I am still learning how to use ANT well, and I wanted to understand if there is some reasonable way to do file tasks in it, similar to Rake and Make:
http://martinfowler.com/articles/rake.html#FileTasks
"With a file you are referring to actual files rather than task names. So 'build/dev/rake.html' and 'dev/rake.xml' are actual files. The html file is the output of this task and the xml file is the input. You can think of a file task as telling the build system how to make the output file - indeed this is exactly the notion in make - you list the output files you want and tell make how to make them.
An important part of the file task is that it's not run unless you need to run it. The build system looks at the files and only runs the task if the output file does not exist or it's modification date is earlier than the input file. File tasks therefore work extremely well when you're thinking of things at a file by file basis."
So in other words, let's say I want to run a custom binary and I only want that binary to run if any of the files have changed. This is related to this question, but I don't want to run the binary at all, not only pass a part of the fileset (i.e. there is only one in the fileset and I don't want the tool to run at all).
The ideal solution would also not be a loooong thing, but rather could be easily applied to any target -- perhaps using some ANT JavaScript or custom task?
Use ant-contrib outofdate task. It has exactly the properties you are asking for. Here is ant-contrib website.
Here is a template on how to integrate it into your build:
<taskdef
resource="net/sf/antcontrib/antlib.xml"
>
<classpath>
<pathelement location="${ant-contrib.jar}"/>
</taskdef>
<outofdate>
<sourcefiles path="dev/rake.xml"/>
<targetfiles path="build/dev/rake.html"/>
<sequential>
... do your work here ...
... will only run if rake.html is older than rake.xml ...
</sequential>
</outofdate>
Related
Using Shake, to create an mp3 (this is just a learning example), I use lame, and then id3v2 to tag it.
If the lame succeeds, but the id3v2 fails, then I'm left with the mp3 file in place; but of course it is "wrong". I was looking for an option to automatically delete target files if a producing command errors, but I can't find anything. I can do this manually by checking the exit code and using removeFiles, or by building in a temporary directory and moving as the last step; but this seems like a common-enough requirement (make does this by default), so I wonder if there's a function or simple technique that I'm just not seeing.
The reason Make does this by default is that if Make has a partial incomplete file on disk, it considers the task to have run successfully and be up to date, which breaks everything. In contrast, Shake records that a task ran successfully in a separate file (.shake.database), so it knows that your mp3 file isn't complete, and will rebuild it next time.
While Shake doesn't need you to delete the file, you might still want to so as to avoid confusing users. You can do that with actionOnException, something like:
let generateMp3 = do cmd "lame" ... ; cmd "id3v2" ...
let deleteMp3 = removeFile "foo.mp3"
actionOnException generateMp3 deleteMp3
I am currently facing an issue which I don't know how to fix. I got the following Julia code:
while true
print(watch_file("test"))
end
So this should get me all the file changes in the directory named "test". At least on windows.
Now thats all well and good, and it kinda works, at least for creating a file or moving a file to that directory. This is an example of what I get:
("New Textfile.txt",Base.FileEvent(true,false,false))
But when I delete or rename that file, I don't get the filename of the file deleted or renamed.
("",Base.FileEvent(true,false,false))
Is there a different method/function I can get the filename with, even when the file is deleted or renamed? Or even better, a way that archives this and is cross-platform-compatible? Any help appreciated.
EDIT: If you could give me an alternative that supports recursive monitoring, that would be even better.
In Linux, Julia 0.4.5 and 0.4.3 watch_file returns file name always. It is a very platform-dependent feature (like in Node.js https://nodejs.org/api/fs.html#fs_caveats) and only manual polling can be truly platform-independent solution.
I'm looking for something that's a bit robust in how it finds files in Emacs. I have a project made up a number of different files, and a lot of them. So, I think maybe Emacs would need to cache a lookup or something like that to make a quick find/open facility to work. It would need to also be configured per project to consider only some directories and exclude others inside of this project, since a number of files and directories are generated and hold a massive amount of text and sometimes a concatenated representation of the rest of the code.
Is there a quick file open/find like IntelliJ's find file, or Sublime's? Something with fuzzy search. But in Emacs? That could help with this problem?
