Bug 509160 (mine_detector) - Review Request: mine_detector – a mine-finding game
Summary: Review Request: mine_detector – a mine-finding game
Keywords:
Status: CLOSED ERRATA
Alias: mine_detector
Product: Fedora
Classification: Fedora
Component: Package Review
Version: rawhide
Hardware: All
OS: Linux
low
medium
Target Milestone: ---
Assignee: David A. Wheeler
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On: PragmARC
Blocks:
TreeView+ depends on / blocked
 
Reported: 2009-07-01 15:30 UTC by Björn Persson
Modified: 2011-07-03 23:33 UTC (History)
6 users (show)

Fixed In Version: mine_detector-6.0-3.fc12
Clone Of:
Environment:
Last Closed: 2010-02-18 22:33:31 UTC
Type: ---
Embargoed:
dwheeler: fedora-review+
gwync: fedora-cvs+


Attachments (Terms of Use)

Description Björn Persson 2009-07-01 15:30:10 UTC
Spec URL: http://www.rombobjörn.se/packages/mine_detector.spec
SRPM URL: http://www.rombobjörn.se/packages/mine_detector-5.0-1.fc11.src.rpm

Description:
Mine Detector is a mine-finding game with somewhat different rules from other mine-finding games. Mine Detector rarely requires guessing. Only at the higher levels may guessing sometimes be the only way to win.

The program uses the PragmAda Reusable Components, review request 509159.

RPMlint output:
mine_detector.spec: E: no-buildroot-tag
Since RPMbuild now ignores the buildroot tag I don't see why I would need one.

I should also mention that I need a sponsor.

By the way, Mine Detector doesn't use a flag icon to mark cells as mines, so it should be politically safe. ;-)

Comment 1 Andrew McNabb 2009-07-16 01:17:20 UTC
Björn, I'm not a sponsored package reviewer, but I have a few thoughts.

1) As you noted, rpmlint gave an error about the buildroot.  I would recommend adding something like "BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-root-%(%{__id_u} -n)".  I don't know anything about the line being ignored, but even if the line just makes it happy, it's can't hurt.

2) The package is called mine_detector, but the packaging guidelines prohibits using underscores.  Consider using a hyphen instead.  See https://fedoraproject.org/wiki/Packaging:NamingGuidelines#Separators

3) It's great that you have a description in two languages.

4) The package lists PragmARC-devel as a build dependency, but there does not seem to be any package with that name in Fedora 11.  As a result, I was unable to build a binary package.

You mentioned that PragmARC-devel is part of a separate request, but I think you should really work towards getting that one accepted before submitting new requests that depend on it.  Good luck.

Comment 2 Björn Persson 2009-07-18 20:32:59 UTC
About the underscore, the naming guidelines also say that "packages where the upstream name naturally contains an underscore are excluded from this". I figured that mine_detector would fall in this category, although I'm not sure how to tell a natural underscore from an unnatural one. The underscore is there in the names of both the source package and the final executable file, so I chose this name to comply with the rule that "the name should match the upstream tarball or project name from which this software came".

The underscore isn't important to me though. I can replace it with a hyphen if an experienced reviewer thinks I should.

Comment 3 Björn Persson 2009-07-23 21:18:12 UTC
Apparently the buildroot tag is still required for bureaucratic reasons, so I added one.

http://www.rombobjörn.se/packages/mine_detector.spec
http://www.rombobjörn.se/packages/mine_detector-5.0-2.fc11.src.rpm

Comment 4 Björn Persson 2009-08-07 23:57:59 UTC
Updated to version 6.0, and updated the URLs because the PragmAda site moved:

http://www.rombobjörn.se/packages/mine_detector.spec
http://www.rombobjörn.se/packages/mine_detector-6.0-1.fc11.src.rpm

Comment 5 Peter Lemenkov 2009-09-27 17:00:45 UTC
unblocking FE-NEEDSPONSOR - I just sponsored Björn Persson.

Comment 6 Björn Persson 2009-11-10 23:10:57 UTC
Fixed to link libgnat and libgcc dynamically:

http://www.rombobjörn.se/packages/mine_detector.spec
http://www.rombobjörn.se/packages/mine_detector-6.0-2.fc11.src.rpm

Comment 7 Björn Persson 2009-11-13 01:21:38 UTC
Now that PargmARC is in I've made a scratch build of Mine Detector. It worked without errors on i586, x86_64, ppc and ppc64:
http://koji.fedoraproject.org/koji/taskinfo?taskID=1804288

Comment 8 David A. Wheeler 2010-01-25 16:07:04 UTC
According to:
 http://fedoraproject.org/wiki/Packaging:LicensingGuidelines
Section "License Clarification":
"In cases where the licensing is unclear, it may be necessary to contact the copyright holders to confirm the licensing of code or content. In those situations, it is _always_ preferred to ask upstream to resolve the licensing confusion by documenting the licensing and releasing an updated tarball. However, this is not always possible to achieve. In such cases, it is acceptable to receive confirmation of licensing via email. A copy of the email, containing full headers, must be included as a source file (marked as %doc) in the package. This file is considered part of the license text. "

Although you've included a "license.txt" file, the package doesn't include the actual email as required.

