PDFConverter Processing multiple postscript files fails around 1000 files

[ronlamb]

When trying to process convert hundreds of postscript files I get the following error sometime between 500 and 1000 files.

The code will start failing with the following error.

NFO : 27-Dec-2019 11:49:47.703: message: Temporary file ...AppData\Local\Temp\ghost4j\f8c87cf6-f505-403d-81d8-65635f12f3b0@13400 cannot be deleted
java.io.IOException: Temporary file ...AppData\Local\Temp\ghost4j\f8c87cf6-f505-403d-81d8-65635f12f3b0@13400 cannot be deleted
at org.ghost4j.util.DiskStore.removeFile(DiskStore.java:132)
at org.ghost4j.converter.PDFConverter.run(PDFConverter.java:265)
at org.ghost4j.converter.AbstractRemoteConverter.convert(AbstractRemoteConverter.java:85)

A retry will work. However after about 22K or so files all retries will fail.

I am using Java 1,8 on Windows using GhostScript 9.5 and Ghost4J 1.0.1

I tried running the same code on Linux, against both Ghostscript 9.25 and 9.5 and both won't even convert a single file.

INFO : 27-Dec-2019 16:54:00.809: Current allocation mode is local
ERROR: 27-Dec-2019 16:54:00.810: GPL Ghostscript 9.25: Unrecoverable error, exit code 1
INFO : 27-Dec-2019 16:54:00.810: message: org.ghost4j.GhostscriptException: Cannot initialize Ghostscript interpreter. Error code is -100
INFO : 27-Dec-2019 16:54:00.921: Error: /undefined in readonlyName
INFO : 27-Dec-2019 16:54:00.921: Operand stack:
INFO : 27-Dec-2019 16:54:00.921: --nostringval-- FontInfo --dict:2/9(L)-- FullName (Times New Roman)
INFO : 27-Dec-2019 16:54:00.921: Execution stack:
INFO : 27-Dec-2019 16:54:00.922: %interp_exit .runexec2 --nostringval-- --nostringval-- --nostringval-- 2 %stopped_push --nostringval
-- --nostringval-- --nostringval-- false 1 %stopped_push 2047 1 3 %oparray_pop 2046 1 3 %oparray_pop 2026 1 3 %oparr
ay_pop 1884 1 3 %oparray_pop --nostringval-- %errorexec_pop .runexec2 --nostringval-- --nostringval-- --nostringval-- 2 %stopp
ed_push --nostringval-- --nostringval--
INFO : 27-Dec-2019 16:54:00.922: Dictionary stack:
INFO : 27-Dec-2019 16:54:00.922: --dict:953/1684(ro)(G)-- --dict:0/20(G)-- --dict:82/200(L)-- --dict:31/60(L)-- --dict:5/20(L)-- --dict:2
/9(L)--
INFO : 27-Dec-2019 16:54:00.922: Current allocation mode is local
ERROR: 27-Dec-2019 16:54:00.922: GPL Ghostscript 9.25: Unrecoverable error, exit code 1

I tried an alternative method to process the file using, to see if I get better messages, but that worked even worse
public void write2() throws GhostscriptException, IOException {
ArrayList args = new ArrayList();
PaperSize paperSize = PaperSize.LETTER;

	args.add("-ps2pdf");
	args.add("-dNOPAUSE");
	args.add("-dBATCH");
	args.add("-dSAFER");
	args.add("-dProcessColorModel=/DeviceRGB");
	args.add("-dPDFSETTINGS=/prepress");
	args.add("-dCompatibilityLevel=1.4");
	args.add("-dPDFX=false");
	args.add("-dDEVICEWIDTHPOINTS=" + paperSize.getWidth());
	args.add("-dDEVICEHEIGHTPOINTS=" + paperSize.getHeight());
	args.add("-sDEVICE=pdfwrite");
	args.add("-dPDFWRDEBUG");
	args.add("-sOutputFile=" + outfile.getAbsolutePath());
	args.add("-q");
	args.add("-f");
	args.add("-");
	
