Description: I export a presentation as pdf. In the pdf when I copy the text from contents into clipboard the Whitespaces are removed. Steps to Reproduce: 1. Open attached presentation in Impress. 2. Export presentation as pdf. 3. Open pdf. 4. Select text from contents and copy to clipboard. 5. Text in the clipboard has no whitespaces. Actual Results: When I paste the text from the clipboard somewhere whitespaces are removed. Example: In step 4 I select "The attributes of the 144000 (9 slides)" After step 5 the clipboard contains "Theattributesofthe144000(9slides)" Expected Results: The whitespaces should be in the clipboard. Example: In step 4 I select "The attributes of the 144000 (9 slides)" After step 5 the clipboard should contain: "The attributes of the 144000 (9 slides)" Reproducible: Always User Profile Reset: No Additional Info: Version: 6.3.4.2 Build ID: 60da17e045e08f1793c57c00ba83cdfce946d0aa CPU threads: 8; OS: Mac OS X 10.14.6; UI render: default; VCL: osx; Locale: en-US (en_DE.UTF-8); UI-Language: en-US Calc: threaded
Created attachment 157160 [details] a single slide powerpoint with a list of contents
what PDF viewer do you use? I can't repro it on Windows in Версия: 6.5.0.0.alpha0+ (x64) ID сборки: 147af9e2cf7f937ed83ab00574b6a418a2cb629e Потоков ЦП: 4; ОС:Windows 10.0 Build 18362; Отрисовка ИП: по умолчанию; VCL: win; Локаль: ru-RU (ru_RU); Язык интерфейса: ru-RU Calc: threaded and I use PDF-Xchange viewer there so possibly it's mac only or problem with only some PDF viewer?
I can export the pdf in two ways: 1. File -> Export As -> Export Directly as PDF 2. File -> Print -> PDF -> Save PDF as Both ways give the same result as reported. PDF viewers I use: - Preview App - PDiff Express 2 App I can reproduce this with both pdf viewers. I found this issue with "PDiff Express 2" when I compared the pdf produced with powerpoint with the pdf produced with Impress because it showed me a diff I didn't expect. The Preview App shows whitespaces but misses them when I copy to clipboard, the "PDiff Express 2" does not show whitespaces. It looks like the text-layer in the pdf has no whitespaces?
I can also reproduce with Version: 6.4.0.2 Build ID: 08d19fecdc7a2298d051e19cfdb7c35544855fc3 CPU threads: 8; OS: Mac OS X 10.14.6; UI render: default; VCL: osx; Locale: en-US (en_DE.UTF-8); UI-Language: en-US Calc: threaded
repro in Версия: 6.5.0.0.alpha0+ ID сборки: b6295e4a1b7735c148174f44f6d28221f4f52302 Потоков ЦП: 4; ОС:Mac OS X 10.15.2; Отрисовка ИП: GL; VCL: osx; Локаль: ru-RU (ru_RU.UTF-8); Язык интерфейса: ru-RU Calc: threaded I used the standard PDF viewer in macOS
I tested PDF from macOS's LibreOffice on Linux machine with Okular PDF viewer. Text pasted from that PDF into document looks fine, it has all spaces. I think it isn't a problem in LibreOffice, but it's a macOS PDF viewer's problem. not a bug? Xisco, any opinions?
Created attachment 157193 [details] PDF from macOS LO
Created attachment 157194 [details] PDF from Powerpoint
(In reply to Roman Kuznetsov from comment #6) > I tested PDF from macOS's LibreOffice on Linux machine with Okular PDF > viewer. Text pasted from that PDF into document looks fine, it has all > spaces. > I think it isn't a problem in LibreOffice, but it's a macOS PDF viewer's > problem. > > not a bug? Xisco, any opinions? I think it is not a problem with macOS PDF viewer, because the whitespaces are there when I use the PDF from Powerpoint which I attached. Notice the difference of the file size: PDF from Powerpoint has 30.7KB and the PDF from LO (see attachment) is 11.53 KB respectively 25 KB when I create it from LO.