Comment 9 David A. Wheeler 2010-01-25 22:51:00 UTC
First, here are some initial impressions after reading the spec file and looking over the code.

First, regarding the package name.  Fedora naming conventions are here:
http://fedoraproject.org/wiki/Packaging:NamingGuidelines
Comment 1 correctly notes that there are issues with using an underscore in package names.  On the other hand, comment 2 has a reasonable rationale for the name the way it is.
I re-read the naming guidelines, which specifically say:
"When naming a package, the name should match the upstream tarball or project name from which this software came....
it is possible that the upstream name does not fall into the Common Character Set . If this is the case, refer to: When Upstream Naming is outside of the specified character set...
The maintainer must NOT use an underscore '_', a plus '+', or a period '.' as a delimiter... packages where the upstream name naturally contains an underscore are excluded from this."

Some may disagree with me, but I think this use of "_" is acceptable, because it is part of the upstream name, just like it is in "tcp_wrappers".

It's weird that it doesn't touch %{SOURCE3} during %prep or %build. it only shows up in %install.   I would expect that to be copied into the build areas during %prep, then copy it out during %install.  By the time you hit %install I expect that only the files in the build would be read from, and that writes would generally only write to (install to) the buildroot.  That could be a real issue if rpmbuild were modified in the future to enforce that... could you change it?

There are no official Ada-specific guidelines for Fedora, so I obviously can't refer to them.  I did peek at the Debian Ada policy, which isn't formally required by Fedora but I thought might be helpful:
http://people.debian.org/~lbrenta/debian-ada-policy.html
However, the technical guidelines from that document focuses on creating libraries, which isn't relevant here.

Although there's no separate licensing file, it's REALLY obvious that it's GPL'ed once you look at the code.  *EVERY* .ads (spec) and .adb (body) created by a human (ignoring b~mine_detector* which is auto-generated) has this boilerplate:
-- This is free software; you can redistribute it and/or modify it under
-- terms of the GNU General Public License as published by the Free Software
-- Foundation; version 2.
The project file "mine_detector.gpr" says:
-- This is free software; you can redistribute it and/or modify it under
-- terms of the GNU General Public License as published by the Free Software
-- Foundation; either version 2, or (at your option) any later version.
However, "mine_detector.gpr" is not from the distributed tarball (and doesn't have code anyway), so I think it'd be safer to simply label the license as "GPLv2", since that's what the code files say.
So I guess I'll retract my comment in comment 8.  In this case, the licensing is NOT unclear, it's just that the license file isn't there as a separate file.
In this case, Fedora applicable policy is that "If the source package does not include the text of the license(s), the packager should contact upstream and encourage them to correct this mistake."

Have you encouraged them to correct this mistake, and re-release a version with an included license file?  You don't have to wait til they do so, just simply make the request.

I'm glad to see that there's a .desktop file.  Great!

Comment 10 David A. Wheeler 2010-01-25 23:25:07 UTC
Using http://fedoraproject.org/wiki/Packaging:ReviewGuidelines

    *  MUST: rpmlint must be run on every package. The output should be posted in the review.[1]

$ cd ~/rpmbuild/
$ rpmlint SPECS/mine_detector.spec SRPMS/mine_detector-6.0-2.fc12.src.rpm RPMS/x86_64/mine_detector-*
3 packages and 1 specfiles checked; 0 errors, 0 warnings.

    * MUST: The package must be named according to the Package Naming Guidelines .

OK.  There's controversy about this because the package name includes "_", but the upstream name includes "_", and I interpret the guidelines as saying that's okay in this case.  See comment 1, comment 2, and comment 9.

    * MUST: The spec file name must match the base package %{name}, in the format %{name}.spec unless your package has an exemption. [2] .
OK

    * MUST: The package must meet the Packaging Guidelines

OK.

This uses %{GNAT_optflags} for the option flags.  This macro is defined in /etc/rpm/macros.gnat, and is derived from %{optflags}, so that's fine.


