librelist archives

« back to archive

Problem installing using easy_install

Problem installing using easy_install

From:
Timothy J. Arrington
Date:
2012-04-15 @ 18:52
Hi,

I just tried to install breathe using easy_install on cygwin, but it
fails with:

Searching for breathe
Reading http://pypi.python.org/simple/breathe/
Reading https://github.com/michaeljones/breathe
Best match: breathe 0.7.5.linux-x86-64
Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
Processing breathe-0.7.5.linux-x86_64.tar.gz
error: Couldn't find a setup script in
/tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz

Is this a known problem? How can I workaround this issue?

Cheers,
 -t

Re: [breathe] Problem installing using easy_install

From:
Michael Jones
Date:
2012-04-15 @ 19:28
Hi,

Thanks for letting me know.  I hadn't realised it had gone wrong.
There seemed to be something wrong with the "dumb binary" entry.  I've
deleted that and tried to install and I think it picks up the "source"
entry instead and works.

Please give it another go and see if you have any luck, otherwise I'll
investigate further.

Thanks for trying breathe, or for trying to try it :)
Michael


On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
<sse3ml@googlemail.com> wrote:
> Hi,
>
> I just tried to install breathe using easy_install on cygwin, but it
> fails with:
>
> Searching for breathe
> Reading http://pypi.python.org/simple/breathe/
> Reading https://github.com/michaeljones/breathe
> Best match: breathe 0.7.5.linux-x86-64
> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
> Processing breathe-0.7.5.linux-x86_64.tar.gz
> error: Couldn't find a setup script in
> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>
> Is this a known problem? How can I workaround this issue?
>
> Cheers,
>  -t

Re: [breathe] Problem installing using easy_install

From:
Timothy J. Arrington
Date:
2012-04-15 @ 19:42
Works now, thanks a lot.

I'll try to get it working during the next week. I'm highly interested
in combining a lot of doxygen documentation into the Sphinx-based
documentation; my only concern right now is formatting and usability.
I have spent quite some time tweaking the Doxygen-Documentation to
look good, and I definitely like the fact that stuff can be found
easily with the built-in doxygen search. I have also lots of content
stored in doxygen pages, which I want to keep if possible.

I'll let you know how I get on.

Cheers,
  -t

Am 15. April 2012 21:28 schrieb Michael Jones <m.pricejones@gmail.com>:
> Hi,
>
> Thanks for letting me know.  I hadn't realised it had gone wrong.
> There seemed to be something wrong with the "dumb binary" entry.  I've
> deleted that and tried to install and I think it picks up the "source"
> entry instead and works.
>
> Please give it another go and see if you have any luck, otherwise I'll
> investigate further.
>
> Thanks for trying breathe, or for trying to try it :)
> Michael
>
>
> On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
> <sse3ml@googlemail.com> wrote:
>> Hi,
>>
>> I just tried to install breathe using easy_install on cygwin, but it
>> fails with:
>>
>> Searching for breathe
>> Reading http://pypi.python.org/simple/breathe/
>> Reading https://github.com/michaeljones/breathe
>> Best match: breathe 0.7.5.linux-x86-64
>> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
>> Processing breathe-0.7.5.linux-x86_64.tar.gz
>> error: Couldn't find a setup script in
>> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>>
>> Is this a known problem? How can I workaround this issue?
>>
>> Cheers,
>>  -t

Re: [breathe] Problem installing using easy_install

From:
Timothy J. Arrington
Date:
2012-04-16 @ 18:03
Ok, so I managed to include breathe into my build process. I've
generated XML documentation using doxygen 1.8.0. What's happening now
is that the build fails. First of all, the documentation
(http://michaeljones.github.com/breathe/quickstart.html) is missing
spaces between the '..' and the command name. In order to work,
commands must be written like this:

.. doxygenindex::

Anyway, the build fails with some parse error (see the attachment.)
Any idea how I can debug this further? In my case, there's quite a
bunch of files with more than 1000 lines, and an initial scan didn't
bring up anything suspicious.