For me neither the LO or PowerPoint produced PDF has whitespace included when I copy from Preview In Acrobat Reader both work Although yes, your PDF from PowerPoint shows correctly. How did you produce it? I'm on 16.34 So at the face of it I would call this maybe not our bug Version: 7.0.0.0.alpha0+ Build ID: 0cb4f304abf6f8dd6b40eb800788d2fe80581813 CPU threads: 4; OS: Mac OS X 10.14.6; UI render: default; VCL: osx; Locale: en-US (en_US.UTF-8); UI-Language: en-US Calc: threaded
Created attachment 157996 [details] screenshot of "pdiff express 2" showing the difference between pdf from lo (left side) and pdf from powerpoint (right side)
Created attachment 157997 [details] screenshot of "pdiff express 2" showing the difference between pdf from lo (left side) and pdf from powerpoint (right side)
(In reply to eisa01 from comment #10) > For me neither the LO or PowerPoint produced PDF has whitespace included > when I copy from Preview > > In Acrobat Reader both work > > Although yes, your PDF from PowerPoint shows correctly. How did you produce > it? I'm on 16.34 > > How did you produce it? File -> Print -> Save PDF > I'm on 16.34 What do you mean? I don't understand. > So at the face of it I would call this maybe not our bug > I attached a screenshot (see attachment 157996 [details]) from PDF comparison with "Pdiff Express 2" that shows the difference of the produced PDF's from LO (= left side) and from Powerpoint (= right side). There you clearly see, the whitespaces are missing in the PDF from LO. When "Pdiff Express 2" says both PDF's are equal then I am happy. I created both PDF's via File -> Print -> Save PDF (I rotated the PDF from LO in Preview before comparison due bug 130007) For the second screenshot (see attachment 157997 [details]) I did it vice versa: I rotated the PDF from Powerpoint before comparison to show that the rotation operation in Preview does not touch the text layer in the PDF. You can get "Pdiff Express 2" from the Apple Store for Mac. The trial version allows to compare the first three pages which is enough for testing purpose of this issue.
Created attachment 157998 [details] screenshot of "pdiff express 2" showing the difference between pdf from lo (left side) and pdf from powerpoint (right side) The PDF from LO is rotated due bug 130007.
(In reply to christoph_egger from comment #13) > > I attached a screenshot (see attachment 157996 [details]) from PDF > comparison with "Pdiff Express 2" that shows the difference of > the produced PDF's from LO (= left side) > and from Powerpoint (= right side). > There you clearly see, the whitespaces are missing in the PDF from LO. > > When "Pdiff Express 2" says both PDF's are equal then I am happy. > > I created both PDF's via File -> Print -> Save PDF > (I rotated the PDF from LO in Preview before comparison due bug 130007) > > For the second screenshot (see attachment 157997 [details]) I did it vice > versa: > I rotated the PDF from Powerpoint before comparison to show that the rotation > operation in Preview does not touch the text layer in the PDF. > > You can get "Pdiff Express 2" from the Apple Store for Mac. > The trial version allows to compare the first three pages which is enough > for testing purpose of this issue. I created Screenshot (see attachment 157998 [details]) of "pdiff express 2" *without* use of Preview before comparison.
In case you ask what is the relation of "pdiff express 2" to the clipboard (in relation to Comment 0): Answer: The content of the clipboard matches with "pdiff express 2" saying what is in the text layout of the pdf. Examples: "pdiff express 2" says "Introduction (1 slide)" and I have "Introduction (1 slide)" in the clipboard. "pdiff express 2" says "Introduc/on(1slide)" and I have "Introduc/on(1slide)" in the clipboard.
Since specifically the "ti" ligature in the word "introduction" is broken in the example this bug might be related to incorrect parsing of escape sequences in one of the programs or incorrect forwarding of escape sequenced text between them. How do the spaces in an export look like if you prefix all spaces in the original document with a dash "/"?