    * MUST: The package must be licensed with a Fedora approved license and meet the Licensing Guidelines

ISSUE: Need to confirm that you've asked upstream to include a separate file with the license in a future edition.

The license of the code (GPLv2) is fine.

    * MUST: The License field in the package spec file must match the actual license. [3]

OK.  The "License:" says GPLv2, and the GPLv2 is specifically included in every source file as a comment.  See comment 9.

    * MUST: If (and only if) the source package includes the text of the license(s) in its own file, then that file, containing the text of the license(s) for the package must be included in %doc.[4]

ISSUE: The package did NOT include a separate license file, and you've ADDED one.  I don't think that meets this requirement.  If there isn't one, then don't add one.  Instead, just ask upstream to (in the future) add one.

    * MUST: The spec file must be written in American English. [5]
OK

    * MUST: The spec file for the package MUST be legible. [6]
OK

    * MUST: The sources used to build the package must match the upstream source, as provided in the spec URL. Reviewers should use md5sum for this task. If no upstream URL can be specified for this package, please see the Source URL Guidelines for how to deal with this.

WARNING: The URL given here has an UNCHANGING version value:
 http://pragmada.x10hosting.com/mine_detector_6.0_src.zip
I recommend doing this instead:
 http://pragmada.x10hosting.com/mine_detector_%{version}_src.zip
This way, updating the RPM spec is far easier.  Otherwise, it'll be easy to change the "version:" value, but not change the URL: entry and get the wrong source file.

The files do match:
be4f2655162cf35208492117456cd479  mine_detector_6.0_src.zip
be4f2655162cf35208492117456cd479  /home/rpmbuilder/rpmbuild/SOURCES/mine_detector_6.0_src.zip

    * MUST: The package MUST successfully compile and build into binary rpms on at least one primary architecture. [7]

OK. x86_64 tested.

    * MUST: If the package does not successfully compile, build or work on an architecture, then those architectures should be listed in the spec in ExcludeArch. Each architecture listed in ExcludeArch MUST have a bug filed in bugzilla, describing the reason that the package does not compile/build/work on that architecture. The bug number MUST be placed in a comment, next to the corresponding ExcludeArch line. [8]

NA.  Don't know of any bad arch.  Koji will detect.


    * MUST: All build dependencies must be listed in BuildRequires, except for any that are listed in the exceptions section of the Packaging Guidelines ; inclusion of those as BuildRequires is optional. Apply common sense.

Used mock to check this:
  mock --rebuild SRPMS/mine_detector-6.0-2.fc12.src.rpm 

    * MUST: The spec file MUST handle locales properly. This is done by using the %find_lang macro. Using %{_datadir}/locale/* is strictly forbidden.[9]

NA.  /locale/ is not used.

    * MUST: Every binary RPM package (or subpackage) which stores shared library files (not just symlinks) in any of the dynamic linker's default paths, must call ldconfig in %post and %postun. [10]

NA

    * MUST: Packages must NOT bundle copies of system libraries.[11]
NA

    * MUST: If the package is designed to be relocatable, the packager must state this fact in the request for review, along with the rationalization for relocation of that specific package. Without this, use of Prefix: /usr is considered a blocker. [12]
NA

    * MUST: A package must own all directories that it creates. If it does not create a directory that it uses, then it should require a package which does create that directory. [13]
OK

    * MUST: A Fedora package must not list a file more than once in the spec file's %files listings. [14]
OK

    * MUST: Permissions on files must be set properly. Executables should be set with executable permissions, for example. Every %files section must include a %defattr(...) line. [15]
OK

    * MUST: Each package must have a %clean section, which contains rm -rf %{buildroot} (or $RPM_BUILD_ROOT). [16]
OK

    * MUST: Each package must consistently use macros. [17]
OK. %{buildroot} style.

    * MUST: The package must contain code, or permissable content. [18]
OK.

    * MUST: Large documentation files must go in a -doc subpackage. (The definition of large is left up to the packager's best judgement, but is not restricted to size. Large can refer to either size or quantity). [19]
NA

    * MUST: If a package includes something as %doc, it must not affect the runtime of the application. To summarize: If it is in %doc, the program must run properly if it is not present. [19]
OK

    * MUST: Header files must be in a -devel package. [20]
NA

    * MUST: Static libraries must be in a -static package. [21]
NA

    * MUST: Packages containing pkgconfig(.pc) files must 'Requires: pkgconfig' (for directory ownership and usability). [22]
    * MUST: If a package contains library files with a suffix (e.g. libfoo.so.1.1), then library files that end in .so (without suffix) must go in a -devel package. [20]
NA