Cheers,
  -t

Am 15. April 2012 21:42 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
> Works now, thanks a lot.
>
> I'll try to get it working during the next week. I'm highly interested
> in combining a lot of doxygen documentation into the Sphinx-based
> documentation; my only concern right now is formatting and usability.
> I have spent quite some time tweaking the Doxygen-Documentation to
> look good, and I definitely like the fact that stuff can be found
> easily with the built-in doxygen search. I have also lots of content
> stored in doxygen pages, which I want to keep if possible.
>
> I'll let you know how I get on.
>
> Cheers,
>  -t
>
> Am 15. April 2012 21:28 schrieb Michael Jones <m.pricejones@gmail.com>:
>> Hi,
>>
>> Thanks for letting me know.  I hadn't realised it had gone wrong.
>> There seemed to be something wrong with the "dumb binary" entry.  I've
>> deleted that and tried to install and I think it picks up the "source"
>> entry instead and works.
>>
>> Please give it another go and see if you have any luck, otherwise I'll
>> investigate further.
>>
>> Thanks for trying breathe, or for trying to try it :)
>> Michael
>>
>>
>> On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
>> <sse3ml@googlemail.com> wrote:
>>> Hi,
>>>
>>> I just tried to install breathe using easy_install on cygwin, but it
>>> fails with:
>>>
>>> Searching for breathe
>>> Reading http://pypi.python.org/simple/breathe/
>>> Reading https://github.com/michaeljones/breathe
>>> Best match: breathe 0.7.5.linux-x86-64
>>> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
>>> Processing breathe-0.7.5.linux-x86_64.tar.gz
>>> error: Couldn't find a setup script in
>>> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>>>
>>> Is this a known problem? How can I workaround this issue?
>>>
>>> Cheers,
>>>  -t

Re: [breathe] Problem installing using easy_install

From:
Timothy J. Arrington
Date:
2012-04-16 @ 18:15
Ah, some debugging revealed it chocked at this line:

<para>This works for arbitary matrices in
O(n³).</para><para><simplesect kind="return"><para>False if a 0-pivot
was encountered. </para></simplesect>

After fixing this, it still breaks on another "special" character, the
degree symbol (°)

<codeline lineno="148"><highlight
class="normal"><sp/><sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight

class="comment">//<sp/>right<sp/>we<sp/>have<sp/>to<sp/>rotate<sp/>by<sp/>30°</highlight><highlight
class="normal"></highlight></codeline>

Is there any way to get Expat to parse this? I definitely don't want
to fix in-source comments to use different characters, and the source
file in question should be correctly UTF-8 encoded. This does not seem
to carry through doxygen though, and I'm not sure where it's going
wrong. It seems that the XML is already incorrectly encoded. Any idea
how I to work around this?