Created attachment 158968 [details] screenshot of "pdiff express 2" showing the difference between pdf from lo (left side) and pdf from powerpoint (right side) with all spaces prefixed with a slash Screenshot made per request in Comment 17
(In reply to Oliver Grimm from comment #17) > Since specifically the "ti" ligature in the word "introduction" is broken in > the example this bug might be related to incorrect parsing of escape > sequences in one of the programs or incorrect forwarding of escape sequenced > text between them. If you think that parsing of the PDF is incorrect how do you explain that parsing of the PDF from Powerpoint is correct? What data is necessary in the PDF export that *any* program has a chance for correct parsing? > How do the spaces in an export look like if you prefix all spaces in the > original document with a dash "/"? See attachment 158968 [details] FYI: I made the requested modification in Powerpoint.
Created attachment 159358 [details] powerpoint presentation with all spaces prefixed with a slash Same as attachment 157160 [details] but with all spaces prefixed with a slash
Ok, first of all you can use Export -> Export as PDF to avoid the rotation bug. However, that does not avoid the issue I also did an export test from LO on Windows, and that resulting file also misses spaces when shown in macOS Preview on Mojave. As such, this is not a mac bug However, in e.g., the Edge PDF viewer I'm able to copy paste the output just fine with spaces. I guess this is a question whether LO produce valid PDF output or not, or if this is a bug in Preview? Second question is then if we should fix the output to be able to be read by Preview or not (the now save as PDF in print dialog produce quite broken output when copy pasted due to the rotation, lots of extra spaces)
Created attachment 160267 [details] PDF from Windows LO 6.4
Created attachment 160268 [details] PDF from macOS LO 6.4 Export as PDF
(In reply to eisa01 from comment #21) > Ok, > first of all you can use Export -> Export as PDF to avoid the rotation bug. > That sounds like you can reproduce bug 130007. Please make a note there.
Created attachment 160307 [details] screenshot of "pdiff express 2" showing the diff between pdf from attachment 160267 [details] (left side) and pdf from attachment 160268 [details] (right side)
(In reply to eisa01 from comment #21) > Ok, > first of all you can use Export -> Export as PDF to avoid the rotation bug. > > However, that does not avoid the issue > > I also did an export test from LO on Windows, and that resulting file also > misses spaces when shown in macOS Preview on Mojave. As such, this is not a > mac bug > > However, in e.g., the Edge PDF viewer I'm able to copy paste the output just > fine with spaces. > > I guess this is a question whether LO produce valid PDF output or not, or if > this is a bug in Preview? Second question is then if we should fix the > output to be able to be read by Preview or not > > (the now save as PDF in print dialog produce quite broken output when copy > pasted due to the rotation, lots of extra spaces) I attached a screenshot from "pdiff express 2" comparing your two pdf's (see screenshot from attachment 160307 [details]). It shows that both pdf's have *no* whitespaces in the text layer. Question: Where does "Edge PDF viewer" get the whitespaces from when you copy the text?
(In reply to eisa01 from comment #21) > Ok, > first of all you can use Export -> Export as PDF to avoid the rotation bug. > > (the now save as PDF in print dialog produce quite broken output when copy > pasted due to the rotation, lots of extra spaces) When I do that from PowerPoint on Mac then the resulting PDF is correct. Can you confirm that?
Created attachment 164300 [details] screenshot of "pdiff express 2" showing the diff between pdf from ppt 2008 (left side) and pdf from ppt 2019 (right side) I recently bought MS Office 2019 for Mac. I produced a pdf and compared it with my older MS Office 2008 version. It is noticable that MS Office 2019 does also remove the whitespaces in the textlayer. I don't know what is the case with MS Office 2013 and 2016 but I think that explains why this is hard to reproduce for many people.
(In reply to christoph_egger from comment #28) > Created attachment 164300 [details] > screenshot of "pdiff express 2" showing the diff between pdf from ppt 2008 > (left side) and pdf from ppt 2019 (right side) > > I recently bought MS Office 2019 for Mac. > I produced a pdf and compared it with my older MS Office 2008 version. > > It is noticable that MS Office 2019 does also remove the whitespaces in the > textlayer. > I don't know what is the case with MS Office 2013 and 2016 but I think > that explains why this is hard to reproduce for many people. I restested with LO version Version: 7.0.0.3 Build ID: 8061b3e9204bef6b321a21033174034a5e2ea88e CPU threads: 8; OS: Mac OS X 10.14.6; UI render: default; VCL: osx Locale: en-US (en_DE.UTF-8); UI: en-US Calc: threaded It still does remove the whitespaces in the textlayer and it still produces "Introdu/on" instead of "Introduction". MS Office 2008 produces "Introduction (1 slide)" MS Office 2019 produces "Introduction(1slide)" LO 7.0.0 produces "Introdu/on(1slide)" The question that rises up: What is the correct behavior according to PDF specification? I figured out LO has the same issue with the word "Revelation": It produces "Revela/on".
