text.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164124804
It says "UTF-8".
Ok.
BTW, this can also be seen in the status line of Geany.
Actually thats the encoding that the file will be saved in, it does not have to be the same as the encoding Geany used to read the file, though thats what it will default to for obvious reasons :)
Both your gists loaded fine and displayed the same characters as the gist did (slight differences due to fonts of course but pretty identical).
Unfortunately your screenshot didn't arrive, better to attach them in directly in github, it might not accept attachments from mail.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164125313
[...] I use the file geany-1.25_setup.exe [...]
I believe this is the one without a bundled GTK+ so you must have GTK+ 2.24 in your `%PATH%` from somewhere else. It should still work though, assuming no weird build options were enabled for that GTK+ build.
I found in my Geany directory DLLs with name 'gtk':
c:/p/geany/bin/libgtk-win32-2.0-0.dll c:/p/gedit/bin/libgtk-win32-2.0-0.dll c:/p/gedit/bin/libgtksourceview-2.0-0.dll
If *these* are the GTK+ libraries, I would conclude that it is version 2.0.
Aside from this, no dll files containing the string 'gtk' in any directory within my PATH. I do have other GTK dlls installed (for Cygwin), but these directories are not in the PATH when I start Geany.
I'm kind of curious if the latest release (1.26) fixes the issue somehow. You might like to try it, and using the one with a GTK+ build bundled in it.
I'l do this next week and report back my findings.
Is that the output when it crashes or just during a normal run? I was hoping before it terminated abnormally, it would print some assertion failures or such messages on its way down.
It was during the normal run. According to Murphys law, I didn't have any crashes *this* time.
Ronald
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164125402
If *these* are the GTK+ libraries, I would conclude that it is version
2.0.
The debug messages you pasted above say "Geany: GTK 2.24.10, GLib 2.28.8"
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164125519
I found in my Geany directory DLLs with name 'gtk':
As @b4n mentioned above, I was mistaken on the file naming. I believe you have the version that includes GTK+, so no worries there.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164160604
Here the attached screenshot (with Geany 1.25):
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164407642
I installed Geany 1.26. As for the unability to display the Japanese text, it is the same as in 1.25. I can't say much about the crashes. Will have to use it for some hours to see whether it happens. Unfortunatel I don't have much editing work to do this week.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164409109
I assume you can display the file in some other application that shows the Japanese characters correctly, but is it using the same font? To see the Geany font see Menu->View->Change Font and check its the same as that being used by the other application.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164423319
On Mon, Dec 14, 2015, at 13:19, elextr wrote:
I assume you can display the file in some other application that shows the Japanese characters correctly, but is it using the same font? To see the Geany font see Menu->View->Change Font and check its the same as that being used by the other application.
Ah! That's it!!!!!! It was the font.
This happens when you work on a Mac most of the time. When on Windows, it is easy to forget, that you need to select yourself a suitable font....
At least we now know, that the crashes and the font problem were not related. BTW, until now, I didn't experience any crash with Geany 1.26 yet. Hope this issue will be gone with the new version ...
Thanks for helping!
Ronald
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-164463901
Just for the record: After a few days without problems, I had two crashes within one hour. Again, I'm not aware that I was doing something peculiar. In one case, I had switched the tab and wanted to start editing, when Geany suddenly crashed. In the other case, I was scrolling the buffer.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-166344037
Hopefully one of the windows spurts can tell you how to run Geany under a debugger so you can get a traceback when it crashes.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-166399538
@elextr it's the same as everywhere else if using msys/cygwin. The problem will be if using Geany release, it most likely won't have debugging symbols compiled in.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-166411097
For what it's worth, I have been plagued by occasional crashes since I started using Geany version 1.26. I work with two different PCs. They both run Windows 7 64-bit; one is new and was set up in late November 2015 (it was in fact setting up this new PC that caused me to download the latest version of Geany). The other computer has been in use with the current configuration (except for the Geany update) for a couple of years at least.
