Apache OpenOffice (AOO) Bugzilla – Issue 83934
binary-only blob in vcl
Last modified: 2008-01-29 14:01:28 UTC
cws beppec56pdf1b adds a new header to vcl: vcl inc/vcl/sRGB-IEC61966-2.1.hxx 1.2 beppec56 i59651 which is bascially a binary-only blob and therefore a binary-only file (strictly speaking) and therefore a P1 (as discussed in HH). Issue 81093 talks about this and even has a clean solution which is blocked by Sun approval... This *has* to be fixed asap.
approval done.
and why is this set to FIXED? Where is this fixed? This is still present in OOH680_m1. reopening. Please fix.
reassign to beppec56
As far as I remember this is going to be fixed by CWS beppec56pdffix02. However how this is P1 I fail to see.
it's about non-free, binary-only things in the source and we agreed in Hamburg that such bugs *are* P1
beppec56: please add this issue to your cws, too
How about a quick reality check ? We're talking about a data file in a well defined format here (an ICC stream). The same way we provide icons, fonts and lots of other binary files, only this one is encapsulated in a header since the tool that would generate said header is not yet checked in. "Nonfree" does not apply in any case, the header in question and obviously the contained file are LGPL 2.1 (courteously provided by beppec56). The original ICC profile is even less restrictive licensed, it can be distributed by anyone. So what exactly about this issue is so outrageous, that it MUST PREVENT the next build ?
> How about a quick reality check ? We're talking about a data file in a well > defined format here (an ICC stream). The same way we provide icons, fonts and > lots of other binary files, ... and it's still software, so it' still has to be free software to be able to be redistributed by Debian at all. And be it only in the source. > only this one is encapsulated in a header since the > tool that would generate said header is not yet checked in. Which is to be fixed. Which this issue is about. > "Nonfree" does not apply in any case, the header in question and obviously the > contained file are LGPL 2.1 (courteously provided by beppec56). ... Eh... I can also take a RFC (which is forbidden to be edited), make a hex out of it and put a LGPL blurb about it. Still would violate the license, and you are not allowed to re-license it anyway unless you are the copyright holder or the license of the original file explicitely allows re-licensing. > The original ICC rofile is even less restrictive licensed, it can be > distributed by anyone. Wow. Distributed. Read the Open Source definition, please. Distribution is only *one* of the possibilities you need to have. You also need to have the right to modify it and distribute modified versions, too. See my comment "------- Additional comments from rene Tue Aug 28 14:17:33 +0000 2007 -------" at Issue 81093: ""provided that the files are not changed [...]" is in the license. > So what exactly about this issue is so outrageous, that it MUST PREVENT the > next build ? That we have transformed non-free files into a blob and have put it into a header used by vcl.
> So what exactly about this issue is so outrageous, that it MUST PREVENT the > next build ? What's so outrageous is that non-conditional stuff gets added relying on non-free software. Of course, this can be fixed via beppecpdffix02. beppec56: What is the status of this cws? It is P1 because it was agreed on an ESC Meeting that it is. That your definition of P1 means "stop the build" is not my problem, I just know this must be fixed ASAP.
Accepted, added to cws beppec56pdffix02. A few lines on the current cws status. On CVS the cws still misses the new 'icc' module, the one part that should go to HEAD, since when I started adding it I quickly realized I was going to mess things up with the CVS data base. So I stopped, and asked for help. Help should be on the way. At my end the cws code builds under Linux and Windows, though under Windows it needs the cygwin g++ compiler, not mentioned anywhere on the build requirement documentation, but requested by the office configure process. Anyway it seems that under Windows g++ it's now required because of guw.exe. I still need help to build it under other platforms though. The current ICC embedded as a 'binary blob' in the OOo code was already generated by me using this library, so it has the LGL 2.1 license and the contents it's generated, not edited, as can be seen analyzing a PDF/A-1a file exported by OOo (both SRC680_m242 and OOH680_m1 contain the this code). So, in the end, the 'blob' it's not an edited version, but a new one, recomputed according to IEC61966-2.1:1999 standard.
P1 on blocking the next build is not "my definition". Have actually ever read the pririty meanings ? Hint: http://www.openoffice.org/scdocs/ddIssues_EnterModify.html#priority However you agreed that this can be fixed in a normal CWS fashion. That automatically makes it a non P1 issue. Moreover it's a duplicate, which you stated yourself even when filing it.
Set Fixed, and reassign to pl. Please verify it in cws beppec56pdffix02
reassigning...
verified in CWS beppec56pdffix02
seen in OOH680m5, closing