I'm having a slight annoyance with the javascript_I18n plugin, which generates js-friendly versions of the I18n translation tables, so that you can localise your javascript. It all works fine, but it works by calling to_json on each locale's translations hash and outputting the results into a file. When you call to_json on a hash the resulting string is all a single lne, which means you end up with files with an enormous line at the end.
This in turn is stopping git from being able to merge any changes, because the merge works on a line by line basis and can't deal with a single massive line with changes in the middle somewhere. It's not a massive problem because i can always just re-generate the js-friendly translation file with a rake task that comes with the plugin (which replaces the mid-merge file with a brand new one which i can just commit in), but it's just a bit annoying. It occurred to me that if the json was output on different lines, instead of all the same line, then it wouldn't be a problem, and it wouldn't even make the file that much bigger, just inserting two characters (\n) per line.
Before i try and hack the resulting string with a gsub to split it onto seperate lines, is there a nicer way to call to_json on a hash and output the results onto seperate lines? Or a nicer way to solve this problem generally? (i can't find much useful in the documentation for javascript_I18n).
Grateful for any advice - max
Not to answer you question, but to give a suggestion:
It would probably be easier to just ignore all the generated js translation files.
This is because even if you separate the translation into multiple lines, there are still chances of merge conflicts, which you probably already resolved once in your yaml file.
I would set it up this way:
1). gitignore all the js translation files.
2). In ActionController, add a before filter to generate the js files automatically on each load in development mode only.
3). Tweak your deploy.rb file to generate the js files after code update.
No more merge conflicts! (at least for js translation files) :D
Aaron Qian
Related
I need to process a (GCS) bucket of files, where each file is compressed and contains a single multi-line JSON record. Also, the name of the file being processed is significant and I need to know it within my transform.
Starting with examples in the docs, TextIO looks pretty close, but it looks like its designed to process each file line-by-line and does not allow me to read the entire file at once. Also, I don't see any way to get the filename that's being processed?
PCollectionTuple results = p.apply(TextIO.Read
.from("gs://bucket/a/*.gz")
.withCompressionType(TextIO.CompressionType.GZIP)
.withCoder(MyJsonCoder.of()))
Looks like I need to write a custom IO reader, or some such? Any tips for best place to start?
You are correct that right now none of the existing classes do exactly what you want. There are 2 reasonable approaches:
Match the filepattern yourself (using IOChannelUtils and IOChannelFactory) and wrap the resulting files into a PCollection<String> where the String will be a filename, using Create.of(filenames). Then apply a ParDo with a function which reads the given filename.
Write your own subclass of Source (there's also FileBasedSource, but it's not quite right for your use case). It would be configured by the filepattern, and splitIntoBundles would match the filepattern and expand into individual sources each corresponding to one file.
I would recommend the first approach because it seems like less code and your use case does not require the full power of Source.
I have an XML file in my app resources folder. I am trying to update that file with new dictionaries dynamically. In other words I am trying to edit an existing XML file to add new keys and values to it.
First of all can we edit a static XML file and add new dictionary with keys and values to it. What is the best way to do this.
In general, you can read an XML file into a document object (choose your language), use methods to modify it (add your new dictionary), and (re-)write it back out to either the original XML file, or a new one.
That's straightforward ... just roll up the ol' sleeves and code it up.
The real problem comes in with formatting in the XML file before and after said additions.
If you are going to 'unix diff' the XML file before and after, then order is important. Some standard XML processors do better with order than others.
If the order changes behind the scenes, and is gratuitously propagated into your output file, you lose standard diffing advantages, such as some gui differs, and some scm diffs (svn, cvs, etc.).
For example, browse to:
Order of XML attributes after DOM processing
They discuss that DOM loses order where SAX does not.
You can also write a custom XML 'diff'er (there may be such off-the-shelf ... for example check out 'http://diffxml.sourceforge.net/') that compares 2 XML documents tag-by-tag, attribute-by-attribute, etc.
Perhaps some standard XML-related tool such as XSLT will allow you to keep the formatting constant without changing tag or attribute order. You'd have to research that.
