[英]Unable to Run application via Terminal but application working fine in XCode
I am working on lilgp which is ac language based tool for Genetic Programming. 我正在研究lilgp ,它是基于交流语言的遗传编程工具。 The problem that i am facing is that I am using XCode for the project and it is working fine and shows correct output via terminal.
我面临的问题是我正在使用XCode进行项目,它工作正常并通过终端显示正确的输出。 But when i try to run the same application in
DerivedData
of my project in XCode i get segmentation fault (11) 但是当我尝试在XCode中的项目的
DerivedData
中运行相同的应用程序时,我得到了分段错误(11)
Then i have checked the console in Utilities for errors which shows error like this 然后我检查了Utilities中的控制台是否有错误,显示错误
Process: Theisis [9325]
Path: /Users/USER/Library/Developer/Xcode/DerivedData/Theisis-gszeehddtmnlkqdbicpeffygvkcw/Build/Products/Release/Theisis
Identifier: Theisis
Version: 0
Code Type: X86-64 (Native)
Parent Process: bash [8987]
Responsible: Terminal [299]
User ID: 501
Date/Time: 2016-09-11 01:05:25.158 +0500
OS Version: Mac OS X 10.11.6 (15G31)
Report Version: 11
Anonymous UUID: 4063B9C3-F525-D9BD-EF5E-358810571673
Sleep/Wake UUID: CA5341A7-C252-4C76-B694-7F2DAE196F79
Time Awake Since Boot: 57000 seconds
Time Since Wake: 1600 seconds
System Integrity Protection: enabled
Crashed Thread: 0 Dispatch queue: com.apple.main-thread
Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x0000000000000068
VM Regions Near 0x68:
-->
__TEXT 0000000100c4e000-0000000100c6a000 [ 112K] r-x/rwx SM=COW /Users/USER/Library/Developer/Xcode/DerivedData/Theisis-gszeehddtmnlkqdbicpeffygvkcw/Build/Products/Release/Theisis
Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0 libsystem_c.dylib 0x00007fff93a8b09e flockfile + 4
1 libsystem_c.dylib 0x00007fff93a8d463 fscanf + 156
2 Theisis 0x0000000100c57853 app_initialize + 195 (app.m:614)
3 Theisis 0x0000000100c4f245 main + 453 (main.m:205)
4 libdyld.dylib 0x00007fff8e0575ad start + 1
Thread 0 crashed with X86 Thread State (64-bit):
rax: 0x00007fff5efb1970 rbx: 0x0000000000000000 rcx: 0x00000b0000000000 rdx: 0x0000000100c6aa8c
rdi: 0x0000000000000000 rsi: 0x0000000100c675d4 rbp: 0x00007fff5efb1860 rsp: 0x00007fff5efb1860
r8: 0x00000000fffffffc r9: 0x00007fff740b1c10 r10: 0x00007fff97709e01 r11: 0x00007fff93a8d3c7
r12: 0x450022575a4d98d4 r13: 0x0000000000000000 r14: 0x0000000100c675d4 r15: 0x0000000000000000
rip: 0x00007fff93a8b09e rfl: 0x0000000000010246 cr2: 0x0000000000000068
Logical CPU: 4
Error Code: 0x00000004
Trap Number: 14
VM Region Summary:
ReadOnly portion of Libraries: Total=176.1M resident=0K(0%) swapped_out_or_unallocated=176.1M(100%)
Writable regions: Total=19.6M written=0K(0%) resident=0K(0%) swapped_out=0K(0%) unallocated=19.6M(100%)
VIRTUAL REGION
REGION TYPE SIZE COUNT (non-coalesced)
=========== ======= =======
Activity Tracing 2048K 2
Kernel Alloc Once 4K 2
MALLOC 9604K 17
Stack 64.0M 3
VM_ALLOCATE 4K 2
__DATA 8300K 141
__LINKEDIT 91.4M 4
__TEXT 84.7M 146
__UNICODE 552K 2
shared memory 8K 3
=========== ======= =======
TOTAL 260.2M 312
However, when i checked app.m:614 it has if (strcmp(c, "regress_asim") != 0) {
where c is a chracter array and this array is working fine from last couple of months and in XCode too. 然而,当我检查app.m:614它有
if (strcmp(c, "regress_asim") != 0) {
其中c是一个chracter数组,这个数组在过去几个月和XCode中工作正常。 Can anybody tell me what am i missing? 谁能告诉我我错过了什么?
