[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#860804: marked as done (RFS: highwayhash/0~20170419-g1f4a24f-1 [ITP] -- tensorflow dependency library)

Your message dated Tue, 25 Apr 2017 07:44:15 +0200
with message-id <20170425054415.bzelvnbf5kjqummw@angband.pl>
and subject line Re: Bug#860804: RFS: highwayhash/0~20170419-g1f4a24f-1 [ITP] -- tensorflow dependency library
has caused the Debian Bug report #860804,
regarding RFS: highwayhash/0~20170419-g1f4a24f-1 [ITP] -- tensorflow dependency library
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@bugs.debian.org
immediately.)


-- 
860804: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=860804
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems
--- Begin Message --- 
Package: sponsorship-requests
Severity: wishlist

  Dear mentors,

  I am looking for a sponsor for my package "highwayhash"

 * Package name    : highwayhash
   Version         : 0~20170419-g1f4a24f-1
   Upstream Author : google
 * URL             : https://github.com/google/highwayhash
 * License         : Apache-2.0
   Section         : science

  It builds those binary packages:

    libhighwayhash-dev - Fast strong hash functions: SipHash/HighwayHash

  To access further information about this package, please visit the
following URL:

  https://mentors.debian.net/package/highwayhash


  Alternatively, one can download the package with dget using this command:

    dget -x https://mentors.debian.net/debian/pool/main/h/highwayhash/highwayhash_0~20170419-g1f4a24f-1.dsc

  More information about hello can be obtained from

http://debomatic-amd64.debian.net/distribution#experimental/highwayhash/0~20170419-g1f4a24f-1/buildlog

  Changes since the last upload:

highwayhash (0~20170419-g1f4a24f-1) experimental; urgency=medium

  * Initial release. (Closes: #848885)

--- End Message ---
--- Begin Message --- 
On Tue, Apr 25, 2017 at 04:48:00AM +0000, Lumin wrote:
> > It does look uploadable, yeah, even though there's a bunch of issues.  It's
> > up to you whether you want to get it good first or to upload present state
> > then improve it incrementally.  Please say what you prefer.
> 
> I updated the package again with the following changes:
> 
>  * changed Multi-Arch of the lib package to same
>  * detect architecture in rules to privide HH_AARCH64 and HH_POWER
> flag to upstream makefile,
>     which is possible to fix the arm64 and ppc64el builds. (oops, I
> have no such machines to test)

Alas, it failed on arm64 because of symbols file.  But then, C++ symbols are
so unpredictable I know of no better way than to run test builds and gather
the output.  This could be done on porterboxes but it takes less human time
to abuse the buildds.  This package is no libreoffice nor acl2, so the
burden on Debian infrastructure is negligible.

I didn't test on ppc* but it's probably same.

>  * added manpage highwayhash.3 (modified based on your snippet)
>  * added autopkgtest script (using modified version of your sanity test code)

> I think this package can be uploaded now. Further fix will be added
> incrementally.

Uploading something that works on amd64 allows packaging whatever you want
this library for, so let's get it past NEW then port it later.

> > On other architectures, the closest it came on x32 (a non-release arch) --
> > builds ok, fails only at dpkg-gensymbols due to symbol mismatches.
> > One warning: left shift count >= width of type [-Wshift-count-overflow]
> > sounds like it might be broken, though[1].
> >
> > On arm64 it fails with:
> > g++: error: unrecognized command line option '-mavx2'
> >
> > On armhf there's both the shift width issue, and:
> > error: #error "Port"
> >
> > On i386, all of the above, plus:
> > error: '_mm_cvtsi64_si128' was not declared in this scope
> 
> The HH_POWER and HH_AARCH64 flags are possible to fix the ppc64{,el} build and
> arm64 build. However, I'm not clear on upstream's attitude towards
> 32-bit architectures.
> 
> Actually I think 32-bit architectures are not my priority too, since
> upstreams of my deep learning packages don't put much attention on 32-bit
> architectures.

M'kay.  Unlike that deep learning thing, an universal hash is useful for
everyone, but it's not mandatory to do the porting.

> So I think this package can be uploaded now.

It's in NEW.

> Thank you for the manpage snippet, I've completed the manpage and
> added it to the -dev package.

I guess you gave me too much credit, I barely started it, copying in the
prototypes.


That's it for now, let's do a round 2 once the symbols files come in.

ᛗᛖᛟᚹ!
-- 
⢀⣴⠾⠻⢶⣦⠀ Meow!
⣾⠁⢠⠒⠀⣿⡁
⢿⡄⠘⠷⠚⠋⠀ Collisions shmolisions, let's see them find a collision or second
⠈⠳⣄⠀⠀⠀⠀ preimage for double rot13!

--- End Message ---
Reply to: