/// <summary>
/// Runs PdfMaker.MakePdf() with the desired arguments. Note that the implementation (as of March 2015)
/// uses an external program to generate the PDF from the HTML file, so it doesn't need to be run on
/// a background thread. The process includes a (possibly overgenerous) timeout, so we don't try to
/// impose one here.
/// </summary>
/// <remarks>
/// Running this on a background thread would be okay, except that on Linux, the interaction between
/// Mono and NUnit and the Bloom method result in the BackgroundWorker.RunWorkerCompleted event
/// never being fired if tests other than those in this file are run along with these tests. This is
/// almost certainly an obscure bug in Mono. Running the method directly as we do here sidesteps that
/// problem. (See https://jira.sil.org/browse/BL-831.)
/// </remarks>
void RunMakePdf(PdfMaker maker, string input, string output, string paperSize, bool landscape, bool saveMemoryMode, bool rightToLeft,
PublishModel.BookletLayoutMethod layout, PublishModel.BookletPortions portion)
{
// Passing in a DoWorkEventArgs object prevents a possible exception being thrown. Which may not
// really matter much in the test situation since NUnit would catch the exception. But I'd rather
// have a nice test failure message than an unexpected exception caught message.
var eventArgs = new DoWorkEventArgs(null);
maker.MakePdf(input, output, paperSize, landscape, saveMemoryMode, rightToLeft, layout, portion, null, eventArgs, null);
}
/// <summary>
/// Runs PdfMaker.MakePdf() with the desired arguments. Note that the implementation (as of March 2015)
/// uses an external program to generate the PDF from the HTML file, so it doesn't need to be run on
/// a background thread. The process includes a (possibly overgenerous) timeout, so we don't try to
/// impose one here.
/// </summary>
/// <remarks>
/// Running this on a background thread would be okay, except that on Linux, the interaction between
/// Mono and NUnit and the Bloom method result in the BackgroundWorker.RunWorkerCompleted event
/// never being fired if tests other than those in this file are run along with these tests. This is
/// almost certainly an obscure bug in Mono. Running the method directly as we do here sidesteps that
/// problem. (See https://jira.sil.org/browse/BL-831.)
/// </remarks>
void RunMakePdf(PdfMaker maker, string input, string output, string paperSize, bool landscape, bool saveMemoryMode, bool rightToLeft,
PublishModel.BookletLayoutMethod layout, PublishModel.BookletPortions portion)
{
// Passing in a DoWorkEventArgs object prevents a possible exception being thrown. Which may not
// really matter much in the test situation since NUnit would catch the exception. But I'd rather
// have a nice test failure message than an unexpected exception caught message.
var eventArgs = new DoWorkEventArgs(null);
maker.MakePdf(new PdfMakingSpecs()
{
InputHtmlPath = input,
OutputPdfPath = output,
PaperSizeName = paperSize,
Landscape = landscape,
SaveMemoryMode = saveMemoryMode,
LayoutPagesForRightToLeft = rightToLeft,
BooketLayoutMethod = layout,
BookletPortion = portion
}, null, eventArgs, null);
}
/// <summary>
///
/// </summary>
/// <param name="inputHtmlPath"></param>
/// <param name="outputPdfPath"></param>
/// <param name="paperSizeName">A0,A1,A2,A3,A4,A5,A6,A7,A8,A9,B0,B1,B10,B2,B3,B4,B5,B6,B7,B8,B9,C5E,Comm10E,DLE,Executive,Folio,Ledger,Legal,Letter,Tabloid</param>
/// <param name="landscape"> </param>
/// <param name="booketLayoutMethod"> </param>
/// <param name="bookletPortion"></param>
/// <param name="worker">If not null, the Background worker which is running this task, and may be queried to determine whether a cancel is being attempted</param>
/// <param name="doWorkEventArgs">The event passed to the worker when it was started. If a cancel is successful, it's Cancel property should be set true.</param>
/// <param name="owner">A control which can be used to invoke parts of the work which must be done on the ui thread.</param>
public void MakePdf(string inputHtmlPath, string outputPdfPath, string paperSizeName, bool landscape, PublishModel.BookletLayoutMethod booketLayoutMethod, PublishModel.BookletPortions bookletPortion, BackgroundWorker worker, DoWorkEventArgs doWorkEventArgs, Control owner)
{
Guard.Against(Path.GetExtension(inputHtmlPath) != ".htm",
"wkhtmtopdf will croak if the input file doesn't have an htm extension.");
// Try up to 4 times. This is a last-resort attempt to handle BL-361.