Cheers,
  -t

Am 16. April 2012 20:03 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
> Ok, so I managed to include breathe into my build process. I've
> generated XML documentation using doxygen 1.8.0. What's happening now
> is that the build fails. First of all, the documentation
> (http://michaeljones.github.com/breathe/quickstart.html) is missing
> spaces between the '..' and the command name. In order to work,
> commands must be written like this:
>
> .. doxygenindex::
>
> Anyway, the build fails with some parse error (see the attachment.)
> Any idea how I can debug this further? In my case, there's quite a
> bunch of files with more than 1000 lines, and an initial scan didn't
> bring up anything suspicious.
>
> Cheers,
>  -t
>
> Am 15. April 2012 21:42 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>> Works now, thanks a lot.
>>
>> I'll try to get it working during the next week. I'm highly interested
>> in combining a lot of doxygen documentation into the Sphinx-based
>> documentation; my only concern right now is formatting and usability.
>> I have spent quite some time tweaking the Doxygen-Documentation to
>> look good, and I definitely like the fact that stuff can be found
>> easily with the built-in doxygen search. I have also lots of content
>> stored in doxygen pages, which I want to keep if possible.
>>
>> I'll let you know how I get on.
>>
>> Cheers,
>>  -t
>>
>> Am 15. April 2012 21:28 schrieb Michael Jones <m.pricejones@gmail.com>:
>>> Hi,
>>>
>>> Thanks for letting me know.  I hadn't realised it had gone wrong.
>>> There seemed to be something wrong with the "dumb binary" entry.  I've
>>> deleted that and tried to install and I think it picks up the "source"
>>> entry instead and works.
>>>
>>> Please give it another go and see if you have any luck, otherwise I'll
>>> investigate further.
>>>
>>> Thanks for trying breathe, or for trying to try it :)
>>> Michael
>>>
>>>
>>> On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
>>> <sse3ml@googlemail.com> wrote:
>>>> Hi,
>>>>
>>>> I just tried to install breathe using easy_install on cygwin, but it
>>>> fails with:
>>>>
>>>> Searching for breathe
>>>> Reading http://pypi.python.org/simple/breathe/
>>>> Reading https://github.com/michaeljones/breathe
>>>> Best match: breathe 0.7.5.linux-x86-64
>>>> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
>>>> Processing breathe-0.7.5.linux-x86_64.tar.gz
>>>> error: Couldn't find a setup script in
>>>> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>>>>
>>>> Is this a known problem? How can I workaround this issue?
>>>>
>>>> Cheers,
>>>>  -t

Re: [breathe] Problem installing using easy_install

From:
Michael Jones
Date:
2012-04-18 @ 07:10
Hi,

I'm afraid I'm not sure myself.  I'd agree that with everything in the
correct encoding you would hope that it would work.  I'll try to take
a look and maybe raise it with the doxygen guys if I can (though I
know them no better than anyone else.)

Thanks for debugging it passed the original error and for pointing out
the issue in the quick start guide, I'll fix that.  Also much respect
for using unicode characters in source code comments... I'm still
stuck in the past in seems.

Cheers,
Michael


On Tue, Apr 17, 2012 at 6:15 AM, Timothy J. Arrington
<sse3ml@googlemail.com> wrote:
> Ah, some debugging revealed it chocked at this line:
>
> <para>This works for arbitary matrices in
> O(n³).</para><para><simplesect kind="return"><para>False if a 0-pivot