    * MUST: In the vast majority of cases, devel packages must require the base package using a fully versioned dependency: Requires: %{name} = %{version}-%{release} [23]
NA

    * MUST: Packages must NOT contain any .la libtool archives, these must be removed in the spec if they are built.[21]
NA

    * MUST: Packages containing GUI applications must include a %{name}.desktop file, and that file must be properly installed with desktop-file-install in the %install section. If you feel that your packaged GUI application does not need a .desktop file, you must put a comment in the spec file with your explanation. [24]
OK

    * MUST: Packages must not own files or directories already owned by other packages. The rule of thumb here is that the first package to be installed should own the files or directories that other packages may rely upon. This means, for example, that no package in Fedora should ever share ownership with any of the files or directories owned by the filesystem or man package. If you feel that you have a good reason to own a file or directory that another package owns, then please present that at package review time. [25]
OK

    * MUST: At the beginning of %install, each package MUST run rm -rf %{buildroot} (or $RPM_BUILD_ROOT). [26]
OK

    * MUST: All filenames in rpm packages must be valid UTF-8. [27]
OK


SHOULD Items:
Items marked as SHOULD are things that the package (or reviewer) SHOULD do, but is not required to do.

    * SHOULD: If the source package does not include license text(s) as a separate file from upstream, the packager SHOULD query upstream to include it. [28]

ISSUE: Email sent, not clear that it met this SHOULD.

    * SHOULD: The description and summary sections in the package spec file should contain translations for supported Non-English languages, if available. [29]

OK.  Done in one case!

    * SHOULD: The reviewer should test that the package builds in mock. [30]
OK.  It does.

    * SHOULD: The package should compile and build into binary rpms on all supported architectures. [31]
Not done.  Koji will do that.

    * SHOULD: The reviewer should test that the package functions as described. A package should not segfault instead of running, for example.

OK.  I did run it briefly, it displayed and responded reasonably to a few clicks.

    * SHOULD: If scriptlets are used, those scriptlets must be sane. This is vague, and left up to the reviewers judgement to determine sanity. [32]
NA

    * SHOULD: Usually, subpackages other than devel should require the base package using a fully versioned dependency. [23]
NA

    * SHOULD: The placement of pkgconfig(.pc) files depends on their usecase, and this is usually for development purposes, so should be placed in a -devel pkg. A reasonable exception is that the main pkg itself is a devel tool not installed in a user runtime, e.g. gcc or gdb. [22]
NA

    * SHOULD: If the package has file dependencies outside of /etc, /bin, /sbin, /usr/bin, or /usr/sbin consider requiring the package which provides the file instead of the file itself. [33]
OK

    * SHOULD: your package should contain man pages for binaries/scripts. If it doesn't, work with upstream to add them where they make sense.[34]

ISSUE. It does NOT include a man page.  Have you queried upstream?

Comment 11 Björn Persson 2010-01-26 08:41:39 UTC
(In reply to comment #9)
> It's weird that it doesn't touch %{SOURCE3} during %prep or %build. it only
> shows up in %install.   I would expect that to be copied into the build areas
> during %prep, then copy it out during %install.  By the time you hit %install I
> expect that only the files in the build would be read from, and that writes
> would generally only write to (install to) the buildroot.  That could be a real
> issue if rpmbuild were modified in the future to enforce that... could you
> change it?

It seemed a bit silly to copy the file in two steps when one step worked just as well, but if it's more future-proof, then I'll do it that way.


(In reply to comment #10)
> ISSUE: Need to confirm that you've asked upstream to include a separate file
> with the license in a future edition.
[...]
> ISSUE: The package did NOT include a separate license file, and you've ADDED
> one.  I don't think that meets this requirement.  If there isn't one, then
> don't add one.  Instead, just ask upstream to (in the future) add one.

There was supposed to be a license file. It was included with version 5 but was left out from version 6 by mistake. Jeffrey Carter said it was a mistake and gave me permission to use the file from version 5. I'm sure he'll include it in the next version. (Unless he forgets it again of course.) I figure it's similar to backporting a bugfix: We carry the fix as a patch until the next version is released with the fix included. The license file isn't in diff format as it's a whole file, but it corrects a mistake in the upstream release like a patch does.