// Most likely that was caused by a race condition in MakePdfUsingGeckofxHtmlToPdfComponent.MakePdf,
// but as it was an intermittent problem and we're not sure that was the cause, this might help.
for (int i = 0; i < 4; i++)
{
new MakePdfUsingGeckofxHtmlToPdfComponent().MakePdf(inputHtmlPath, outputPdfPath, paperSizeName, landscape,
owner, worker, doWorkEventArgs);
if (doWorkEventArgs.Cancel || (doWorkEventArgs.Result != null && doWorkEventArgs.Result is Exception))
{
return;
}
if (File.Exists(outputPdfPath))
{
break; // normally the first time
}
}
if (!File.Exists(outputPdfPath) && owner != null)
{
// Should never happen, but...
owner.Invoke((Action)(() =>
{
// Review: should we localize this? Hopefully the user never sees it...don't want to increase burden on localizers...
MessageBox.Show(
"Bloom unexpectedly failed to create the PDF. If this happens repeatedy please report it to the developers. Probably it will work if you just try again.",
"Pdf creation failed", MessageBoxButtons.OK);
}));
}
if (bookletPortion != PublishModel.BookletPortions.AllPagesNoBooklet)
{
//remake the pdf by reording the pages (and sometimes rotating, shrinking, etc)
MakeBooklet(outputPdfPath, paperSizeName, booketLayoutMethod);
}
}
/// <summary>
///
/// </summary>
/// <param name="pdfPath">this is the path where it already exists, and the path where we leave the transformed version</param>
/// <param name="incomingPaperSize"></param>
/// <param name="booketLayoutMethod"></param>
private void MakeBooklet(string pdfPath, string incomingPaperSize, PublishModel.BookletLayoutMethod booketLayoutMethod)
{
//TODO: we need to let the user chose the paper size, as they do in PdfDroplet.
//For now, just assume a size double the original
PageSize pageSize;
switch (incomingPaperSize)
{
case "A3":
pageSize = PageSize.A2;
break;
case "A4":
pageSize = PageSize.A3;
break;
case "A5":
pageSize = PageSize.A4;
break;
case "A6":
pageSize = PageSize.A5;
break;
case "B5":
pageSize = PageSize.B4;
break;
case "Letter":
pageSize = PageSize.Letter; //TODO... what's reasonable?
break;
case "HalfLetter":
pageSize = PageSize.Letter;
break;
case "Legal":
pageSize = PageSize.Legal; //TODO... what's reasonable?
break;
default:
throw new ApplicationException("PdfMaker.MakeBooklet() does not contain a map from " + incomingPaperSize + " to a PdfSharp paper size.");
}
using (var incoming = new TempFile())
{
File.Delete(incoming.Path);
File.Move(pdfPath, incoming.Path);
LayoutMethod method;
switch (booketLayoutMethod)
{
case PublishModel.BookletLayoutMethod.NoBooklet:
method = new NullLayoutMethod();
break;
case PublishModel.BookletLayoutMethod.SideFold:
method = new SideFoldBookletLayouter();
break;
case PublishModel.BookletLayoutMethod.CutAndStack:
method = new CutLandscapeLayout();
break;
case PublishModel.BookletLayoutMethod.Calendar:
method = new CalendarLayouter();
break;
default:
throw new ArgumentOutOfRangeException("booketLayoutMethod");
}
var paperTarget = new PaperTarget("ZZ" /*we're not displaying this anyhwere, so we don't need to know the name*/, pageSize);
var pdf = XPdfForm.FromFile(incoming.Path); //REVIEW: this whole giving them the pdf and the file too... I checked once and it wasn't wasting effort...the path was only used with a NullLayout option
method.Layout(pdf, incoming.Path, pdfPath, paperTarget, /*TODO: rightToLeft*/ false, ShowCropMarks);
}
}
/// <summary>
///
/// </summary>
/// <param name="inputHtmlPath"></param>
/// <param name="outputPdfPath"></param>
/// <param name="paperSizeName">A0,A1,A2,A3,A4,A5,A6,A7,A8,A9,B0,B1,B10,B2,B3,B4,B5,B6,B7,B8,B9,C5E,Comm10E,DLE,Executive,Folio,Ledger,Legal,Letter,Tabloid</param>
/// <param name="landscape"> </param>
/// <param name="booketLayoutMethod"> </param>
/// <param name="bookletPortion"></param>
/// <param name="doWorkEventArgs"> </param>
/// <param name="getIsLandscape"></param>
public void MakePdf(string inputHtmlPath, string outputPdfPath, string paperSizeName, bool landscape, PublishModel.BookletLayoutMethod booketLayoutMethod, PublishModel.BookletPortions bookletPortion, DoWorkEventArgs doWorkEventArgs)