> was encountered. </para></simplesect>
>
> After fixing this, it still breaks on another "special" character, the
> degree symbol (°)
>
> <codeline lineno="148"><highlight
> class="normal"><sp/><sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight
> 
class="comment">//<sp/>right<sp/>we<sp/>have<sp/>to<sp/>rotate<sp/>by<sp/>30°</highlight><highlight
> class="normal"></highlight></codeline>
>
> Is there any way to get Expat to parse this? I definitely don't want
> to fix in-source comments to use different characters, and the source
> file in question should be correctly UTF-8 encoded. This does not seem
> to carry through doxygen though, and I'm not sure where it's going
> wrong. It seems that the XML is already incorrectly encoded. Any idea
> how I to work around this?
>
> Cheers,
>  -t
>
> Am 16. April 2012 20:03 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>> Ok, so I managed to include breathe into my build process. I've
>> generated XML documentation using doxygen 1.8.0. What's happening now
>> is that the build fails. First of all, the documentation
>> (http://michaeljones.github.com/breathe/quickstart.html) is missing
>> spaces between the '..' and the command name. In order to work,
>> commands must be written like this:
>>
>> .. doxygenindex::
>>
>> Anyway, the build fails with some parse error (see the attachment.)
>> Any idea how I can debug this further? In my case, there's quite a
>> bunch of files with more than 1000 lines, and an initial scan didn't
>> bring up anything suspicious.
>>
>> Cheers,
>>  -t
>>
>> Am 15. April 2012 21:42 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>> Works now, thanks a lot.
>>>
>>> I'll try to get it working during the next week. I'm highly interested
>>> in combining a lot of doxygen documentation into the Sphinx-based
>>> documentation; my only concern right now is formatting and usability.
>>> I have spent quite some time tweaking the Doxygen-Documentation to
>>> look good, and I definitely like the fact that stuff can be found
>>> easily with the built-in doxygen search. I have also lots of content
>>> stored in doxygen pages, which I want to keep if possible.
>>>
>>> I'll let you know how I get on.
>>>
>>> Cheers,
>>>  -t
>>>
>>> Am 15. April 2012 21:28 schrieb Michael Jones <m.pricejones@gmail.com>:
>>>> Hi,
>>>>
>>>> Thanks for letting me know.  I hadn't realised it had gone wrong.
>>>> There seemed to be something wrong with the "dumb binary" entry.  I've
>>>> deleted that and tried to install and I think it picks up the "source"
>>>> entry instead and works.
>>>>
>>>> Please give it another go and see if you have any luck, otherwise I'll
>>>> investigate further.
>>>>
>>>> Thanks for trying breathe, or for trying to try it :)
>>>> Michael
>>>>
>>>>
>>>> On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
>>>> <sse3ml@googlemail.com> wrote:
>>>>> Hi,
>>>>>
>>>>> I just tried to install breathe using easy_install on cygwin, but it
>>>>> fails with:
>>>>>
>>>>> Searching for breathe
>>>>> Reading http://pypi.python.org/simple/breathe/
>>>>> Reading https://github.com/michaeljones/breathe
>>>>> Best match: breathe 0.7.5.linux-x86-64
>>>>> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
>>>>> Processing breathe-0.7.5.linux-x86_64.tar.gz
>>>>> error: Couldn't find a setup script in
>>>>> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>>>>>
>>>>> Is this a known problem? How can I workaround this issue?
>>>>>
>>>>> Cheers,
>>>>>  -t

