man/epydoc.1 | 48 24 + 24 - 0 !
man/epydocgui.1 | 8 4 + 4 - 0 !
2 files changed, 28 insertions(+), 28 deletions(-)

 make minor tweaks to manpage syntax.

man/epydoc.1 | 2 1 + 1 - 0 !
man/epydocgui.1 | 2 1 + 1 - 0 !
2 files changed, 2 insertions(+), 2 deletions(-)

 fix minor typos in manpages, as pointed out by lintian.

man/epydocgui.1 | 1 1 + 0 - 0 !
1 file changed, 1 insertion(+)

 mention debian contributions in the manpage.

epydoc/gui.py | 2 0 + 2 - 0 !
1 file changed, 2 deletions(-)

 remove shebang (#!) from .py files to comply with policy.

epydoc/apidoc.py | 2 1 + 1 - 0 !
1 file changed, 1 insertion(+), 1 deletion(-)

 get rid of string exceptions.
 One of the changes brought by Python 2.6 is the removal of string
 exceptions.  A mass bug filing identified Epydoc as having potential
 problems.  I later spot-checked all of the exceptions in the code, and I
 believe this is the only one that we have to worry about.
Bug-Debian: http://bugs.debian.org/585290

epydoc/markup/restructuredtext.py | 24 13 + 11 - 0 !
1 file changed, 13 insertions(+), 11 deletions(-)

 handle problems encountered with docutils 0.6.
 The problem here is that the child.data element does not always exist any
 more.  Apparently, the child element is sometimes a string instead.  So, we
 work around it by only executing the code in question if child.data can be
 referenced. Thanks to Thomas Hille for research and the initial patch.
Bug-Debian: http://bugs.debian.org/561793

epydoc/docparser.py | 29 28 + 1 - 0 !
1 file changed, 28 insertions(+), 1 deletion(-)

 fix the tokenizer so comment docstrings work with python 2.6.
Bug: https://sourceforge.net/tracker/index.php?func=detail&aid=2585292&group_id=32455&atid=405618
Bug-Debian: http://bugs.debian.org/590112

epydoc/docwriter/latex.py | 2 2 + 0 - 0 !
1 file changed, 2 insertions(+)

 do a better job of handling latex special entities.
 Mike's original problem was handling special << >> quote characters, i.e.
 from a French keyboard.  He solved this problem by adding the fontenc
 package.  In testing, I found that this change caused a HUGE slowdown for
 PostScript and PDF output, i.e. on the order of several minutes for dvips
 to render the result.  After digging around on Google, I added the lmodern
 package and that seems to have solved the problems I observed.
Bug-Debian: http://bugs.debian.org/614570

doc/index.html | 6 2 + 4 - 0 !
1 file changed, 2 insertions(+), 4 deletions(-)

 remove sourceforge logo and replace with "sourceforge".
 This fixes Lintian warning "privacy-breach-logo".  The IMG URL is considerd
 a privacy breach by Debian because it associates a user's browser info with
 use of the package.

epydoc/cli.py | 8 7 + 1 - 0 !
epydoc/docwriter/html.py | 18 14 + 4 - 0 !
man/epydoc.1 | 6 6 + 0 - 0 !
3 files changed, 27 insertions(+), 5 deletions(-)

 add --no-include-build-time option to allow reproducible builds.
 This patch was contributed by the Debian Reproducible Builds effort [1].
 [1]: https://wiki.debian.org/ReproducibleBuilds
Bug: https://sourceforge.net/p/epydoc/bugs/367/
Bug-Debian: https://bugs.debian.org/783326

epydoc/docwriter/html.py | 10 9 + 1 - 0 !
epydoc/docwriter/latex.py | 10 8 + 2 - 0 !
man/epydoc.1 | 26 26 + 0 - 0 !
3 files changed, 43 insertions(+), 3 deletions(-)

 use source_date_epoch, if available, rather than the current date.
 A lot of Debian packages rely on Epydoc during their build process.  By
 default, Eypdoc-generated documentation is "unreproducible", meaning that the
 content of the generated files changes from build to build even if the source
 tree does not. 
 .
 This patch changes Epydoc to use a deterministic date taken from environment
 variable SOURCE_DATE_EPOCH, instead of using the current date.  This helps
 make the Epydoc-generated documentation more predictible.
 .
 This patch was contributed by the Debian Reproducible Builds effort [1].
 [1]: https://wiki.debian.org/ReproducibleBuilds/TimestampsProposal
Bug: https://sourceforge.net/p/epydoc/bugs/368/
Bug-Debian: http://bugs.debian.org/790899

epydoc/markup/pyval_repr.py | 15 12 + 3 - 0 !
1 file changed, 12 insertions(+), 3 deletions(-)

 hide memory references in the output.
 A lot of Debian packages rely on Epydoc during their build process.  By
 default, Eypdoc-generated documentation is "unreproducible", meaning that the
 content of the generated files changes from build to build even if the source
 tree does not. 
 .
 Epydoc uses repr() on objects that don't always override the default one. As a
 consequence, the documentation shows memory references on the build system,
 e.g. <module.Foo instance at 0x7f3aff4299e0>.  This patch detects objects with
 no overriding of __repr__ and provides a fallback that does not show the
 reference.  It also takes care of not calling default __repr__ on functions. 
 .
 This patch was contributed by the Debian Reproducible Builds effort [1].
 [1]: https://wiki.debian.org/ReproducibleBuilds
Bug: https://sourceforge.net/p/epydoc/bugs/369/
Bug-Debian: http://bugs.debian.org/795826

epydoc/docwriter/html.py | 7 5 + 2 - 0 !
1 file changed, 5 insertions(+), 2 deletions(-)

 make class and module ordering predictable.
 A lot of Debian packages rely on Epydoc during their build process.  By
 default, Eypdoc-generated documentation is "unreproducible", meaning that the
 content of the generated files changes from build to build even if the source
 tree does not. 
 .
 Class trees written by Epydoc list classes in an order that varies across
 builds.  This patch applies a total order on classes (orders on class name,
 then module, then package, then parent package, etc.) instead of a partial one
 (only class name).  It also adds ordering for modules.
 .
 This patch was contributed by the Debian Reproducible Builds effort [1].
 [1]: https://wiki.debian.org/ReproducibleBuilds
Bug: https://sourceforge.net/p/epydoc/bugs/370/
Bug-Debian: http://bugs.debian.org/795835

epydoc/docbuilder.py | 4 2 + 2 - 0 !
epydoc/docwriter/html.py | 3 2 + 1 - 0 !
2 files changed, 4 insertions(+), 3 deletions(-)

 make subclass traversal and input file handling deterministic

epydoc/markup/pyval_repr.py | 2 2 + 0 - 0 !
1 file changed, 2 insertions(+)

 remove memory addresses for weakref.* instances
 Instances of the weakref.WeakKeyDictionary, weakref.WeakValueDictionary
 and weakref.WeakSet classes have a __repr__ string which contains their
 memory address. This should be skipped when printing their values for
 reproducibility of epydoc output.