(In reply to christoph_egger from comment #29) > (In reply to christoph_egger from comment #28) > > Created attachment 164300 [details] > > screenshot of "pdiff express 2" showing the diff between pdf from ppt 2008 > > (left side) and pdf from ppt 2019 (right side) > > > > I recently bought MS Office 2019 for Mac. > > I produced a pdf and compared it with my older MS Office 2008 version. > > > > It is noticable that MS Office 2019 does also remove the whitespaces in the > > textlayer. > > I don't know what is the case with MS Office 2013 and 2016 but I think > > that explains why this is hard to reproduce for many people. > > > I restested with LO version > > Version: 7.0.0.3 > Build ID: 8061b3e9204bef6b321a21033174034a5e2ea88e > CPU threads: 8; OS: Mac OS X 10.14.6; UI render: default; VCL: osx > Locale: en-US (en_DE.UTF-8); UI: en-US > Calc: threaded > > It still does remove the whitespaces in the textlayer > and it still produces "Introdu/on" instead of "Introduction". > > > MS Office 2008 produces "Introduction (1 slide)" > MS Office 2019 produces "Introduction(1slide)" > LO 7.0.0 produces "Introdu/on(1slide)" > > > The question that rises up: > What is the correct behavior according to PDF specification? > > I figured out LO has the same issue with the word "Revelation": > It produces "Revela/on". Restested with version Version: 7.1.0.0.alpha0+ Build ID: <buildversion> CPU threads: 8; OS: Mac OS X 10.14.6; UI render: default; VCL: osx Locale: en-US (en_DE.UTF-8); UI: en-US Calc: threaded I downloaded it from: https://dev-builds.libreoffice.org/daily/master/MacOSX-x86_64@tb81-TDF/2020-08-14_07.26.35/LibreOfficeDev_7.1.0.0.alpha0_MacOS_x86-64.dmg Same result as with 7.0.0.3 above
Created attachment 166928 [details] screenshot of "pdiff express 2" showing the difference between pdf from lo (left side) and pdf from powerpoint 2008 (right side) Replace screenshot from attachment 157998 [details] with a newer version now that bug 130007 is addressed
Created attachment 166929 [details] screenshot of "pdiff express 2" showing the difference between pdf from lo (left side) and pdf from powerpoint 2008 (right side) with all spaces prefixed with a slash Replace screenshot from attachment 158968 [details] with a new screenshot now that bug 130007 is addressed
Created attachment 166930 [details] screenshot of "pdiff express 2" showing the difference between pdf from lo (left side) and pdf from powerpoint 2019 (right side) with all spaces prefixed with a slash Same screenshot as attachment 166929 [details] but against powerpoint 2019
This bug is unreadable. I tried but I gave up. If there's a bug, looks like Mac only, or I didn't see where it was confirmed for other OS. christoph, please do not quote yourself. With new info like "Repro 7.0" you should obsolete same previous "Repro 6.3". Instead, make a complete comment that minimally summarizes the issue (you may refer to attachment 123), explain why this is not NotOurBug (use Adobe Reader for PDF), then please go from beginning, leave just valuable comments and mark all other "obsolete".