Re: [breathe] Problem installing using easy_install

From:
Timothy J. Arrington
Date:
2012-04-18 @ 15:58
Hi Michael,

thanks. I'll try to get a bug report to the Doxygen guys as well if I
can make a small repro case. Actually, the character is not even
unicode per se. It's already in the ANSI set (ANSI Character 176 or
0xB0) but this is no longer in the lower 7-bit UTF-8 plane. My guess
is the character gets passed through as-is instead of being encoded
properly as UTF-8. I'll try to figure it out, thanks for trying on
your end as well.

Cheers,
  -t

Am 18. April 2012 09:10 schrieb Michael Jones <m.pricejones@gmail.com>:
> Hi,
>
> I'm afraid I'm not sure myself.  I'd agree that with everything in the
> correct encoding you would hope that it would work.  I'll try to take
> a look and maybe raise it with the doxygen guys if I can (though I
> know them no better than anyone else.)
>
> Thanks for debugging it passed the original error and for pointing out
> the issue in the quick start guide, I'll fix that.  Also much respect
> for using unicode characters in source code comments... I'm still
> stuck in the past in seems.
>
> Cheers,
> Michael
>
>
> On Tue, Apr 17, 2012 at 6:15 AM, Timothy J. Arrington
> <sse3ml@googlemail.com> wrote:
>> Ah, some debugging revealed it chocked at this line:
>>
>> <para>This works for arbitary matrices in
>> O(n³).</para><para><simplesect kind="return"><para>False if a 0-pivot
>> was encountered. </para></simplesect>
>>
>> After fixing this, it still breaks on another "special" character, the
>> degree symbol (°)
>>
>> <codeline lineno="148"><highlight
>> class="normal"><sp/><sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight
>> 
class="comment">//<sp/>right<sp/>we<sp/>have<sp/>to<sp/>rotate<sp/>by<sp/>30°</highlight><highlight
>> class="normal"></highlight></codeline>
>>
>> Is there any way to get Expat to parse this? I definitely don't want
>> to fix in-source comments to use different characters, and the source
>> file in question should be correctly UTF-8 encoded. This does not seem
>> to carry through doxygen though, and I'm not sure where it's going
>> wrong. It seems that the XML is already incorrectly encoded. Any idea
>> how I to work around this?
>>
>> Cheers,
>>  -t
>>
>> Am 16. April 2012 20:03 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>> Ok, so I managed to include breathe into my build process. I've
>>> generated XML documentation using doxygen 1.8.0. What's happening now
>>> is that the build fails. First of all, the documentation
>>> (http://michaeljones.github.com/breathe/quickstart.html) is missing
>>> spaces between the '..' and the command name. In order to work,
>>> commands must be written like this:
>>>
>>> .. doxygenindex::
>>>
>>> Anyway, the build fails with some parse error (see the attachment.)
>>> Any idea how I can debug this further? In my case, there's quite a
>>> bunch of files with more than 1000 lines, and an initial scan didn't
>>> bring up anything suspicious.
>>>
>>> Cheers,
>>>  -t
>>>
>>> Am 15. April 2012 21:42 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>>> Works now, thanks a lot.
>>>>
>>>> I'll try to get it working during the next week. I'm highly interested
>>>> in combining a lot of doxygen documentation into the Sphinx-based
>>>> documentation; my only concern right now is formatting and usability.
>>>> I have spent quite some time tweaking the Doxygen-Documentation to
>>>> look good, and I definitely like the fact that stuff can be found
>>>> easily with the built-in doxygen search. I have also lots of content
>>>> stored in doxygen pages, which I want to keep if possible.
>>>>
>>>> I'll let you know how I get on.
>>>>
>>>> Cheers,
>>>>  -t
>>>>
>>>> Am 15. April 2012 21:28 schrieb Michael Jones <m.pricejones@gmail.com>:
>>>>> Hi,
>>>>>
>>>>> Thanks for letting me know.  I hadn't realised it had gone wrong.
>>>>> There seemed to be something wrong with the "dumb binary" entry.  I've
>>>>> deleted that and tried to install and I think it picks up the "source"
>>>>> entry instead and works.
>>>>>
>>>>> Please give it another go and see if you have any luck, otherwise I'll
>>>>> investigate further.
>>>>>
>>>>> Thanks for trying breathe, or for trying to try it :)
>>>>> Michael
>>>>>
>>>>>
>>>>> On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
>>>>> <sse3ml@googlemail.com> wrote:
>>>>>> Hi,
>>>>>>
>>>>>> I just tried to install breathe using easy_install on cygwin, but it
>>>>>> fails with:
>>>>>>
>>>>>> Searching for breathe
>>>>>> Reading http://pypi.python.org/simple/breathe/
>>>>>> Reading https://github.com/michaeljones/breathe
>>>>>> Best match: breathe 0.7.5.linux-x86-64
>>>>>> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
>>>>>> Processing breathe-0.7.5.linux-x86_64.tar.gz
>>>>>> error: Couldn't find a setup script in
>>>>>> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>
>>>>>> Is this a known problem? How can I workaround this issue?
>>>>>>
>>>>>> Cheers,
>>>>>>  -t

Re: [breathe] Problem installing using easy_install

From:
Timothy J. Arrington
Date:
2012-04-18 @ 16:54
Hi,

with lots of more wrestling, I finally made sure that each and every
file is properly UTF-8 encoded. It seems that doxygen passes through
whatever comes in, and if it is not UTF8, it gets mangled in the
process.

Ok, so I've built now my documentation using breathe. It worked fine
for my several-hundred-classes project, albeit it took some time :)
The documentation looks pretty decent, but -- it's all on one page,
and the layout is not quite how I would like to it be. For instance, I
would like to have a short list of all functions/members for each
class serving as an index to the longer function descriptions. There's
also very little CSS markup I can use for styling. And finally, how
the heck can I put each class into its own file?