Here's the relevant email conversation:

------------------------

From: PragmAda Software Engineering <pragmada.com>
To: Bjorn
Subject: PragmAda Announcements
   
FYI, here's a copy of a msg I posted on Team-Ada and c.l.a mentioning you:

1. The PragmAda Software Engineering web site has moved. You may now
find it at

http://pragmada.x10hosting.com/

2. The e-mail address for PragmAda has also changed. It is now

pragmada.com

3. There is a new, source-only version of Mine Detector. V6.0 is
functionally identical to V5.0, but builds with GtkAda 2.8, 2.10, and
2.14 (tested on Linux only). Thanks to Pascal Malaise and Björn Persson
for suggestions about the changes needed. The actual changes in V6.0
were made by Pascal Malaise.

Mine Detector is at

http://pragmada.x10hosting.com/mindet.html

-- 
Jeffrey R. Carter, President
PragmAda Software Engineering

------------------------

From: Björn Persson <bjorn>
To: PragmAda Software Engineering <pragmada.com>
Subject: Re: PragmAda Announcements
   
söndagen den 2 augusti 2009 23:07:23 skrev du:
> FYI, here's a copy of a msg I posted on Team-Ada and c.l.a mentioning you:
>
> 1. The PragmAda Software Engineering web site has moved. You may now
> find it at
>
> http://pragmada.x10hosting.com/
>
> 2. The e-mail address for PragmAda has also changed. It is now
>
> pragmada.com
>
> 3. There is a new, source-only version of Mine Detector. V6.0 is
> functionally identical to V5.0, but builds with GtkAda 2.8, 2.10, and
> 2.14 (tested on Linux only). Thanks to Pascal Malaise and Björn Persson
> for suggestions about the changes needed. The actual changes in V6.0
> were made by Pascal Malaise.
>
> Mine Detector is at
>
> http://pragmada.x10hosting.com/mindet.html

Thanks for notifying me. I will update my Fedora packages (which are currently 
waiting in the review queue).

Have you removed license.txt on purpose? The Fedora project prefers to have a 
license file in every package, and wants me to ask you to include one. It's OK 
to say no, but since you had a license file I'm curious to know why you removed 
it.

Björn Persson

------------------------

From: PragmAda Software Engineering <pragmada.com>
To: Björn Persson <bjorn>
Subject: Re: PragmAda Announcements
   
Björn Persson wrote:
> 
> Have you removed license.txt on purpose? The Fedora project prefers to have a 
> license file in every package, and wants me to ask you to include one. It's OK 
> to say no, but since you had a license file I'm curious to know why you removed 
> it.

No, I just forgot to include it. You may reuse the one in V5.0.

-- 
Jeffrey R. Carter, President
PragmAda Software Engineering

------------------------

So, now that you have all the information, do you still think I should package Mine Detector without the license file?


> WARNING: The URL given here has an UNCHANGING version value:
>  http://pragmada.x10hosting.com/mine_detector_6.0_src.zip
> I recommend doing this instead:
>  http://pragmada.x10hosting.com/mine_detector_%{version}_src.zip
> This way, updating the RPM spec is far easier.  Otherwise, it'll be easy to
> change the "version:" value, but not change the URL: entry and get the wrong
> source file.

Yeah well, Jeffrey's site has jumped around a few times. As you can see above he moved it to a new domain at the same time as he released Mine Detector 6. If he moves it again I'll have to change the URL anyway, and it's much easier to copy the URL and download the zipfile if you don't have to do macro substitution by hand. This was a choice I made but it's no big deal. I can use the macro if you want me to.


> ISSUE. It does NOT include a man page.  Have you queried upstream?    

No. There was no written policy on man pages when I last touched this package. I suppose I could learn the basics of writing man pages and then make one, but I don't see that it could become more than a stub. Mine Detector has no configuration and takes no command line parameters. All the instructions you need are displayed when you press the "Rules" button, and I don't see a point in duplicating that text in a man page.

Comment 12 David A. Wheeler 2010-01-26 15:29:56 UTC
(In reply to comment #11)

>> It's weird that it doesn't touch %{SOURCE3} during %prep or %build. it only shows up in %install...
>It seemed a bit silly to copy the file in two steps when one step worked just
>as well, but if it's more future-proof, then I'll do it that way.

I don't see any guideline requiring it, it just seemed odd.
I'll leave that to your discretion, it's not a requirement.


(In reply to comment #10)
> There was supposed to be a license file. It was included with version 5 but was left out from version 6 by mistake. Jeffrey Carter said it was a mistake and
gave me permission to use the file from version 5... Here's the relevant email conversation...

Okay, thanks for the clarification.  Licensing is a big deal in Fedora, we want to make sure that everything is 100% okay legally.

Please *add* the email conversation to the package (as a single file).  Include the headers for each email.  You already have the emails, so that should be easy.  I think this is important to do.

