I have a DCL script on VMS which calls a perl script. Is there a VMS/DCL command I can use that will tell me every file handle opened by the perl script?
Set default to the disk the app runs from (or you might have to try each disk in succession if it's a really large or distributed app). Then the command is
show device/files/nosystem
If you're on a more recent version of VMS and the lists are too long, you can pipe it with a search by doing this:
pipe show device/files/nosystem | search sys$input (name of perl script)
You need to find the documentation for undocumented VMS features :-)
Seriously I think that set watch might do what you want. If you issue
$ set watch file/class=(all,nodump)
$ perl yourperlscript.pl
You will get loads of output that will hopefully include what you want. I havent done it for years, you probably tune the options to fine tune it. See
http://www.parsec.com/openvms/undocumented.php?page=13
Jason, I need more clarification for a). Are you saying that you want to run your perl script in a batch file and have the batch file monitor the files being accessed by the perl script? Or something else?
Hmm, not sure about that. Maybe add a linux tag to your post so that some linux people can see this and chime in. I'm not sure why your perl program wouldn't know what files it opened. It's your program, wouldn't it access the files you told it to access? Or if you're computing the filenames somehow (which I've done in cobol, but still know at least which directory to find them in, and what naming scheme they use), you'd still have clues like what I mention. Also, since it's your program, and if you're computing the filenames, coudn't you also make your Perl program output it's own little report of what the files were? Like, just after it computes the filename, have it copy the name string to a separate report file.
Related
Is there a way to do so ? I searched Google but couldn't find any answer, so I guess the answer would be no. Is there anything close ? If not, would it be easy to extend red-lang to do so ?
From http://www.rebol.com/docs/setup.html
Startup Scripts
When REBOL starts it will automatically run the rebol.r and user.r files, if they exist.
The system looks for these files first in the current directory (or the directory of the script being run), then in the directory that contains the REBOL executable program.
Note that REBOL/Core runs fine without the rebol.r and user.r files. They simply provide an easy way to include additional code and data on startup, such as your network preferences.
If you compile your own Red interpreter you can add an autoloading file, maybe in console.red after system/console/init "Red Console" and before system/console/launch Best advice is to ask on the https://gitter.im/red/help site to ask for help. I guess this was already discussed.
When am trying to open any .EXE file am getting information in encoded form. Any idea how to see the content of an .EXE file ????
I need to know what Database tables are used in the particular .EXE.
Ah, now we are getting closer to the real question.
It is probably much more productive to ask the targeted databases about the SQL queries being execute during the run, or a top-ten shortly afterwards.
The table-names might not be hard-coded recognizably as such in the executable.
They might be obtained by a lookup, and some fun pre-fixing or other transformation might be in place.
Admittedly they like are clear text.
Easiest is probably to just transfer to a Unix server and use STRINGS on the image.
I want to include the source here with but that failed, and I cannot find how to attach a file. Below you'll find a link OpenVMS macro program source for a STRINGS like tool. Not sure how long the link will survive.
Just read for instructions, save (strings.mar), compile ($ MACRO strings), link ($link strings), and activate ($ mcr sys$login:strings image_to_test.exe)
OpenVMS Macro String program text
Good luck!
Hein
Use analyze/image to view the contents of an executable image file.
I'm guessing you are trying to look in the EXE because you do not have access to the source. I do something like this:
$ dump/record/byte/hex/out=a.a myexe.exe
Then look at a.a with any text editor (132 columns). The linker groups string literals together, and they are mostly near the beginning of the EXE, so you don't have to look to far into the file. Of course this only helps if the database references are string literals.
The string literal might be broken across a block (512 byte) boundary, so if you use search in your editor, try looking for substrings.
Aksh - you are chasing your tail on this one. Its a false dawn. Even if you could (and you can't) find the database tables, you will need the source of the .exe to do anything sensible with it, or the problem you are trying to solve. Its possible to write a program which just lists all the tables in a database without reading any of 'em. So you could spend and awful lot of effort and get nowhere. Hope this helps
I'm trying to convince fellow developers to not include $ in file names but couldn't come up any argument other than some Perl scripts may need to be updated. Any suggestions?
A quick glance at Wikipedia's Filename page reveals that only zOS and Commodore DOS disallow the user of $ appearing in certain file names. I can see why you would have trouble coming up with a reason not to use them.
I currently have a directory with 98,000 individual archive transaction files. I need to search those files for user input strings and have the option to open the files as it finds them or at the end of the search. I'm using Notepad++ currently and, while functional, it's quite slow. I thought about writing my own, but I am only familiar with .NET and I'm a beginner. Also, I'm not sure how efficient that would be compared to NP++.
This tool would be used again and again so the dev time would definitely be worth it if it came to that. Is there some other tool out there that's already developed that would accomplish this?
Agent Ransack
I've been using it for years.
I recommend you using Astrogrep, a grep utility for Windows. You can open files as it finds them, and it shows you the line where the match was found, without having to open the file.
Assuming the archive transaction files are plain text, you can download CYGWIN which is an environment providing UNIX tools for Windows.
Once that's done, you can open a new Cygwin Bash Shell, then do cd 'c:\\foo' to get into the directory with your files, then do grep -F -r "my string" * to find your text. (The -F means it searches for that literal string as opposed to a regular expression and -r means recursive.)
Possibly overkill, but you could index the folder using Lucene, keep the index uptodate (as transaction files are added) and then searches will take trivial amounts of time, you can target the file, line and word number of each match for a given search string
There are many programs out there that will allow you to pack a few files together and generate an executable that has the necessary code to extract them. Somehow, those files are residing inside the executable. I am interested in doing the same thing; how is this done?
FYI, I'm interested primarily in Windows .exe files, if it makes a difference.
Look at this: Article
You could probably save a file/files in a resource, compile it into the exe then use some code in the exe to extract it out to a file.
ie:
http://www.codeproject.com/KB/winsdk/binaryresources.aspx
Self extracting .exe files are usually archive files (zip, rar, tar, etc) concatenated together with a small program that does the extraction then executes the program that was extracted from the archive.
A really sophisticated one could extract the archive into memory and then jump to the extracted code and run it, but back in the old days, that sort of thing was easier to do.
If you wanted to write your own in Windows, you would create a small console application that did the extraction, and you would include the 'real' program in the console programs' resources.
There are also products like pkzip and winzip which do it for you. If your time is worth anything, those would be more efficient.
UPX is a well known packer for Windows .EXE which can be found here on WikiPedia. And here is the main site on sourceforge for UPX.
Hope this helps,
Best regards,
Tom.