I have some problems with an rpm spec file that is listing the same file multiple times. For this spec we do some normal compilation and then we have script that copies everything to the buildroot. Within this buildroot we have a lot of generic scripts that need to be installed on the final system, so we just list this directory.
However the problem is, that one of the scripts might be changed and configuration options might be changed within the script. So we list this script with different attributes as %config. However this means the script is defined multiple times with conflicting attributes, so rpmbuild complains and does not include the script at all in the installation package.
Is there a good way to handle this problem and to tell rpmbuild to only use the second definition, or do we have to seperate the script into two parts, one containing the configuration and one containing the actual logic?
Instead of specifying the directory, you can create a file list and then prune duplicate files from that.
So where you have something like
%files
%dir foo
%config foo/scriptname
You modify those parts to
find $RPM_BUILD_ROOT -type f | sed -e "s|^$RPM_BUILD_ROOT||" > filelist
sed -i "\|^foo/scriptname$|d" filelist
%files -f filelist
%config foo/scriptname
You can also use %{buildroot} in place of $RPM_BUILD_ROOT.
Related
I have a muttrc file which sources a secondary file mutt-secrets which resides in its same directory. But I have what appear to be two conflicting needs:
Be free to reference the muttrc file from any working directory
Be free to move it (and mutt-secrets) without having to edit muttrc to change the source path for mutt-secrets
At present, the first line of my muttrc says: source mutt-secrets. That works fine when I run mutt from within the directory where the two file reside, but if I run mutt from elsewhere and reference muttrc with a -F flag, then mutt can find muttrc, but muttrc can't find mutt-secrets.
How can I solve this?
Use absolute paths. For example:
source ~/.mutt/mutt-secrets
TL;DR one-line solution:
source `lsof -c mutt -Fn | grep '/muttrc$' | sed 's|^n||; s|/muttrc$||;'`/mutt-secrets
or, if you want to reuse the muttrc directory, you can save it to a custom variable:
set my_muttrc_dir = `lsof -c mutt -Fn | grep '/muttrc$' | sed 's|^n||; s|/muttrc$||'`
source $my_muttrc_dir/mutt-secrets
If you want to see the output of the command when you launch mutt, you can put this line in your muttrc:
echo `lsof -c mutt -Fn | grep '/muttrc$' | sed 's|^n||; s|/muttrc$||'`
Assumptions: the Mutt process is called mutt and the Mutt's initialization file is called muttrc. Furthermore, you could get in trouble if you have more than one Mutt instance running (for example if you launch in parallel two or more Mutt instances with different initialization files, because the command may select the wrong path).
Explanation
The idea is to look for muttrc full path in the list of open files by Mutt. We can get this list using lsof command (it has to be installed in your system), then extract the full path by parsing the lsof output with grep and sed commands.
This approach is viable because Mutt's initialization files support the use of external command's output with backticks (``). When Mutt encounter and execute our command enclosed in backticks (``), it is in the process of reading the muttrc file, so the muttrc file appears in the list of currently open files by Mutt. This enables us to use the lsof command.
lsof parameters
-c mutt: list open files of process named mutt;
-Fn: for each element, print only the name (it is the path in our case). Because of lsof output format, the path will be prefixed with the character n.
grep and sed
We use grep to select the line which contains muttrc file path, assuming the filename is exactly muttrc. Then we clean the lsof output with sed by both removing the n character at the beginning of the line and the /muttrc string from the end of the line. This way we get the path of the directory containing the muttrc file.
There is a cleaner solution?
Mutt expands relative paths inside initialization files from its current working directory, i.e. from the directory you launch Mutt. It supports a mechanism that allows path's expansion relatives to something different, but the "initialization file directory" or something similar are not available. See here.
I neither found a way to get the -F <path> option you pass to the mutt command inside the initialization file.
References
backticks in Mutt's initialization file;
current directory;
_mutt_buffer_expand_path, source code
source_rc, source code
source_rc call, source code
Tested with: Mutt 2.0.5, lsof 4.93.2, GNU grep 3.7, GNU sed 4.7.
I need to rsync files from a bucket to a local machine everyday, and the bucket contains 20k files. I need to download only the changed files that end with *some_naming_convention.csv .
What's the best way to do that? using a wildcard in the download source gave me an error.
I don't think you can do that with Rsynch. As Christopher told you, you can skip files by using the "-x" flag, but no just synch those [1]. I created a public Feature Request on your behalf [2] for you to follow updates there.
As I say in the FR, IMHO I consider this to not follow the purpose of rsynch, as it's to keep folders/buckets synchronise, and just synchronising some of them don't fall in that purpose.