You could make an argument that this isn't strictly necessary.  In particular, the source files already reference the license.  But unlike *normal* patches, adding a patch that inserts a software license that wasn't there before is certain to raise concerns.  I think it's important to make it absolutely clear to any recipient that this is correct.  And at that point, it doesn't matter how you interpret the guidelines... you've covered all the bases.


Also: Usually files given as Source: have the package name prefixed.
That way, simultaneous builds won't cause trouble.
I think you should do that; be sure to rename them back to their
expected name.

>> WARNING: The URL given here has an UNCHANGING version value:

> Yeah well, Jeffrey's site has jumped around a few times... I can use
the macro if you want me to.

Ah.  Okay, that makes sense.  The guidelines say "Having macros in a Source: or Patch: line is a matter of style."


>> ISSUE. It does NOT include a man page.  Have you queried upstream?    

> No. There was no written policy on man pages when I last touched this package.... All the instructions you
need are displayed when you press the "Rules" button, and I don't see a point
in duplicating that text in a man page.    

It's not a requirement, merely a "Should".  And I agree, for a simple GUI game where the rules are already presented, there's no need for a man page.  So don't worry about that.


So, the most critical thing at this point is to add a file with the email conversation on licensing, as you've done in this bug report, so that everyone can see there is absolutely NO problem.  Also, I think you should modify your files that go into sources so that they're prefixed by the name (so multiple simultaneous builds have no problem).

Comment 13 David A. Wheeler 2010-01-26 22:25:22 UTC
BTW, a koji build shows that this package builds on all Fedora architectures, as expected.  Here are the koji results, just so that they're documented as part of this review request.

$ koji build --scratch dist-f12 ./mine_detector-6.0-2.fc12.src.rpm 
Uploading srpm: ./mine_detector-6.0-2.fc12.src.rpm
[====================================] 100% 00:00:02  23.25 KiB   8.04 KiB/sec
Created task: 1946859
Task info: http://koji.fedoraproject.org/koji/taskinfo?taskID=1946859
Watching tasks (this may be safely interrupted)...
1946859 build (dist-f12, mine_detector-6.0-2.fc12.src.rpm): open (x86-04.phx2.fedoraproject.org)
  1946860 buildArch (mine_detector-6.0-2.fc12.src.rpm, ppc): open (ppc08.phx2.fedoraproject.org)
  1946862 buildArch (mine_detector-6.0-2.fc12.src.rpm, ppc64): free
  1946861 buildArch (mine_detector-6.0-2.fc12.src.rpm, x86_64): open (x86-03.phx2.fedoraproject.org)
  1946863 buildArch (mine_detector-6.0-2.fc12.src.rpm, i686): free
  1946862 buildArch (mine_detector-6.0-2.fc12.src.rpm, ppc64): free -> open (ppc05.phx2.fedoraproject.org)
  1946863 buildArch (mine_detector-6.0-2.fc12.src.rpm, i686): free -> open (xb-01.phx2.fedoraproject.org)
  1946861 buildArch (mine_detector-6.0-2.fc12.src.rpm, x86_64): open (x86-03.phx2.fedoraproject.org) -> closed
  0 free  4 open  1 done  0 failed
  1946863 buildArch (mine_detector-6.0-2.fc12.src.rpm, i686): open (xb-01.phx2.fedoraproject.org) -> closed
  0 free  3 open  2 done  0 failed
  1946860 buildArch (mine_detector-6.0-2.fc12.src.rpm, ppc): open (ppc08.phx2.fedoraproject.org) -> closed
  0 free  2 open  3 done  0 failed
  1946862 buildArch (mine_detector-6.0-2.fc12.src.rpm, ppc64): open (ppc05.phx2.fedoraproject.org) -> closed
  0 free  1 open  4 done  0 failed
1946859 build (dist-f12, mine_detector-6.0-2.fc12.src.rpm): open (x86-04.phx2.fedoraproject.org) -> closed
  0 free  0 open  5 done  0 failed

Comment 14 Björn Persson 2010-01-27 00:04:46 UTC
(In reply to comment #12)
> >> It's weird that it doesn't touch %{SOURCE3} during %prep or %build. it only shows up in %install...
> >It seemed a bit silly to copy the file in two steps when one step worked just
> >as well, but if it's more future-proof, then I'll do it that way.
> 
> I don't see any guideline requiring it, it just seemed odd.
> I'll leave that to your discretion, it's not a requirement.

So one way seems silly to me and the other seems odd to you. Neither is a good basis for decisions. The question that matters is whether it works by design or by accident. I'll leave it as it is for now. If I find any documentation that says it shouldn't be done my way, then I will change it.