My biggest problems right now:
* Each class into one file: How can I change this, or where do I have
to start extending Breathe to fix this?
* Each page into one page: Same as above.
* Where to extend breathe to start tweaking the output?
* The search ain't working, if I search for a class using the
Sphinx-built-in-search, it doesn't find anything

Besides this, getting Breathe to work at all is damn cool, and the
output looks like it could be much better than Doxygen with less
tweaking.

Cheers,
  --t

Am 18. April 2012 17:58 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
> Hi Michael,
>
> thanks. I'll try to get a bug report to the Doxygen guys as well if I
> can make a small repro case. Actually, the character is not even
> unicode per se. It's already in the ANSI set (ANSI Character 176 or
> 0xB0) but this is no longer in the lower 7-bit UTF-8 plane. My guess
> is the character gets passed through as-is instead of being encoded
> properly as UTF-8. I'll try to figure it out, thanks for trying on
> your end as well.
>
> Cheers,
>  -t
>
> Am 18. April 2012 09:10 schrieb Michael Jones <m.pricejones@gmail.com>:
>> Hi,
>>
>> I'm afraid I'm not sure myself.  I'd agree that with everything in the
>> correct encoding you would hope that it would work.  I'll try to take
>> a look and maybe raise it with the doxygen guys if I can (though I
>> know them no better than anyone else.)
>>
>> Thanks for debugging it passed the original error and for pointing out
>> the issue in the quick start guide, I'll fix that.  Also much respect
>> for using unicode characters in source code comments... I'm still
>> stuck in the past in seems.
>>
>> Cheers,
>> Michael
>>
>>
>> On Tue, Apr 17, 2012 at 6:15 AM, Timothy J. Arrington
>> <sse3ml@googlemail.com> wrote:
>>> Ah, some debugging revealed it chocked at this line:
>>>
>>> <para>This works for arbitary matrices in
>>> O(n³).</para><para><simplesect kind="return"><para>False if a 0-pivot
>>> was encountered. </para></simplesect>
>>>
>>> After fixing this, it still breaks on another "special" character, the
>>> degree symbol (°)
>>>
>>> <codeline lineno="148"><highlight
>>> class="normal"><sp/><sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight
>>> 
class="comment">//<sp/>right<sp/>we<sp/>have<sp/>to<sp/>rotate<sp/>by<sp/>30°</highlight><highlight
>>> class="normal"></highlight></codeline>
>>>
>>> Is there any way to get Expat to parse this? I definitely don't want
>>> to fix in-source comments to use different characters, and the source
>>> file in question should be correctly UTF-8 encoded. This does not seem
>>> to carry through doxygen though, and I'm not sure where it's going
>>> wrong. It seems that the XML is already incorrectly encoded. Any idea
>>> how I to work around this?
>>>
>>> Cheers,
>>>  -t
>>>
>>> Am 16. April 2012 20:03 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>>> Ok, so I managed to include breathe into my build process. I've
>>>> generated XML documentation using doxygen 1.8.0. What's happening now
>>>> is that the build fails. First of all, the documentation
>>>> (http://michaeljones.github.com/breathe/quickstart.html) is missing
>>>> spaces between the '..' and the command name. In order to work,
>>>> commands must be written like this:
>>>>
>>>> .. doxygenindex::
>>>>
>>>> Anyway, the build fails with some parse error (see the attachment.)
>>>> Any idea how I can debug this further? In my case, there's quite a
>>>> bunch of files with more than 1000 lines, and an initial scan didn't
>>>> bring up anything suspicious.
>>>>
>>>> Cheers,
>>>>  -t
>>>>
>>>> Am 15. April 2012 21:42 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>>>> Works now, thanks a lot.
>>>>>
>>>>> I'll try to get it working during the next week. I'm highly interested
>>>>> in combining a lot of doxygen documentation into the Sphinx-based
>>>>> documentation; my only concern right now is formatting and usability.
>>>>> I have spent quite some time tweaking the Doxygen-Documentation to
>>>>> look good, and I definitely like the fact that stuff can be found
>>>>> easily with the built-in doxygen search. I have also lots of content
>>>>> stored in doxygen pages, which I want to keep if possible.
>>>>>
>>>>> I'll let you know how I get on.
>>>>>
>>>>> Cheers,
>>>>>  -t
>>>>>
>>>>> Am 15. April 2012 21:28 schrieb Michael Jones <m.pricejones@gmail.com>:
>>>>>> Hi,
>>>>>>
>>>>>> Thanks for letting me know.  I hadn't realised it had gone wrong.
>>>>>> There seemed to be something wrong with the "dumb binary" entry.  I've
>>>>>> deleted that and tried to install and I think it picks up the "source"
>>>>>> entry instead and works.
>>>>>>
>>>>>> Please give it another go and see if you have any luck, otherwise I'll
>>>>>> investigate further.
>>>>>>
>>>>>> Thanks for trying breathe, or for trying to try it :)
>>>>>> Michael
>>>>>>
>>>>>>
>>>>>> On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
>>>>>> <sse3ml@googlemail.com> wrote:
>>>>>>> Hi,
>>>>>>>
>>>>>>> I just tried to install breathe using easy_install on cygwin, but it
>>>>>>> fails with:
>>>>>>>
>>>>>>> Searching for breathe
>>>>>>> Reading http://pypi.python.org/simple/breathe/
>>>>>>> Reading https://github.com/michaeljones/breathe
>>>>>>> Best match: breathe 0.7.5.linux-x86-64
>>>>>>> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
>>>>>>> Processing breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>> error: Couldn't find a setup script in
>>>>>>> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>>
>>>>>>> Is this a known problem? How can I workaround this issue?
>>>>>>>
>>>>>>> Cheers,
>>>>>>>  -t