	System.out.println("XXX  sOutputFile=" + outfile.getAbsolutePath());
	Ghostscript gs = Ghostscript.getInstance();
	
	synchronized (gs) {
		try (
			InputStream is = new ByteArrayInputStream(((ByteArrayOutputStream) os).toByteArray());
		) {
			gs.setStdIn(is);
    		String[] s = new String[args.size()];
    		s = args.toArray(s);
    		
    		int i=0;
    		for (String str : s) {
    			System.out.println("arg[" + i + "] = " + str);
    			i++;
    		}
    		
    			gs.initialize(args.toArray(new String[args.size()]));
    	}
	}
}

And I get a similar initialize error I get in Linux,
INFO : 27-Dec-2019 11:58:40.300: **** DSC comment: /Header
INFO : 27-Dec-2019 11:58:40.300: << /EPSF 0 /DSC_struct -dsc_data_struct- >>
INFO : 27-Dec-2019 11:58:40.301: **** DSC comment: /Creator
INFO : 27-Dec-2019 11:58:40.301: << /Creator (Designer Batch Version:15.0.23 Build Number:337) /DSC_struct -dsc_data_struct- >>
INFO : 27-Dec-2019 11:58:40.301: **** DSC comment: /CreationDate
INFO : 27-Dec-2019 11:58:40.301: << /CreationDate ((12/11/19) Time(16:46:39)) /DSC_struct -dsc_data_struct- >>
INFO : 27-Dec-2019 11:58:40.301: **** DSC comment: /Title
INFO : 27-Dec-2019 11:58:40.302: << /Title ((RunName:Report_BW StreamName:REPORT_PS)) /DSC_struct -dsc_data_struct- >>
INFO : 27-Dec-2019 11:58:40.411: **** DSC comment: /Page
INFO : 27-Dec-2019 11:58:40.412: << /PageNum 17 /DSC_struct -dsc_data_struct- >>
INFO : 27-Dec-2019 11:58:40.458: **** DSC comment: /Page
INFO : 27-Dec-2019 11:58:40.458: << /PageNum 18 /DSC_struct -dsc_data_struct- >>
INFO : 27-Dec-2019 11:58:40.486: **** DSC comment: /Page
INFO : 27-Dec-2019 11:58:40.486: << /PageNum 19 /DSC_struct -dsc_data_struct- >>
INFO : 27-Dec-2019 11:58:40.513: **** DSC comment: /Page
INFO : 27-Dec-2019 11:58:40.514: << /PageNum 20 /DSC_struct -dsc_data_struct- >>
ERROR: 27-Dec-2019 11:58:40.589: Extract failed: Cannot initialize Ghostscript interpreter. Error code is -100

I can manually run gswin64c, or gs in Linux on the command line all the files no problem, but would like to automate the conversion in both Windows and Linux.

[ronlamb]

The code to write the file using PDFConverter is:
public class PDFWriter extends FileWriter {

/**
 * PDF Converter is static - Repeatably creating and destroying the converter causes errors and
 * after about 26K files fails.
 */
private static final Logger logger = Logger.getLogger(FileWriter.class);
private static final PDFConverter converter;

static {
	logger.info("Initializing PDF Converter");
	converter = new PDFConverter();
	converter.setMaxProcessCount(0);
	converter.setPDFSettings(PDFConverter.OPTION_PDFSETTINGS_PREPRESS);
}

public PDFWriter(File outFile) throws IOException {
	super(outFile);
	
	os = new ByteArrayOutputStream();
}

public void write() throws IOException, ConverterException, DocumentException, GhostscriptException  {
	Document doc = new PSDocument();

	try (
		InputStream is = new ByteArrayInputStream(((ByteArrayOutputStream) os).toByteArray())
	) {
    	doc.load(is);
	}
	
	try (FileOutputStream fos = new FileOutputStream(outfile);)
	{
    	converter.convert(doc, fos);
    	/* remove the instance */
	} finally { 
		try {
			Ghostscript.deleteInstance();
		} catch (GhostscriptException e) {
			
		}
	}
}

}