> Please *add* the email conversation to the package (as a single file).

Done. I also added the word "file" to the license statement in mine_detector.gpr, to clarify that that license statement covers that file only, not Mine Detector itself. I hope I've made everything perfectly clear now.

> Also: Usually files given as Source: have the package name prefixed.
> That way, simultaneous builds won't cause trouble.
> I think you should do that; be sure to rename them back to their
> expected name.

OK, done.

New version:
http://www.rombobjörn.se/packages/mine_detector.spec
http://www.rombobjörn.se/packages/mine_detector-6.0-3.fc11.src.rpm

Comment 15 David A. Wheeler 2010-01-27 17:23:24 UTC
>> Also: Usually files given as Source: have the package name prefixed.
>> That way, simultaneous builds won't cause trouble.
>> I think you should do that; be sure to rename them back to their
>> expected name.

>OK, done.

Not quite; "license.txt" doesn't have a prefix.  If someone simultaneously built this and another package that ALSO had a "license.txt", using a shared directory, you could have a race condition where the license.txt of that OTHER package would be incorporated into the final build for THIS package.  Eeek.

Best to change this to something like mine_detector_license.txt,
then in %prep do:
cp -p %{SOURCE4} license.txt

Comment 16 David A. Wheeler 2010-01-27 17:26:52 UTC
Whups, never mind, I was looking at the wrong .spec file.  Stupid cache.  Just a second, let me review what you *really* did :-).

Comment 17 David A. Wheeler 2010-01-27 17:49:09 UTC
Thanks!

Sadly, the new SRPM creates a new rpmlint error:
 rpmlint mine_detector.spec ../RPMS/x86_64/mine_detector-6.0-3.fc12.x86_64.rpm ../SRPMS/mine_detector-6.0-3.fc12.src.rpm 
mine_detector.x86_64: W: file-not-utf8 /usr/share/doc/mine_detector-6.0/license_clarification.mbox

The guidelines require fixing rpmlint errors unless rpmlint is wrong:
  http://fedoraproject.org/wiki/Packaging/Guidelines#rpmlint

I suggest changing .mbox to .txt, and invoking "iconv" during %prep or %build to convert the file to UTF-8.  That way, people can get the original file in its original encoding if they want to (it'll be in the SRPM), but the installed file will be in UTF-8.  I suspect few people will use mail readers to read this license clarification material.  Instead, they'll probably pull up their favorite pager or text editor, which won't really pay attention the character encoding markup in this file.  Switching to UTF-8 and using a .txt extension will it easier on those users.

Comment 18 Toshio Ernie Kuratomi 2010-01-27 19:19:51 UTC
Note:  This is fine:
  desktop-file-install --dir=%{buildroot}%{_datadir}/applications %{SOURCE3}

Sometimes people will do something like this: 
  mkdir -p %{buildroot}/%{docdir}/%{name}-%{version}
  cp %{SOURCE3} %{buildroot}/%{docdir}/%{name}-%{version}

Than can run into problems if there's also %doc directives in %files:
  %doc README COPYING

as rpm will remove the docdir files created in %install.  That's one reason you might see some packages installing files unneeded for the build into the builddir in %prep.

Comment 19 David A. Wheeler 2010-01-27 20:01:17 UTC
Regarding comment 18:
Thanks for the note.   I've hit the %doc problem myself when creating packages.  I've modified the wiki to note this:
 https://fedoraproject.org/wiki/How_to_create_an_RPM_package
please feel free to improve that page!

Comment 20 Björn Persson 2010-01-30 21:32:56 UTC
(In reply to comment #18)
Thanks for the information Toshio.

(In reply to comment #17)
> Sadly, the new SRPM creates a new rpmlint error:
>  rpmlint mine_detector.spec ../RPMS/x86_64/mine_detector-6.0-3.fc12.x86_64.rpm
> ../SRPMS/mine_detector-6.0-3.fc12.src.rpm 
> mine_detector.x86_64: W: file-not-utf8
> /usr/share/doc/mine_detector-6.0/license_clarification.mbox

Strictly speaking it's a warning, not an error.

> I suggest changing .mbox to .txt, and invoking "iconv" during %prep or %build
> to convert the file to UTF-8.

I am STRONGLY against shipping the emails with incorrect MIME headers. If someone does load them in an MUA they shall be displayed correctly!

