[Reproducible-builds] Bug#822268: transdecoder: please make the build reproducible

[Dhole]

Source: transdecoder 
Version: 2.0.1+dfsg-2 
Severity: wishlist
Tags: patch
User: reproducible-builds at lists.alioth.debian.org
Usertags: randomness
X-Debbugs-Cc: reproducible-builds at lists.alioth.debian.org

Hi,

While working on the "reproducible builds" effort [1], we have noticed
that transdecoder could not be built reproducibly.

When generating some sample files (sample_data/transcripts.fasta) a Perl
script is used (cufflinks_gtf_genome_to_cdna_fasta.pl) that loops
through a dictionary producing unreproducible output.

The attached patch fixes this by setting the env var PERL_HASH_SEED to 0
before calling this script to get a reproducible output when looping
through the hash keys. Once applied, transdecoder can be built
reproducibly in our current experimental framework.

 [1]: https://wiki.debian.org/ReproducibleBuilds

Regards,
-- 
Dhole
-------------- next part --------------
diff -Nru transdecoder-2.0.1+dfsg/debian/changelog transdecoder-2.0.1+dfsg/debian/changelog
--- transdecoder-2.0.1+dfsg/debian/changelog	2015-12-29 00:54:39.000000000 +0100
+++ transdecoder-2.0.1+dfsg/debian/changelog	2016-04-22 20:42:48.000000000 +0200
@@ -1,3 +1,11 @@
+transdecoder (2.0.1+dfsg-2.1) UNRELEASED; urgency=medium
+
+  * Non-maintainer upload.
+  * Set a hash seed for Perl before generating sample_data/transcripts.fasta
+    to get a reproducible output. 
+
+ -- Eduard Sanou <dhole at openmailbox.org>  Fri, 22 Apr 2016 20:42:02 +0200
+
 transdecoder (2.0.1+dfsg-2) unstable; urgency=medium
 
   [ Andreas Tille ]
diff -Nru transdecoder-2.0.1+dfsg/debian/patches/reproducible-sample_data.patch transdecoder-2.0.1+dfsg/debian/patches/reproducible-sample_data.patch
--- transdecoder-2.0.1+dfsg/debian/patches/reproducible-sample_data.patch	1970-01-01 01:00:00.000000000 +0100
+++ transdecoder-2.0.1+dfsg/debian/patches/reproducible-sample_data.patch	2016-04-22 20:43:56.000000000 +0200
@@ -0,0 +1,16 @@
+Description: Make the sample data reproducible
+ Set a fixed hash seed for Perl before generating sample_data/transcripts.fasta
+ to get reproducible results
+Author: Eduard Sanou <dhole at openmailbox.org>
+
+--- transdecoder-2.0.1+dfsg.orig/sample_data/runMe.sh
++++ transdecoder-2.0.1+dfsg/sample_data/runMe.sh
+@@ -26,7 +26,7 @@ fi
+ ../util/cufflinks_gtf_to_alignment_gff3.pl transcripts.gtf > transcripts.gff3
+ 
+ ## generate transcripts fasta file
+-../util/cufflinks_gtf_genome_to_cdna_fasta.pl transcripts.gtf test.genome.fasta > transcripts.fasta 
++PERL_HASH_SEED=0 ../util/cufflinks_gtf_genome_to_cdna_fasta.pl transcripts.gtf test.genome.fasta > transcripts.fasta 
+ 
+ ## Extract the long ORFs
+ ../TransDecoder.LongOrfs -t transcripts.fasta
diff -Nru transdecoder-2.0.1+dfsg/debian/patches/series transdecoder-2.0.1+dfsg/debian/patches/series
--- transdecoder-2.0.1+dfsg/debian/patches/series	2015-12-29 00:54:39.000000000 +0100
+++ transdecoder-2.0.1+dfsg/debian/patches/series	2016-04-22 20:43:08.000000000 +0200
@@ -1,3 +1,4 @@
 cd-hit-est-rename
 fix-whatis
 reproducible-build
+reproducible-sample_data.patch
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/reproducible-builds/attachments/20160422/92e0113c/attachment.sig>