Crashes occur irregularly, on average maybe 2 - 3 times a week, on both PCs. When I used Geany version 1.23 on the older PC (which I did since shortly after it was released, until early December 2015 - I never used any version between 1.23 and 1.26), I never had a single crash.
I use Geany as my primary editor and use it heavily, editing a considerable number of source code files, scripts etc. every day. I tend to have Geany running all the time, opening, modifying, saving and closing files in separate tabs within one instance of Geany.
I can't see a pattern where some specific action, file, or type of file trigs the crash. My files are C and C++ code, Python scripts and the like, with UTF-8 encoding.
I installed Geany with the bundled GTK+, and I don't use any plugins.
--- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/issues/795#issuecomment-178594595
I too experience geany crashing. I'm on Windows 10 using geany 1.33 built on or after Feb 25 2018. Geany crashes for editing markdown files too and can't pinpoint a way to reproduce it. During crashes even though I save my work I noticed when I startup geany my changes weren't saved at all after the crash. Maybe geany is crashing due to some issue with saving files.
@graph 1) run without NO plugins enabled, 2) if you saved (File->Save or ctrl+s) and it returned successfully then the save is done, that shouldn't be changed unless your file system is broken, is its on an external windows share?
This one I can't reproduce, it just happens randomly. Only on windows 10, randomly. I use geany on mac & linux and on those 2 geany works wonderfully.
To all windows users, AFAIK none of the regular Geany contributors use Geany on Windows regularly, so we don't see intermittent problems, we need you to provide information to debug those problems. But as @codebrainz noted, that may not provide sufficient information if there are no symbols. Somebody running their own build with debug symbols is the best way of providing the information needed for the Geany team to help you.
Heres stacktrace from the release version of geany 1.33 on windows 10. Probably not useful since it's release build.
``` libgdk-win32-2.0-0.dll!6c55e3a9() [Frames below may be incorrect and/or missing, no symbols loaded for libgdk-win32-2.0-0.dll] libgdk-win32-2.0-0.dll!6c55fae2() libgdk-win32-2.0-0.dll!6c545964() libgtk-win32-2.0-0.dll!61ad1735() libgtk-win32-2.0-0.dll!61a4e0aa() libgobject-2.0-0.dll!63c45f03() libgobject-2.0-0.dll!63c57d17() libgobject-2.0-0.dll!63c5fbdd() libgobject-2.0-0.dll!63c60937() libgtk-win32-2.0-0.dll!61b695d3() libgtk-win32-2.0-0.dll!61b4c500() libgobject-2.0-0.dll!63c45f03() libgobject-2.0-0.dll!63c582cd() libgobject-2.0-0.dll!63c600e7() libgobject-2.0-0.dll!63c60937() libgtk-win32-2.0-0.dll!61991104() libgobject-2.0-0.dll!63c45e2c() libgobject-2.0-0.dll!63c583b4() msvcrt.dll!73f4bd49() libglib-2.0-0.dll!687e2d90() libglib-2.0-0.dll!687d6281() libglib-2.0-0.dll!687d6281() libgobject-2.0-0.dll!63c45e2c() libgtk-win32-2.0-0.dll!61ab0120() libgobject-2.0-0.dll!63c45e2c() libgobject-2.0-0.dll!63c583b4() libgobject-2.0-0.dll!63c600e7() libgobject-2.0-0.dll!63c60937() libgtk-win32-2.0-0.dll!61b6d921() libgtk-win32-2.0-0.dll!61a78b16() libgobject-2.0-0.dll!63c45e2c() libgobject-2.0-0.dll!63c583b4() libgobject-2.0-0.dll!63c600e7() libgobject-2.0-0.dll!63c60937() libgtk-win32-2.0-