I used the mbox format because I got the impression that it's important to keep the emails authentic. If UTF-8 is more important than authenticity, then there are three ways I could make RPMlint happy:

1: I could apply Quoted-Printable to the bodies of the first and the third email and change the Content-Transfer-Encoding headers accordingly. Quoted-Printable would then be used consistently throughout the file, and RPMlint would be satisfied as the whole file would be pure ASCII. When viewed with a pager or text editor the file would look equally bad to all users regardless of their locale, but all three emails would be displayed correctly in a good MUA.

2: I could convert the bodies of the first and the third email to UTF-8 and change the Content-Type headers accordingly, so that at least two of the occurrences of my name would be fully readable to those users who use UTF-8 in their locales. (Note that which encoding a pager or text editor will assume depends on the user's locale setting. Not everyone is using UTF-8 in their locale.) I'd have to leave the middle email Quoted-Printable-encoded, because any changes would invalidate the PGP signature. The address headers would also have to remain Quoted-Printable-encoded. All three emails would still be displayed correctly in a good MUA.

3: I could abandon the mbox format and merge the emails and the explanatory text that I put in README.Fedora into a single text file. The file would not contain authentic emails but rather plaintext representations of the emails. I would then remove all transfer encoding from both bodies and address headers and convert everything to UTF-8. I would also remove the MIME headers and the PGP signature as they would no longer apply. This would make all three emails as easy to read as possible to users viewing them with a pager or text editor in a UTF-8 locale, but it would no longer be possible to view them with an MUA.

Another approach could be to declare that this is one of those cases where RPMlint is wrong.

Comment 21 David A. Wheeler 2010-01-31 19:27:35 UTC
Okay, you've convinced me!  Go ahead and leave in the .mbox file.  As you noted, the rpmlint report is a warning not an error.

Looking back at the guidelines, The guidelines require that spec files and filenames be encoded in UTF-8, but the guildelines don't require that arbitrary files be in UTF-8.  I think it's important to transition to UTF-8 if there's no easy way for readers to determine the correct encoding (e.g., .txt files), so I think it makes sense for rpmlint to do this check.  But clearly, in .mbox format you DO get that information, and I think you've got a good argument for keeping it in the original format (trying to keep the authentic info about licensing).

Thanks for adding the mbox stuff.  I'm really sensitive to the licensing stuff, and I think this makes it abundantly clear that there is no problem.

Comment 22 David A. Wheeler 2010-01-31 20:02:07 UTC
You've addressed all the issues noted above.  In particular, I think the addition of the mbox file that documents the licensing is an improvement.  Thanks.

APPROVED.

Comment 23 Björn Persson 2010-01-31 21:27:15 UTC
Thanks for the review David! Your reviews are indeed very thorough.

New Package CVS Request
=======================
Package Name: mine_detector
Short Description: Mine Detector, a mine-finding game
Owners: rombobeorn
Branches: F-11 F-12
InitialCC:

Comment 24 Kevin Fenzi 2010-02-01 20:46:30 UTC
CVS done (by process-cvs-requests.py).

Comment 25 Fedora Update System 2010-02-02 20:39:33 UTC
mine_detector-6.0-3.fc11 has been pushed to the Fedora 11 testing repository.  If problems still persist, please make note of it in this bug report.
 If you want to test the update, you can install it with 
 su -c 'yum --enablerepo=updates-testing update mine_detector'.  You can provide feedback for this update here: http://admin.fedoraproject.org/updates/F11/FEDORA-2010-1372

Comment 26 Fedora Update System 2010-02-02 20:44:02 UTC
mine_detector-6.0-3.fc12 has been pushed to the Fedora 12 testing repository.  If problems still persist, please make note of it in this bug report.
 If you want to test the update, you can install it with 
 su -c 'yum --enablerepo=updates-testing update mine_detector'.  You can provide feedback for this update here: http://admin.fedoraproject.org/updates/F12/FEDORA-2010-1403

Comment 27 Fedora Update System 2010-02-18 22:33:24 UTC
mine_detector-6.0-3.fc11 has been pushed to the Fedora 11 stable repository.  If problems still persist, please make note of it in this bug report.

Comment 28 Fedora Update System 2010-02-18 22:35:47 UTC
mine_detector-6.0-3.fc12 has been pushed to the Fedora 12 stable repository.  If problems still persist, please make note of it in this bug report.

Comment 29 Björn Persson 2011-07-03 23:24:36 UTC
Package Change Request
======================
Package Name: mine_detector
New Branches: el6
Owners: rombobeorn

Comment 30 Gwyn Ciesla 2011-07-03 23:33:52 UTC
Git done (by process-git-requests).


Note You need to log in before you can comment on or make changes to this bug.