Re: [Reproducible-builds] dropping texlive from our archive
Hi Mattia, thanks for "picking up texlive-bin"…! On Tue, May 31, 2016 at 09:52:43PM +, Mattia Rizzolo wrote: > 1+2 were already tried, with very poor results, with both upstream and > debian maintainer not interested at all at even listening to us. I think this summary is too harsh. > I think that before going to make a choice on it (even if that choice is > "prod upstream some more") we should try to move aside the patched > texlive-bin package from our archive and see how many packages are > actually affected by this. I think that's reasonable indeed. -- cheers, Holger signature.asc Description: Digital signature ___ Reproducible-builds mailing list Reproducible-builds@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/reproducible-builds
Re: [Reproducible-builds] dropping texlive from our archive
On Wed, Jun 01, 2016 at 10:16:40AM +0200, Alexis Bienvenüe wrote: > Note that when the TeX primitives do not honour S_D_E, xetex output is > not reproducible (even if \today is not used) because of the /Creator > string which includes a date that is build from these primitives : > > https://anonscm.debian.org/cgit/debian-tex/texlive-bin.git/tree/texk/web2c/xetexdir/xetex.web#n13872 > > So solution 3 should include a bugfix to make xetex's Creator string > either strip the build date either honour S_D_E. Consider that that's a thing in the PDF header. We should be able to have that thing not to use the same primitives, or anyway have that follows SDE. Such a part alone might have a chance to be accepted upstream, as IIUC their main concern was about changing the actual visible output, but they should be fine with changing the headers (as they already did in other parts). -- regards, Mattia Rizzolo GPG Key: 66AE 2B4A FCCF 3F52 DA18 4D18 4B04 3FCD B944 4540 .''`. more about me: https://mapreri.org : :' : Launchpad user: https://launchpad.net/~mapreri `. `'` Debian QA page: https://qa.debian.org/developer.php?login=mattia `- signature.asc Description: PGP signature ___ Reproducible-builds mailing list Reproducible-builds@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/reproducible-builds
Re: [Reproducible-builds] dropping texlive from our archive
Hi. Le 31/05/2016 23:52, Mattia Rizzolo a écrit : > The current status is: nearly everything was merged, but one patch. > https://anonscm.debian.org/git/reproducible/texlive-bin.git/tree/debian/patches/source-date-epoch-tex-primitives-defaults-to-1?id=c6dede514f6c724228ad6bbaa90531b27e78f1eb > That patch is basically what makes \today reproducible follow S_D_E > always instead of when somebody enables it with Yet Another Env Var. Note that when the TeX primitives do not honour S_D_E, xetex output is not reproducible (even if \today is not used) because of the /Creator string which includes a date that is build from these primitives : https://anonscm.debian.org/cgit/debian-tex/texlive-bin.git/tree/texk/web2c/xetexdir/xetex.web#n13872 So solution 3 should include a bugfix to make xetex's Creator string either strip the build date either honour S_D_E. Regards, Alexis Bienvenüe. ___ Reproducible-builds mailing list Reproducible-builds@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/reproducible-builds
[Reproducible-builds] dropping texlive from our archive
Hi humans. I'd like to take a deeper stab at texlive. The current status is: nearly everything was merged, but one patch. https://anonscm.debian.org/git/reproducible/texlive-bin.git/tree/debian/patches/source-date-epoch-tex-primitives-defaults-to-1?id=c6dede514f6c724228ad6bbaa90531b27e78f1eb That patch is basically what makes \today reproducible follow S_D_E always instead of when somebody enables it with Yet Another Env Var. As HW42 pointed out on IRC what's left to us is: 1) convice upstream 2) convince Debian maintainer 3) manualy fix packages 4) export a tex specific variable in debhelper In my opinion: 1+2 were already tried, with very poor results, with both upstream and debian maintainer not interested at all at even listening to us. We could still try again though. 4 is a very very ugly thing which would set a very bad precendent for debhelper. 3 is instead a reasonable choice: according to my codesearch.d.n lookup there are 149 packages¹ with '\today' somewhere in their code. I expect that way less than half of them are actually going to effect output, but I'm going to put my idea of worst case here to half: 75 packages. It would still be a reasonable enough number of packages to fix. I think that before going to make a choice on it (even if that choice is "prod upstream some more") we should try to move aside the patched texlive-bin package from our archive and see how many packages are actually affected by this. ¹ the list of the resulting packages are attached to this email, would be nice if somebody could check it against reproducible.json (thing that I'd do before dropping texlive-bin to be able to compare the results). -- regards, Mattia Rizzolo GPG Key: 66AE 2B4A FCCF 3F52 DA18 4D18 4B04 3FCD B944 4540 .''`. more about me: https://mapreri.org : :' : Launchpad user: https://launchpad.net/~mapreri `. `'` Debian QA page: https://qa.debian.org/developer.php?login=mattia `- signature.asc Description: PGP signature ___ Reproducible-builds mailing list Reproducible-builds@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/reproducible-builds