Let's restart the bug report that people can follow again: Description: I export a presentation as pdf. In the pdf when I copy the text from contents into clipboard the Whitespaces are removed or content changed. Steps to Reproduce: 1. Open attached presentation in Impress. 2. Export presentation as pdf. 3. Open pdf. 4. Select text from contents and copy to clipboard. 5. Text in the clipboard has no whitespaces. Actual Results: When I paste the text from the clipboard somewhere whitespaces are removed or content changed. Example: In step 4 I select "Introduction (1 slide)" After step 5 the clipboard contains "Introduc/on (1 slide)" or "Introduction(1slide)" Expected Results: The whitespaces should be in the clipboard and content unchanged. Example: In step 4 I select "Introduction (1 slide)" After step 5 the clipboard should contain: "Introduction (1 slide)" Reproducible: Always User Profile Reset: No Which PDF viewers do I use? 1. Preview App 2. PDiff Express 2 3. Adobe Acrobat Reader for Mac When I create the PDF with Powerpoint 2008 for Mac and follow the steps to reproduce then I get "Introduction (1 slide)" into the clipboard with all three PDF viewers. This is the expected result. When I create the PDF with Powerpoint 2019 for Mac and follow the steps to reproduce then I get "Introduction(1slide)" into the clipboard with Preview App and PDiff Express 2. With Acrobat Reader I get "Introduction (1 slide)" into the clipboard. The removal of the whitespace content is unexpected. When I create the PDF with LO 7.1+ and follow the steps to reproduce then I get "Introdu/on(1slide)" into the clipboard with Preview App and PDiff Express 2. With Adobe Acrobat Reader for Mac I get "Introduc/on (1 slide)" into the clipboard. This content change is unexpected. In Comment 17 there is the question what happens with a slash prefix to each whitespace. I created Attachment 159385 [details], Attachment 166929 [details] and attachment 166930 [details] in response to that question. The test results concerning the question in Comment 17 are these: When I create the PDF with Powerpoint 2008 for Mac and follow the steps to reproduce then I get "Introduction/ (1/ slide)" into the clipboard with all three PDF viewers. This is the expected result. When I create the PDF with Powerpoint 2019 for Mac and follow the steps to reproduce then I get "Introduction/(1/slide)" into the clipboard with Preview App and PDiff Express 2. With Adobe Acrobat I get "Introduction/ (1/ slide)" into the clipboard. The removal of the whitespace content is unexpected. When I create the PDF with LO 7.1+ and follow the steps to reproduce then I get "Introduc/on(1slide)" with Preview App and with "PDiff Express 2" into the clipboard. With Adobe Acrobat I get "Introduc/on/ (1/ slide)" into the clipboard. This content change is unexpected.
(In reply to christoph_egger from comment #35) > Let's restart the bug report that people can follow again: > > Description: > I export a presentation as pdf. In the pdf when I copy the text from > contents into clipboard the Whitespaces are removed or content changed. > > Steps to Reproduce: > 1. Open attached presentation in Impress. > 2. Export presentation as pdf. > 3. Open pdf. > 4. Select text from contents and copy to clipboard. > 5. Text in the clipboard has no whitespaces. > > Actual Results: > When I paste the text from the clipboard somewhere whitespaces are removed > or content changed. > > Example: > In step 4 I select "Introduction (1 slide)" > After step 5 the clipboard contains "Introduc/on (1 slide)" > or "Introduction(1slide)" Tested with attachment 157160 [details] and Okular and I don't see the problem NixOS Version: 7.3.0.0.alpha0+ / LibreOffice Community Build ID: 67e47070a7580a17804adce812cc2f98bfe7b51f CPU threads: 16; OS: Linux 5.13; UI render: default; VCL: x11 Locale: fi-FI (fi_FI.UTF-8); UI: en-US Calc: threaded