{
Guard.Against(Path.GetExtension(inputHtmlPath) != ".htm",
"wkhtmtopdf will croak if the input file doesn't have an htm extension.");
MakeSimplePdf(inputHtmlPath, outputPdfPath, paperSizeName, landscape, doWorkEventArgs);
if (doWorkEventArgs.Cancel || (doWorkEventArgs.Result != null && doWorkEventArgs.Result is Exception))
{
return;
}
if (bookletPortion != PublishModel.BookletPortions.AllPagesNoBooklet)
{
//remake the pdf by reording the pages (and sometimes rotating, shrinking, etc)
MakeBooklet(outputPdfPath, paperSizeName, booketLayoutMethod);
}
}
/// <summary>
///
/// </summary>
/// <param name="inputHtmlPath"></param>
/// <param name="outputPdfPath"></param>
/// <param name="paperSizeName">A0,A1,A2,A3,A4,A5,A6,A7,A8,A9,B0,B1,B10,B2,B3,B4,B5,B6,B7,B8,B9,C5E,Comm10E,DLE,Executive,Folio,Ledger,Legal,Letter,Tabloid</param>
/// <param name="landscape">true if landscape orientation, false if portrait orientation</param>
/// <param name="saveMemoryMode">true if PDF file is to be produced using less memory (but more time)</param>
/// <param name="layoutPagesForRightToLeft">true if RTL, false if LTR layout</param>
/// <param name="booketLayoutMethod">NoBooklet,SideFold,CutAndStack,Calendar</param>
/// <param name="bookletPortion">None,AllPagesNoBooklet,BookletCover,BookletPages,InnerContent</param>
/// <param name="worker">If not null, the Background worker which is running this task, and may be queried to determine whether a cancel is being attempted</param>
/// <param name="doWorkEventArgs">The event passed to the worker when it was started. If a cancel is successful, it's Cancel property should be set true.</param>
/// <param name="owner">A control which can be used to invoke parts of the work which must be done on the ui thread.</param>
public void MakePdf(string inputHtmlPath, string outputPdfPath, string paperSizeName, bool landscape, bool saveMemoryMode, bool layoutPagesForRightToLeft,
PublishModel.BookletLayoutMethod booketLayoutMethod, PublishModel.BookletPortions bookletPortion, BackgroundWorker worker, DoWorkEventArgs doWorkEventArgs, Control owner)
{
// Try up to 4 times. This is a last-resort attempt to handle BL-361.
// Most likely that was caused by a race condition in MakePdfUsingGeckofxHtmlToPdfComponent.MakePdf,
// but as it was an intermittent problem and we're not sure that was the cause, this might help.
for (int i = 0; i < 4; i++)
{
new MakePdfUsingGeckofxHtmlToPdfProgram().MakePdf(inputHtmlPath, outputPdfPath, paperSizeName, landscape, saveMemoryMode,
owner, worker, doWorkEventArgs);
if (doWorkEventArgs.Cancel || (doWorkEventArgs.Result != null && doWorkEventArgs.Result is Exception))
{
return;
}
if (RobustFile.Exists(outputPdfPath))
{
break; // normally the first time
}
}
if (!RobustFile.Exists(outputPdfPath) && owner != null)
{
// Should never happen, but...
owner.Invoke((Action)(() =>
{
// Review: should we localize this? Hopefully the user never sees it...don't want to increase burden on localizers...
MessageBox.Show(
"Bloom unexpectedly failed to create the PDF. If this happens repeatedy please report it to the developers. Probably it will work if you just try again.",
"Pdf creation failed", MessageBoxButtons.OK);
}));
}
try
{
if (bookletPortion != PublishModel.BookletPortions.AllPagesNoBooklet)
{
//remake the pdf by reording the pages (and sometimes rotating, shrinking, etc)
MakeBooklet(outputPdfPath, paperSizeName, booketLayoutMethod, layoutPagesForRightToLeft);
}
else
{
// Just check that we got a valid, readable PDF. (MakeBooklet has to read the PDF itself,
// so we don't need to do this check if we're calling that.)
// If we get a reliable fix to BL-932 we can take this 'else' out altogether.
CheckPdf(outputPdfPath);
}
}
catch (KeyNotFoundException e)
{
// This is characteristic of BL-932, where Gecko29 fails to make a valid PDF, typically
// because the user has embedded a really huge image, something like 4000 pixels wide.
// We think it could also happen with a very long book or if the user is short of memory.