Projectile can probably do what you're after. It describes itself as a "project interaction library" with facilities for finding project files quickly.
Try projectile: https://github.com/bbatsov/projectile (see its fancy UI, helm-projectile). You'll have the command projectile-find-file. It is based on projects (they are defined by a .git/.gh/… or a .projectile).
permanent caching ? Yes
filter out directories ? Yes (with a command or a config into the .projectile)
fuzzy search ? Yes, a few: emacs'default, ido, ido-fuzzy, grizzl or helm.
you install it simply with M-x package-install RET projectile RET.
See this EmacsWiki page, which is is a jumping-off place for multiple answers to your question.
Emacs has a built-in file-name cache -- see (emacs) File Name Cache and this page.
See also Emacs bookmarks, and in particular, Bookmark+. You can bookmark any file or set of files. You can bookmark a Dired buffer, including its omit set, markings, and included subdirs. You can bookmark a set of such Dired buffers. You can aggregate bookmarks and use them to perform actions that set up environments etc. They can be triggered in various ways. You can bookmark Emacs desktops. You can tag bookmarks and files & dirs with free-form tags, which lets you organize them flexibly into overlapping sets.
See also this page about project support with Icicles.
how to write an exec task in ant to replace two values in .m file.
Because I want to replace the version number in iphone app, including my build number which i build using hudson, i remotely access the macos and get the app to be copied on to my windows
system, so i generated the build using ant exec tasks, but now i want to replace my build number into the concern file before compiling,
so how can i use exec task to replace certain values in a file.
I know propertyfile task which replaces the value in a property file, but it doesn't work for
this.
If you have a template file, you can use the <copy> task with a nested <filterset> to replace certain values. If not, you can still use <copy>, this time with a nested filter chain containing a <replaceregex> token filter. If you want to edit in place, you can move the newly created file to the old location afterwards.
The use of <exec> is almost certainly a bad idea for your application, as it would make your build rely on external applications which usually are not available on all platforms, thus breaking portability of your build.
I am trying to use a MEL script to load ANIMATION.MA file that references CHARACTER_RIG.MA. The CHARACTER_RIG.MA and ANIMATION.MA files are produced by someone else and supplied to me. The ANIMATION.MA is looking for N:/Project/Maya//char/character/CHARACTER_RIG.MA
If I open ANIMATION.MA from Maya, or use the equivalent MEL command I always get prompted with:
"Reference File Not Found"
Reference File Not Found: N:/Project/Maya//char/character/CHARACTER_RIG.MA.
[Abort File Read] [Skip] [Browse...] [Retry]
If I tap browse, and select the CHARACTER_RIG.MA then it opens perfectly. I can see it created a reference in the Reference Editor that has the Unresolved Path (N:/...) , the Resolved Path (/my/path) and the namespace and the namespaceRN.
My question is, how do I do the equivalent of the "Browse..." from MEL? I tried pre-creating a reference, but it doesn't let me set the unresolved path, so when I load the ANIMATION.MA it keeps prompting in MAYA.
file -f -options "v=0" -typ "mayaAscii" -o "/Source/project/assets/anims/ANIMATION.MA"
If you know the directory where the file is, then you can use the dirmap command. The command dirmap allows you to remap directory structures if your disk configuration changes. So in this case it would look like:
dirmap -en true;
dirmap -m "N:/Project/Maya//char/character" "/my/path";
Possibly more manageable if you have lots of mappings to do especially when moving form a windows machine to a *nix one. However it is much more useful to define your project structure because then things just work when you move, tough this may not be the best of choice for shared assets.
I ended up finding several solutions:
Rename the RIG.MA file to match the filename in ANIM.MA (they were different in my case) and put it in one of the search or project folders that MAYA uses and it will automatically find it.
or
Programatically through code (or manually) edit the ANIM.MA file to remap the file/folder of the RIG.MA to where you want to load it from. Note: You
also need to remap any other files, such as textures. I did this with
perl -pi -e 's/\Qold-path\E/\Qnew-path\E/g' ANIMATION.MA
HTH someone else.
Quick and easy, File, Project, Set and select the folder where meshes or whatever it is.