I try to backup the linux kernel. But the .tar.gz file is so large which is about 2GB. Why the .tar.gz file I donwloaded before is only about 100MB?
Related
I am planning to run U-boot on an ARM based (Cortex-R5 BIG ENDIAN) embedded system.
U-Boot will run as a second level bootloader on top of a 1st level bootloader
developed using IAR Embedded Workbench for ARM (EWARM).
U-Boot is built on a Linux host computer.
The first level bootloader is developed on a Windows host computer.
I just managed to download the U-Boot ELF file using EWARM but there is no source debugging.
Have not copied the source files, but the ELF file of course contains linux style paths.
Is there any tool that converts an ELF file with linux style paths
to an ELF file with Windows style path?
"objcopy --only-keep-debug foo foo.dbg" takes a full ELF file
and copies just the debug stuff to foo.dbg
objcopy --strip-debug foo then strips the ELF file.
objcopy --add-gnu-debuglink=foo.dbg foo adds the connection to the debug info.
No objcopy option to convert paths as far as I can tell.
Is there any tool which copies a full directory tree, to NTFS handling symbolic links?
Since this is only for source level debug, making duplicates
instead of following the links would be OK.
My best idea is to copy the file system over the network using SAMBA.
Have not tried, but I suspect SAMBA would ignore the links and just make duplicates.
The Windows computer is connected to a company network and the VPN connection
disables all other networks, so I have to use a third computer for this...
Maybe I could SAMBA mount the linux machine on itself....
Ideas appreciated.
Is there any tool which copies a full directory tree, to NTFS handling symbolic links?
You have two options:
Copy the directory to a FAT partition. Symbolic links will be replaced by the actual files. Both Linux and Windows can read and write FAT.
If you are on the same computer: Mount the NTFS partition in Linux. If Windows uses bitlocker you will need the complete key (the PIN is not enough).
I have downloaded the latest kernel from kernel.org . What I want is to compile it to get the executable file. I need the file to run some commands on it (like strings) to figure what how to identify the linux kernel from its binary file.
All the tutorials on the internet are for upgrading/changing the kernel but I only want the binary file.
I think what you want are vmlinux and *.ko elf files, but you must compile your kernel first. The bzImage is compressed so you cannot run strings on it.
This might be a stupid question, but if boost is meant to be cross-platform, why does each platform need its own version?
If I want to package a (header-only) subset of boost as a dependency with my project, do I need to include both windows and linux versions?
Line endings.
Windows IDEs typically prefer to edit all files in CR+LF (windows) line-ends.
All other platforms use LF (Unix) style line ends.
From old documentation:
.zip file
The .zip format is widely supported by both free decoders and
commercial compress/archive utilities. If you don't already have a
.zip file decoder, download one from the Info-ZIP web site, which
supplies versions for many operating systems. Text file line endings
in the .zip file are as supplied by each library developer. This
works fine for Windows, but not for Unix/Linux. The .tar.gz and
.tar.bz2 files supply Unix/Linux friendly line endings.
.tar.gz and .tar.bz2 files
The .tar.gz format is widely supported on Unix/Linux platforms. Some
Windows compress/archive utilities can read the format as well.
Because the gzip format compresses the archive as a single file rather
than compressing each file individually, the .tar.gz file is smaller
that the .zip file.
The .tar.bz2 format is becoming widely available on Unix/Linux
platforms and is built into many tar utilities. This format differs
for the .tar.gz format in the compression used, which is considerably
better and therefore creates smaller files.
Text file line endings in the .tar.gz and .tar.bz2 files have been
converted to newlines for ease of use on Unix/Linux platforms.
I have pdftk installed on a shared file server, and it works fine from the server itself.
Is it possible to run pdftk from another computer that has access to the file server, but doesn't have pdftk installed itself?
Any help would be much appreciated.
pdftk doesn't need to be "installed", you just run the binary executable file. If I recall correctly, it does have a dependency on the libgcc library, so this would need to be present somewhere the system can automatically find it (like /windows/system32 on windows, or /lib | /lib64 on linux systems). Then you would just need to make sure the fileserver shares the folder that contains the pdftk binary executable and proper permissions are given.
I have a 120MB bundle that when is compressed (zip) is only 30MB.
I, however, so not want to distribute a zip file. I want to distribute a bundle in a dmg.
Are there any tools I can use that will enable me to compress my bundle, but also preserve the bundle file structure so that no 'unzip' tool is needed by end-users?
Just discovered "hdiutil -format UDBZ" which creates a bzip2 compressed image.