// The resulting corruption of the PDF file takes the form of a syntax error in an embedded
// object so that the parser finds an empty string where it expected a 'generationNumber'
// (currently line 106 of Parser.cs). This exception is swallowed but leads to an empty
// externalIDs dictionary in PdfImportedObjectTable, and eventually a new exception trying
// to look up an object ID at line 121 of that class. We catch that exception here and
// suggest possible actions the user can take until we find a better solution.
SIL.Reporting.ErrorReport.NotifyUserOfProblem(e,
LocalizationManager.GetString("PublishTab.PdfMaker.BadPdf", "Bloom had a problem making a PDF of this book. You may need technical help or to contact the developers. But here are some things you can try:")
+ Environment.NewLine + "- "
+ LocalizationManager.GetString("PublishTab.PdfMaker.TryRestart", "Restart your computer and try this again right away")
+ Environment.NewLine + "- "
+
LocalizationManager.GetString("PublishTab.PdfMaker.TrySmallerImages",
"Replace large, high-resolution images in your document with lower-resolution ones")
+ Environment.NewLine + "- "
+ LocalizationManager.GetString("PublishTab.PdfMaker.TryMoreMemory", "Try doing this on a computer with more memory"));
}
}
/// <summary>
///
/// </summary>
/// <param name="pdfPath">this is the path where it already exists, and the path where we leave the transformed version</param>
/// <param name="incomingPaperSize"></param>
/// <param name="booketLayoutMethod"></param>
/// <param name="layoutPagesForRightToLeft"></param>
private void MakeBooklet(string pdfPath, string incomingPaperSize, PublishModel.BookletLayoutMethod booketLayoutMethod, bool layoutPagesForRightToLeft)
{
//TODO: we need to let the user chose the paper size, as they do in PdfDroplet.
//For now, just assume a size double the original
PageSize pageSize;
switch (incomingPaperSize)
{
case "A3":
pageSize = PageSize.A2;
break;
case "A4":
pageSize = PageSize.A3;
break;
case "A5":
pageSize = PageSize.A4;
break;
case "A6":
pageSize = PageSize.A5;
break;
case "B5":
pageSize = PageSize.B4;
break;
case "Letter":
pageSize = PageSize.Ledger;
break;
case "HalfLetter":
pageSize = PageSize.Letter;
break;
case "QuarterLetter":
pageSize = PageSize.Statement; // ?? Wikipedia says HalfLetter is aka Statement
break;
case "Legal":
pageSize = PageSize.Legal; //TODO... what's reasonable?
break;
case "HalfLegal":
pageSize = PageSize.Legal;
break;
default:
throw new ApplicationException("PdfMaker.MakeBooklet() does not contain a map from " + incomingPaperSize + " to a PdfSharp paper size.");
}
using (var incoming = new TempFile())
{
RobustFile.Delete(incoming.Path);
RobustFile.Move(pdfPath, incoming.Path);
LayoutMethod method;
switch (booketLayoutMethod)
{
case PublishModel.BookletLayoutMethod.NoBooklet:
method = new NullLayoutMethod();
break;
case PublishModel.BookletLayoutMethod.SideFold:
// To keep the GUI simple, we assume that A6 page size for booklets
// implies 4up printing on A4 paper. This feature was requested by
// https://jira.sil.org/browse/BL-1059 "A6 booklets should print 4
// to an A4 sheet". The same is done for QuarterLetter booklets
// printing on Letter size sheets.
if (incomingPaperSize == "A6")
{
method = new SideFold4UpBookletLayouter();
pageSize = PageSize.A4;
}
else if (incomingPaperSize == "QuarterLetter")
{
method = new SideFold4UpBookletLayouter();
pageSize = PageSize.Letter;
}
else
{
method = new SideFoldBookletLayouter();
}
break;
case PublishModel.BookletLayoutMethod.CutAndStack:
method = new CutLandscapeLayout();
break;
case PublishModel.BookletLayoutMethod.Calendar:
method = new CalendarLayouter();
break;
default:
throw new ArgumentOutOfRangeException("booketLayoutMethod");
}
var paperTarget = new PaperTarget("ZZ" /*we're not displaying this anyhwere, so we don't need to know the name*/, pageSize);
var pdf = XPdfForm.FromFile(incoming.Path); //REVIEW: this whole giving them the pdf and the file too... I checked once and it wasn't wasting effort...the path was only used with a NullLayout option
method.Layout(pdf, incoming.Path, pdfPath, paperTarget, layoutPagesForRightToLeft, ShowCropMarks);
}
}