There is a possible "workaround" by using gsutil cp to copy files and -n to skip the ones that already exist. The whole command for your case should be:
gsutil -m cp -n <bucket>/*some_naming_convention.csv <directory>
Other option, maybe a little bit more far-fetched is to copy/move those files to a folder and then use that folder to rsynch.
I hope this works for you ;)
Original Answer
From here, you can do something like gsutil rsync -r -x '^(?!.*\.json$).*' gs://mybucket mydir to rsync all json files. The key is the ?! prefix to the pattern you actually want.
Edit
The -x flag excludes a pattern. The pattern ^(?!.*\.json$).* uses negative look-ahead to specify patterns not ending in .json. It follows that the result of the gsutil rsync call will get all files which end in .json.
Rsync lets you include and exclude files matching patterns.
For each file rsync applies the first patch that matches, some if you want to sync only selected files then you need to include those, and then exclude everything else.
Add the following to your rsync options:
--include='*some_naming_convention.csv' --exclude='*'
That's enough if all your files are in one directory. If you also want to search sub folders then you need a little bit more:
--include='*/' --include='*some_naming_convention.csv' --exclude='*'
This will duplicate all the directory tree, but only copy the files you want. If that leaves empty directories you don't want then add --prune-empty-dirs.
what i want to achieve
I want to pack a .clang-format file with my software, which only gets used if the user doesn't provide their own.
My idea so far is to use clang-format -style=file which will find the users .clang-format if it is "located in one of the parent directories of the source file". (Which might be the case because it is in their home directory, or because they are using my package as a git submodule, or because they manually placed it after downloading my software).
If no .clang-format file is found, then the one shipped with my software should be used - e.g. by symlinking to mypackage/auxiliary_files/.clang-format in the top level. I want to do this somewhat automatized from cmake directly or by calling a shell script from cmake (or other).
workflow in short
clang-format finds .clang-format: do nothing (as shell script exit 0; as cmake script set(I-FOUND-CLANG-FORMAT 1) or something similar)
clang-format doesn't find .clang-format: call ln -s auxiliary_files/.clang-format
what I don't want to do
I don't want to write a loop which replicates the search behaviour of clang-format (going one directory up to look for .clang-format until either it's found or the top most directory is reached) to have built-in 100% compatibility.
options i've seen but didn't figure out how to use for my goal
clang-format does not appear to return a non-zero status code if no style file has been provided although style=file is set, even if -fallback-style=none is set. I also don't see in the printout if the fallback option or the regular option is used.
I can't rely on the umask since my machine does not use umask to set permissions. Is there a way to specify that all sub-directories (and their sub-directories etc) of some root directory all have a certain permission, and similarly, that all sub-files of the same root directory have another type of permission in the %files section of the spec file.
If not, I'll have to run some external bash scrip to get the spec file syntax for each individual file, and copy that output to the %files section of the spec file, which will be highly tedious.
If you look at the various references online, %defattr() takes a lesser-known fourth parameter for directories.
extracting ZIP files that have been created using:
PKSFX(R) Version 2.50 FAST! Self Extract Utility for Windows 95/NT 4-15-1998
Is there an option to specify which folder the files contained in the EXE should be extracted to?
By typing "/help" I was able to find some useful options (like Overwrite, Exclude, etc.), but I cannot see anything related to the Output Folder.
Any suggestion?
Thank you!
I found the following over at ousob.com. It appears as part of a larger article which seems to have the documentation for the whole PKSFX suite. It looks like you run the executable and then specify any flags followed by the path to extract to and, optionally, specific named files from within the archive to be extracted if you don't want them all.
PKSFX Command Summary
Command line syntax and options are the same as PKUNZIP. The options
listed below are offered by a self-extracting file.
Syntax: pksfx [option] [d:\path] [file]* [file]*
PKSFX indicates the name of the self-extracting file you are executing.
No options are necessary if you wish to simply extract all files to the
current directory. If you wish to extract the contents to another
directory, indicate the drive and/or path the files should extract to. To
extract only particular files, indicate the name(s) at the end of the
command line.
Options:
#listfile.lst Specify list of files for extraction*
-c[m] Extract to console*
-d Re-create directories stored in .ZIP file*
-l Display software license agreement*
-n Extract only newer files*
-o Overwrite existing files
-p[a|b][c][#] Extract to printer*
-s<password> Decrypt with password*
-t Test file integrity*
* Indicates options not available in the PKSFX Junior and PKUNZIP Junior
programs.
Options should be placed after the self-extracting files name. If
multiple options are used they must be separated by a space and each must
have its own option indicator (either - or /).
All options behave in the same manner as they do with PKUNZIP, with one
exception. The -s option cannot accept keyboard entry. A password
entered with a self-extracting file must be able to be entered from the
command line.