GNOME Bugzilla – Bug 320237
bug in the decoding of quoted printable
Last modified: 2006-08-12 13:06:51 UTC
Please describe the problem: The decoding of quoted-printable strings ist not working correctle in addressbook/libebook/e-vcard.c The bug can be very easily verified with the program dump-vcard in addressbook/tests/vcard. Steps to reproduce: 1. In addressbook/tests/vcard run: ./dump-vcard 5.vcf > 5.vcf.decoded 2. Open 5.vcf.decoded in a hex-editor. 3. Compare the result of the decoding with the original strings. Actual results: Each quoted-printable sequence of the form =ab is translated in one 16bit value instead of in one 8bit value. Expected results: Each =ab is one 8bit value 0xab in the output. Does this happen every time? yes Other information: This patch: diff -u -r evolution-data-server/addressbook/libebook/e-vcard.c evolution-data-server.patched/addressbook/libebook/e-vcard.c --- evolution-data-server/addressbook/libebook/e-vcard.c 2005-09-15 13:49:20.000000000 +0200 +++ evolution-data-server.patched/addressbook/libebook/e-vcard.c 2005-10-30 10:14:38.000000000 +0100 @@ -236,7 +236,7 @@ */ } else if (isxdigit(a) && isxdigit (b)) { - gunichar c; + char c; a = tolower (a); b = tolower (b); @@ -244,7 +244,7 @@ c = (((a>='a'?a-'a'+10:a-'0')&0x0f) << 4) | ((b>='a'?b-'a'+10:b-'0')&0x0f); - g_string_append_unichar (str, c); + g_string_append_c (str, c); } else {
Created attachment 54081 [details] [review] Patch witch fixes the bug.
also see bug 322125; confirming
I'm a little bit frustrated, because this bug is in the 1.6 release. Does it make any sense to submit bug reports and fixes? Or is this work worth nothing? Do the Gnome-Devlopers ignore other languages than english?! Felix
*** Bug 322125 has been marked as a duplicate of this bug. ***
Reading the source, W/o testing the fix or knowing more, Approving this patch for commit to both HEAD and the gnome-2-14 branch. Addressbook hackers : Please take a look if I should know better. Felix : Let me know if you cannot commit this directly and if you want me to.
I have no write acess to cvs, so i cannot commit it myself. Please do you commit it. Thanks!
Hi Felix, this fix is already covered by the following commit by Devashish which was made four weeks ago: http://cvs.gnome.org/viewcvs/evolution-data-server/addressbook/libebook/e-vcard.c?r1=1.26&r2=1.27 This has been only committed to CVS HEAD (unstable evolution version 2.7.x) though, not to the stable 2.6 version, as it breaks API. So this will definitely be fixed in the next stable release, Evolution 2.8. Thanks a lot for your patience. It does make sense to report bugs and to attach patches here and I would like to encourage you to report issues to help improving Evolution and the GNOME Desktop, but the Evolution Team perhaps simply lacks human resources - at least we do encounter a huge amount of traffic in terms of patches and bugs, more than we manage to handle as we would like to. :-/ Harish, Devashish: Should the attached patch (which does not break API) be committed to the stable branch so it will be included in Evolution 2.6.3? Devashish already committed a more general patch (which includes this partial fix) to CVS HEAD.
andre no need to commit this patch , as a more general patch is already committed to head. Anyway thanks for the patch Felix. Closing this as fixed.
great! i'm looking forward to evolution 2.8. thanks.
devashish: my question was if this patch should be committed to *stable* as it does not change API but improves the 2.6.x situation.
ok andre commit this to stable
damn, i slept away the time. obsolete now, as 2.6.3 is already out. :-/