<html> <head> <meta content="text/html; charset=UTF-8" http-equiv="Content-Type"> </head> <body bidimailui-detected-decoding-type="UTF-8" bgcolor="#FFFFFF" text="#000000"> On 09/28/2011 04:12 PM, Albert Astals Cid wrote: <blockquote cite="mid:4278039.aBFlyhzbAN@bluebox" type="cite"> <pre wrap="">Brian any chance you can have a look at this regression?</pre> </blockquote> hello albert, brian, other friends, i would happily collaborate to check the rtl issues once will brian spare some time and show the reason of his patch. but since this patch of brian is dealing with text selection, we could already allow page_get_text to directly invoke TextPage::getText that returns the text of the entire page as expected, rather than invoking poppler_page_get_selected_text, as it currently does, but wrong in my case, at least. here's a patch to head in this spirit. could you consider applying it? thanks in advance, alex n.b. i dare ask you to consider making a new 0.16.x release too. debian maintainers prefer stable releases :) . <blockquote cite="mid:4278039.aBFlyhzbAN@bluebox" type="cite"> <pre wrap=""> Albert A Dimecres, 28 de setembre de 2011, alex bodnaru vàreu escriure: </pre> <blockquote type="cite"> <pre wrap="">hello albert, brian, other friends i've did the research regarding the bug reported here, using git bisect, and found the following: i will look at the commit. db014ffb357e760d9397544c5a8fe747cdb497ab is the first bad commit commit db014ffb357e760d9397544c5a8fe747cdb497ab Author: Brian Ewins <a class="moz-txt-link-rfc2396E" href="mailto:brian.ewins@gmail.com"><brian.ewins@gmail.com></a> Date: Mon Nov 23 08:58:19 2009 +0000 Select top right to bottom left in RTL mode This makes pure RTL selection work. Bidi is not handled at all. Rendering of the selection is poor and the dumped text appears to still be in reverse order to me. :040000 040000 fbc5ddcd87a559cd94f20119eaee2af2fa9dc257 :e58e22d3707422029f1ca753868164eb22cf8bb4 M poppler On 09/27/2011 04:52 AM, alex bodnaru wrote: hello, On 09/26/2011 03:26 PM, Albert Astals Cid wrote: A Dissabte, 24 de setembre de 2011, alex bodnaru vàreu escriure: hello albert, friends, Hi about the "recent" defect, it got to me since the *recent* debian upgrade of libpoppler* from 0.12.4 strait to 0.16.7. after comparing the releases in between, i found the problem occured from 0.13.2 to 0.13.3. diff between these 2 is attached. That is not recent at all, that is older than a year ;-) no comment ;) In that release there were the following commits that touched TextOutputdev, I do not know if you know how to compile from git but if you do it would be great if you could try going back to 9c5612f6e013a8698eff6531ec388a7e6c1fb89a db014ffb357e760d9397544c5a8fe747cdb497ab b1d43fa052d9160c4f319a67415ecf3ebf2cf9b3 f83b677a8eb44d65698b77edb13a5c7de3a72c0f a2191a4d45e0abaec97c19aacae37c4c5824bd36 345ed51af9b9e7ea53af42727b91ed68dcc52370 12d83931ae1b899b70c7ea5c01f03f123b1bb9a8 thanks a lot. i'll certainly checkout each of them to get closer to the problem. And compile for each of them and see in which of those the bug is present and in which of them is not. Albert P.S: You still send html email ;-) hope to also have fixed this now. sorry again. best regards, alex just please look at the glib/demo/poppler-glib-demo get text output from the attached pdf, even of the fist page. On 09/18/2011 05:09 PM, Albert Astals Cid wrote: Please do not email me, email the list. A Diumenge, 18 de setembre de 2011, vàreu escriure: On 09/18/2011 02:41 PM, Albert Astals Cid wrote: A Diumenge, 18 de setembre de 2011, alex bodnaru vàreu escriure: hello friends, Hi thanks a lot albert for considering my problem. I am not considering your problem, I am complaining about the lack of information in your original mail ;-) i'm using poppler through python (that invokes glib interface). a recent change (probably together with get_text separation) broke the glib interface. what does recent mean? 0.16.7? 0.17.x? git master? 0.16.7. So 0.16.7 does not work, which is the version you know it works? Albert P.S: Would it be possible for you not to send HTML email? Albert thanks again, alex i can't load the entire page text with get_text (see the glib demo) of one pdf i have, but pdftotext does output the entire text. my pdf is attached. i apology for the language, but i promise it's a non offending cadastre report. please see that not all text lines are being output by get_text. could you help? thanks in advance, alex _______________________________________________ > poppler mailing list > > <a class="moz-txt-link-abbreviated" href="mailto:poppler@lists.freedesktop.org">poppler@lists.freedesktop.org</a> > > <a class="moz-txt-link-freetext" href="http://lists.freedesktop.org/mailman/listinfo/poppler">http://lists.freedesktop.org/mailman/listinfo/poppler</a> _______________________________________________ poppler mailing list <a class="moz-txt-link-abbreviated" href="mailto:poppler@lists.freedesktop.org">poppler@lists.freedesktop.org</a> <a class="moz-txt-link-freetext" href="http://lists.freedesktop.org/mailman/listinfo/poppler">http://lists.freedesktop.org/mailman/listinfo/poppler</a> _______________________________________________ poppler mailing list <a class="moz-txt-link-abbreviated" href="mailto:poppler@lists.freedesktop.org">poppler@lists.freedesktop.org</a> <a class="moz-txt-link-freetext" href="http://lists.freedesktop.org/mailman/listinfo/poppler">http://lists.freedesktop.org/mailman/listinfo/poppler</a> </pre> </blockquote> <pre wrap="">_______________________________________________ poppler mailing list <a class="moz-txt-link-abbreviated" href="mailto:poppler@lists.freedesktop.org">poppler@lists.freedesktop.org</a> <a class="moz-txt-link-freetext" href="http://lists.freedesktop.org/mailman/listinfo/poppler">http://lists.freedesktop.org/mailman/listinfo/poppler</a> </pre> </blockquote> </body> </html>