[ https://issues.apache.org/jira/browse/PDFBOX-5022?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17236485#comment-17236485 ]
Tres Finocchiaro edited comment on PDFBOX-5022 at 11/20/20, 10:34 PM: ---------------------------------------------------------------------- [~tilman] I had just assumed it was some type of circular reference, but I don't know the PDF specification well enough to understand. Slightly off-topic, but in regards to the out-of-memory stuff, is there a way (perhaps through an enhancement) to ignore the transparency-forces-rasterization behavior? As the printers get higher and higher quality, this can really be a detriment. Assuming this request is preposterous, could this be (easily) done programmatically using the API, so that downstream care can be taken to avoid transparency-forced rasterization at all costs? This technique may be common for some (e.g. manuals , brochures), but it's very rare for others (e.g. reports, labels, folios, invoices) and having control over this would help guard the memory issues, albeit downstream. was (Author: tresf): [~tilman] I had just assumed it was some type of circular reference, but I don't know the PDF specification well enough to understand. Slightly off-topic, but in regards to the out-of-memory stuff, is there a way (perhaps through an enhancement) to ignore the transparency-forces-rasterization behavior? As the printers get higher and higher quality, this can really be a detriment. Assuming this request is preposterous, could this be (easily) done programmatically using the API, so that downstream care can be taken to avoid transparency-forced rasterization at all costs? This technique may be common for some (e.g. manuals , brochures), but it's very rare for others (e.g. reports, labels, folios, invoices) and having control over this would help guard the memory issues. > Recursion is too deep > --------------------- > > Key: PDFBOX-5022 > URL: https://issues.apache.org/jira/browse/PDFBOX-5022 > Project: PDFBox > Issue Type: Bug > Components: Rendering > Affects Versions: 2.0.21 > Reporter: Tres Finocchiaro > Priority: Major > Attachments: crash_pdf.pdf > > > *OS:* Windows 10 x86_64 > *Java:* 1.8.0u271 x86_64 > The attached A4 PDF is created with TCPDF 6.2.26. > * At 600dpi, the attached PDF prints over > 4GB (-Xmx4096m) heap and crashes > the JVM. > * At 300dpi, the attached PDF prints under 2GB (-Xmx2048m) heap but shows > the following warning(s). > Google Chrome and Adobe can print this PDF without issue. The crash only > occurs using PDFBOX. > * Calling "*PDFDebugger*" does NOT crash the JVM. > * Calling "*PrintPDF*" DOES crash the JVM. > For reference, the public mailing list where this was reported: > [https://groups.google.com/g/qz-print/c/REV9fP3rqeM/m/OW1Gf2QZBQAJ] > > {code:java} > Nov 19, 2020 10:53:46 PM > org.apache.pdfbox.contentstream.operator.graphics.DrawObject process > SCHWERWIEGEND: recursion is too deep, skipping form XObject > Nov 19, 2020 10:53:46 PM > org.apache.pdfbox.contentstream.operator.graphics.DrawObject process > SCHWERWIEGEND: recursion is too deep, skipping form XObject > Nov 19, 2020 10:53:47 PM > org.apache.pdfbox.contentstream.operator.graphics.DrawObject process > SCHWERWIEGEND: recursion is too deep, skipping form XObject{code} > -- This message was sent by Atlassian Jira (v8.3.4#803005) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@pdfbox.apache.org For additional commands, e-mail: dev-h...@pdfbox.apache.org