BTW, a related problem is the config (.ini) file problem ... many common processors flippantly announce that the write-order may not agree with the read-order.
Unfortunately this question may be a bit vague, I have a problem that I am finding difficult to describe, it is intermittent and I cannot reproduce it myself, I am just hoping that someone else has seen something like it before.
My application has quite a lot of text and ini files that get written when it closes down. Typically this would be in response to a Close event, but may also be triggered by a WM_ENDSESSION. Unfortunately at the moment I am not sure if both or only one of these events can result in the problem I am about to describe, because I have been unable to reproduce this problem myself.
The issue I have is that for some users some of the text and ini files end up being written as NULLs. The file sizes end up looking about right, but instead of text, every character is written as a x00. So instead of 500 bytes of regular ASCII text I end up with 500 x00's. I also have an application log file that can sometimes end up with nulls written to it also. However the logging of x00's to the log file does not necessarily correspond to the exact same time as x00's were written to the config files.
For my files I am using TmemIniFile or TstringList which means that ultimately a Tstrings.SaveToFile is being called for all of my config files.
sl:=TstringList.Create;
try
SourceList.GetSpecificSubset(sl);
AppLogLogLine('Commands: Saving Always Available list. List has '+inttostr(sl.Count)+' commands.');
sl.SaveToFile(fn);
finally
sl.Free;
end;
But then I also have instance where I would already have a TstringList in memory and I just call SaveToFile on it. For TmemIniFile the structure would look similar to above. In some instances I may have an outer loop to write multiple lists. Some of those will result in files being written correctly, some will be full of 00's.
EDIT: GetSpecificSubset is simply a function that will populate "sl" with a list of command names. I have "GetAllUsersCommands", "GetHiddenCommands", "GetAlwaysVisibleCommands" etc. Note that my log file also writes this kind of thing, as a check for how big those lists are:
16/10/2013 11:17:49 AM: Commands: Saving Any User list. List has 8 commands.
16/10/2013 11:17:49 AM: Commands: Saving Always Visible list. List has 17 commands.
16/10/2013 11:17:49 AM: Commands: Saving Always Hidden list. List has 2 commands.
I accidentally left the logging line out of the code above. So this log line is the last thing written before calling Tstrings.SaveToFile, and at that point it thinks it has data. Even if somehow each line of text were NULLs, I would still expect to see x13x10 in the files, but that is not happening.
Here's a screen cap from a HEX editor:
EDIT 2: I just realised I left off a very important piece of information. This is only intermittent. It works 99% of the time. When saving files at shutdown it might not even be all files. Even if I have a loop saving multiple similar files, some may work fine and others may fail.
I am wondering if there is an easy way of keeping different i18n files in sync, so when adding a key/value in messsage_aa.properties would result in the same line in a message_bb.properties file? But this offcourse using a different value..
I would really like to have the same key on the same line number in each of my message_xx.properties file..
Any suggestion?
This feature depends on your IDE.
As an example, I use ResourceBundle plugin in Eclipse that partially do what you are asking.
"Partially", because it keeps the same order but not necessary the same line (i.e. it doesn't skip line to match one file with another).
I'm using odt file as some kind of template and Libre Office as tool to create this template. It usually works fine except one thing.
Let assume our odt file has a paragraph of text.
There is my text.
XML file may or may not look (seems random) like this (messy, not very good thing for for parsing or as a template):
<text:p text:style-name="P7">There is</text:p><text:p text:style-name="P7"> my text<text:p text:style-name="P7">.</text:p></text:p>
Sometimes it's (again seems random) like this (expected result, makes sense after all):
<text:p text:style-name="P7">There is my text.</text:p>
Is there any way to get rid superfluous xml tags? Or at least can user see a raw document in LibreOffice/OpenOffice to manually remove redundancy?
The key is to provide easy tool for a user, to detect and fix artefacts like this.
Have you tried Ctrl-M? If all formatting is defined in styles and style formatting is not manually overridden, it should not disturb the formatting but should remove redundant tags.
A tedious user process would be to cut and paste-special as text and apply style again.
Finally, a macro would definitely do the trick.