0.dll!61b6d921() libgtk-win32-2.0-0.dll!6199b9e2() libgobject-2.0-0.dll!63c45e2c() libgobject-2.0-0.dll!63c583b4() libgobject-2.0-0.dll!63c4a6ee() libgobject-2.0-0.dll!63c6ce42() libgobject-2.0-0.dll!63c4473f() libgtk-win32-2.0-0.dll!61b76c14() libgobject-2.0-0.dll!63c45f03() libgobject-2.0-0.dll!63c6cb0f() libgobject-2.0-0.dll!63c45f03() libgobject-2.0-0.dll!63c57d17() libgobject-2.0-0.dll!63c600e7() ntdll.dll!7730e5fc() mswsock.dll!71faefaa() libgdk-win32-2.0-0.dll!6c5415dc() libglib-2.0-0.dll!687f3921() libglib-2.0-0.dll!687f3cd8() libglib-2.0-0.dll!687f4172() libgtk-win32-2.0-0.dll!61a4b5e9() libgeany-0.dll!6a182d9a() geany.exe!00401620() geany.exe!004013cb() kernel32.dll!74bc62c4() ntdll.dll!77300fa9() ntdll.dll!77300f74() ```
@graph thanks, but as noted above, without symbols its not much use. Also note there is no Geany code after we call gtk, which is probably the mainloop call but again without symbols we can't tell for sure.
I understand no worries. Do you know if the nightly builds have symbols? I assume they are stripped too, but it would make more sense for them to have since it's nightly. I haven't tried a nightly build yet.
Not sure if the windows nightly has symbols, nightlys were instituted before the advent of Travis CI to simply test that the repo compiled each day, not actually to produce runnable product, although thats an accidental side effect. So it may have symbols and it may not, maybe the nightly maintainer knows.
The builds from Travis CI are not available to download or so, they are just tests for successful compilation. The nightly builds on https://nightly.geany.org are currently not usable on Windows due to different C and GTK runtime environments used for compilation. So these builds are also only compilation tests.
I attached a hand made binary with full debug symbols and disabled compiler optimization. The binary was created by the same scripts and the same machine used for release binaries and is signed with the same key as the release builds.
Hopefully you get a better backtrace with this builds.
[geany_win32_debug.zip](https://github.com/geany/geany/files/1977464/geany_win32_debug.zip)
Thanks eht16!. I've been using this debug build for a while now and no crashing yet 😮
the debug build crashed
``` libgdk-win32-2.0-0.dll!00c4e3a9() [Frames below may be incorrect and/or missing, no symbols loaded for libgdk-win32-2.0-0.dll] libgdk-win32-2.0-0.dll!00c4fae2() libgdk-win32-2.0-0.dll!00c35964() libgtk-win32-2.0-0.dll!61ad196f() libgtk-win32-2.0-0.dll!61a4e0aa() libgobject-2.0-0.dll!63c45f03() libgobject-2.0-0.dll!63c57d17() libgobject-2.0-0.dll!63c5fbdd() libgobject-2.0-0.dll!63c60937() libgtk-win32-2.0-0.dll!61b695d3() libgdk-win32-2.0-0.dll!00c55bc9() libgdk-win32-2.0-0.dll!00c5222b() libgdk-win32-2.0-0.dll!00c52b17() libgtk-win32-2.0-0.dll!619d014e() libgdk-win32-2.0-0.dll!00c315dc() libglib-2.0-0.dll!687f3921() libglib-2.0-0.dll!687f3cd8() libglib-2.0-0.dll!687f4172() libgtk-win32-2.0-0.dll!61a4b5e9() libgeany-0.dll!6a182d9a() geany.exe!00401620() geany.exe!004013cb() kernel32.dll!76d462c4() ntdll.dll!77a00fa9() ntdll.dll!77a00f74() ```
I don't know why visual studio is not giving the symbols.
I don't know why visual studio is not giving the symbols.
As far as I know Visual Studio doesn't understand Dwarf debug info generated by non-MS compilers.
github-comments@lists.geany.org