(In reply to Buovjaga from comment #36) > (In reply to christoph_egger from comment #35) > > Let's restart the bug report that people can follow again: > > > > Description: > > I export a presentation as pdf. In the pdf when I copy the text from > > contents into clipboard the Whitespaces are removed or content changed. > > > > Steps to Reproduce: > > 1. Open attached presentation in Impress. > > 2. Export presentation as pdf. > > 3. Open pdf. > > 4. Select text from contents and copy to clipboard. > > 5. Text in the clipboard has no whitespaces. > > > > Actual Results: > > When I paste the text from the clipboard somewhere whitespaces are removed > > or content changed. > > > > Example: > > In step 4 I select "Introduction (1 slide)" > > After step 5 the clipboard contains "Introduc/on (1 slide)" > > or "Introduction(1slide)" > > Tested with attachment 157160 [details] and Okular and I don't see the > problem > > NixOS > Version: 7.3.0.0.alpha0+ / LibreOffice Community > Build ID: 67e47070a7580a17804adce812cc2f98bfe7b51f > CPU threads: 16; OS: Linux 5.13; UI render: default; VCL: x11 > Locale: fi-FI (fi_FI.UTF-8); UI: en-US > Calc: threaded Noone on windows can reproduce either. This really is Mac OS X only. I can still reproduce on Version: 7.1.5.2 / LibreOffice Community Build ID: 85f04e9f809797b8199d13c421bd8a2b025d52b5 CPU threads: 12; OS: Mac OS X 10.15.7; UI render: default; VCL: osx Locale: en-GB (en_DE.UTF-8); UI: en-US Calc: threaded and on Version: 7.3.0.0.alpha0+ / LibreOffice Community Build ID: 4677345e3695bac158bb04048b4d5c608ed764b4 CPU threads: 12; OS: Mac OS X 10.15.7; UI render: default; VCL: osx Locale: en-US (en_DE.UTF-8); UI: en-US Calc: threaded
When opening the PDF in the LibreOffice, PDF Expert 2.5.18, Crome 92.0.4515.107 the error is not reproduced. I was able to reproduce the bug in OS X 11.0 (999.4) standard viewer when copying text missing spaces. If you remove the numbering and go through all the steps again, the error disappears. Additional Info: Version: 7.3.0.0.alpha0+ / LibreOffice Community Build ID: 246f24ebc8227345f13784d9e2f055d813f3d24f CPU threads: 12; OS: Mac OS X 10.15.7; UI render: default; VCL: osx Locale: ru-RU (ru_UA.UTF-8); UI: en-US Calc: threaded Version: 7.1.5.2 / LibreOffice Community Build ID: 85f04e9f809797b8199d13c421bd8a2b025d52b5 CPU threads: 12; OS: Mac OS X 10.15.7; UI render: default; VCL: osx Locale: ru-RU (ru_UA.UTF-8); UI: en-US Calc: threaded
(In reply to Denys Prokhorov from comment #38) > When opening the PDF in the LibreOffice, PDF Expert 2.5.18, Crome > 92.0.4515.107 the error is not reproduced. > > I was able to reproduce the bug in OS X 11.0 (999.4) standard viewer when > copying text missing spaces. > If you remove the numbering and go through all the steps again, the error > disappears. This is interesting discovery. I can confirm: With the numbering I get: Introduc/on(1slide) Without the numbering I get: Introduc/on (1 slide) Indeed, the problem with the whitespaces disappears. This problem with whitespaces is related to the numbering format. The other problem with the 'ti' -> '/' change is unrelated to it.
(In reply to Buovjaga from comment #36) > Tested with attachment 157160 [details] and Okular and I don't see the > problem > > NixOS > Version: 7.3.0.0.alpha0+ / LibreOffice Community > Build ID: 67e47070a7580a17804adce812cc2f98bfe7b51f > CPU threads: 16; OS: Linux 5.13; UI render: default; VCL: x11 > Locale: fi-FI (fi_FI.UTF-8); UI: en-US > Calc: threaded As per comment #21 this is present when exported from Windows and opened in Preview on macOS. It did not have the problem in Edge, so I guess Okular also doesn't have the problem It's a question whether it's rather a bug in Preview though, as e.g., opening the PDF in Firefox for macOS and copying out works fine Still present with PDFs produced in 7.5 if anyone wonders
All the spaces are there in the PDF, LibreOffice is exporting them correctly. Text extraction from PDF is mess and every PDF viewer has its own undocumented behavior and little bugs, there is not much we can do about it. The same goes for the ti ligature, all information needed to gets its textual representation is in the PDF, but the PDiff tool is ignoring it. The PDF from PowerPoint is behaving better by cheer luck, and I don’t see anything specific in its output that makes it behave differently. Unless someone can point to a specific PDF feature present in PowerPoint output and missing from ours, there is nothing to be done here.