UPDATE: 更新:
1) The Resolved Project settings for both Debug and Release are same (except for the binary paths for sure) 1)Debug和Release的Resolved Project设置是相同的(确保二进制路径除外)
2) Code not only works in XCode but is also working in eclipse and creating a working binary. 2)代码不仅适用于XCode,而且还在eclipse中工作并创建工作二进制文件。
3) The part of the code that crashes in that the array c
is: 3)代码崩溃的部分,数组
c
是:
Declared as 宣称为
char c[100];
Initillized as 初始化为
strncpy(c, "Equation_Default_data.csv", sizeof(c));
The code block where it crashes 崩溃的代码块
if (!startfromcheckpoint) {
oprintf( OUT_PRG, 50, "not starting from checkpoint file.\n");
param = get_parameter("app.fitness_cases");
if (param == NULL)
fitness_cases = 200;
else {
fitness_cases = atoi(param);
if (fitness_cases < 0)
error( E_FATAL_ERROR,
"invalid value for \"app.fitness_cases\".");
}
FILE *in_file = fopen(c, "r");
fscanf(in_file, "%d", &fitness_cases);
if (strcmp(c, "regress_asim") != 0) { //Line 614
app_y_desired = (double *) MALLOC(fitness_cases * sizeof(double));
app_fitness_cases[0] = (double *) MALLOC(
fitness_cases * sizeof(double));
app_fitness_cases[1] = (double *) MALLOC(
fitness_cases * sizeof(double));
app_fitness_cases[2] = (double *) MALLOC(
fitness_cases * sizeof(double));
app_fitness_cases[3] = (double *) MALLOC(
fitness_cases * sizeof(double));
memset(app_fitness_cases[2], 0, fitness_cases * sizeof(double));
memset(app_fitness_cases[3], 0, fitness_cases * sizeof(double));
memset(app_y_desired, 0, fitness_cases * sizeof(double));
}
app_fitness_importance = (int *) MALLOC(fitness_cases * sizeof(int));
//Asim Code
double x, y;
for (i = 0; i < fitness_cases; ++i) {
fscanf(in_file, "%lf", &x);
fscanf(in_file, "%lf", &y);
app_fitness_cases[0][i] = x;
app_fitness_cases[1][i] = y;
if (strcmp(c, "regress_asim") != 0) {
app_y_desired[i] = y;
}
app_fitness_importance[i] = checkImportance(x);
}
fclose(in_file);
datapointsPerImportance = (int*) MALLOC((max_datapoint_importance+1)*sizeof(int));
memset(datapointsPerImportance, 0, (max_datapoint_importance+1)*sizeof(int));
for (i = 0; i < fitness_cases; ++i) {
// printf("%d : %d\n",i,checkImportance(app_fitness_cases[0][i]));
datapointsPerImportance[checkImportance(app_fitness_cases[0][i])]=datapointsPerImportance[checkImportance(app_fitness_cases[0][i])]+1;
}
for(int i=0;i<=max_datapoint_importance;i++)
{
printf("Importance %d =%d\n",i,datapointsPerImportance[i]);
}
/*oprintf( OUT_PRG, 50, "%d fitness cases:\n", fitness_cases);
for (i = 0; i < fitness_cases; ++i) {
x = (random_double() * 2.0) - 1.0;
// change this line to modify the goal function.
y = x * x * x * x + x * x * x + x * x + x;
app_fitness_cases[0][i] = x;
app_fitness_cases[1][i] = y;
// oprintf( OUT_PRG, 50, " x = %12.5lf, y = %12.5lf\n", x, y);
}*/
} else {
oprintf( OUT_PRG, 50, "started from checkpoint file.\n");
}
Without seeing the code, this is like repairing a car in absolute darkness. 没有看到代码,这就像在绝对黑暗中修理汽车。 Try it.
试试吧。
However, I'll give it a try. 但是,我会试一试。
In the stack dump, it shows that you are using fscanf
to read data from a file. 在堆栈转储中,它显示您正在使用
fscanf
从文件中读取数据。 Later, strcmp
crashes. 后来,
strcmp
崩溃了。
The obvious idea is that fscanf
reads more characters than can fit in the char array, which leaves no space for the closing \\0
. 显而易见的想法是,
fscanf
读取的字符数多于char数组中可容纳的字符数,因此不会为关闭\\0
留下空间。 strcmp
(or any other code, doesn't really matter) the runs past the end of the string, until it hits something sensible, and then it crashes. strcmp
(或任何其他代码,并不重要)运行超过字符串的结尾,直到它击中一些合理的东西,然后它崩溃。
If this is the reason, making the buffer longer would fix it (temporarily). 如果这是原因,使缓冲区更长时间(暂时)将修复它。 Set it something huge, like 4096, to verify the approach, and then find the real fix.
设置一些巨大的东西,比如4096,来验证方法,然后找到真正的修复。
If you ask why the same program would work in one place, but not in the other - you are probably reading a different file, so the content of the file has different lengths; 如果你问为什么同一个程序在一个地方工作,而不是在另一个地方 - 你可能正在读一个不同的文件,所以文件的内容有不同的长度; or you simple reading the file in a different directory, which has a longer name;
或者您只是在不同的目录中读取文件,该目录的名称较长; any of those can be the reason for the char array overflow.
任何这些都可能是char数组溢出的原因。
Second idea: The dump line VM Regions Near 0x68
shows the memory area where it dumped, and its content looks like a part of a ls result or a path/filename ( /Users/USER/Library/Developer/Xcode/DerivedData/Theisis-gsze....
), so potentially your variable for that path is too short. 第二个想法:转储行
VM Regions Near 0x68
显示它转储的内存区域,其内容看起来像ls结果或路径/文件名的一部分( /Users/USER/Library/Developer/Xcode/DerivedData/Theisis-gsze....
),因此该路径的变量可能太短。 Make that one longer too and try. 让那个更长,然后尝试。
Edit: c[100]
is too small, the path shown there is 115 chars . 编辑:
c[100]
太小,显示的路径有115个字符 。 Note that .\\Thesis
will be replaced by \\Users\\...
by the OS, which then is longer than 100 chars. 请注意
.\\Thesis
OS将替换为\\Users\\...
,然后超过100个字符。
Declare c[260]
(or even larger) and you will be good. 声明
c[260]
(甚至更大),你会很好。
According to these lines in the debug output: 根据调试输出中的这些行:
Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0 libsystem_c.dylib 0x00007fff93a8b09e flockfile + 4
1 libsystem_c.dylib 0x00007fff93a8d463 fscanf + 156
What is causing the crash is the fscanf, not the strcmp. 导致崩溃的原因是fscanf,而不是strcmp。 Check the return values from fopen, it must be NULL in your test case (most probably the file that is being read hasnt been found; when running the program from XCode, the working directory may not be the same as you were expecting).
检查来自fopen的返回值,在测试用例中它必须为NULL(很可能是找不到正在读取的文件;从XCode运行程序时,工作目录可能与您预期的不同)。
In your case, this should suffice: 在你的情况下,这应该足够:
FILE *in_file = fopen(c, "r");
if (in_file == NULL) {
perror( "The following IO error occurred" );
error( E_FATAL_ERROR, "IO error" );
}
int r = fscanf(in_file, "%d", &fitness_cases);
if (r != 1) {
perror( "The following IO error occurred" );
fclose( in_file );
error( E_FATAL_ERROR, "Read Error" );
}
Assuming that your error function jumps out of the function. 假设您的错误函数跳出函数。 If not, use some
return
s; 如果没有,请使用一些
return
;
You need to check return values in C 您需要在C中检查返回值
FILE *in_file = fopen(c, "r");
/* insert error checking code here */
fscanf(in_file, "%d", &fitness_cases);
/* otherwise it fails here when fscanf tries to read from NULL */
if (strcmp(c, "regress_asim") != 0) { //Line 614
From the looks at the rest of the posted code, it looks like error checking is an issue (MALLOC?) so the copying of the file path (not posted?) probably has the same issues and overflows the "c" buffer, causing fopen to return NULL and causing fscanf to fail. 从发布代码的其余部分看,看起来错误检查是一个问题(MALLOC?)所以复制文件路径(未发布?)可能有相同的问题并溢出“c”缓冲区,导致fopen返回NULL并导致fscanf失败。 CHECK RETURN VALUES... if strncpy returns a string that doesn't have '\\0' in c[sizeof(c)-1] - you cannot use it to open a file.
检查返回值...如果strncpy返回c [sizeof(c)-1]中没有'\\ 0'的字符串 - 则不能使用它来打开文件。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.