Re: [breathe] Problem installing using easy_install

From:
Timothy J. Arrington
Date:
2012-04-18 @ 18:15
I just did some more testing, here's the findings:

* The output seems to contain everything that Doxygen produces as
well, so I'm fine with that. However, some classes get documented
which are not included in the Doxygen documentation, in my case,
private Impl classes basically produce noise. There doesn't seem a way
to remove them.
* No inheritance graph: I can live with that. It's rarely useful anyway.
* Enumeration formatting is a bit off: If there is no description, it
still produces value - -- the trailing hyphen is irritating. I could
hack something to resolve this issue relatively quickly by modifying
the renderer in renderer/.../compound.py, in the
EnumvalueTypeSubRenderer class. However, I'm not sure if this class is
generated or open for modification at all. There is definitely a lot
of pre-generated stuff, and I would be more happy to just "inject"
modifications instead of modifying existing classes.
* When using .. doxygenclass:: I have to provide the full class name
including the namespace, but for the documentation, I would love to
strip the common namespace prefix. That should be easy to do if I can
hook the title output, but I couldn't find where exactly this is
happening.
* Function parameters get reformatted somewhere; I would like to
manipulate it so I can change const Id & id to <bold>const</bold> Id&
id for instance. Same for various instances of typedefs, where typedef
is not highlighted.
* The output is packed in <div> instances for Sphinx 1.1.3; I wonder
if this can be augmented a bit so I can figure out which parts have
been generated by breathe, so I can apply a different formatting. For
enums, I have seen that there is a breatheenumvalues, but there is
certainly a lack of CSS selectors for formatting everything.
* The printed documentation has a lot of empty space.
* No easy way to get a list of all classes, functions, etc.
* Groups seem to be ignored? I have multiple pages nested in groups,
but I couldn't find a way how to extract those. The most interesting
feature here is to have a list of all members in each group, and nest
the groups properly, so I can retain a hierarchical index (this works
painlessly in doxygen.)

However, that's it actually, otherwise, Breathe seems like the perfect
solution for projects which use Sphinx for documentation. I like the
output and being indepedent of Doxygen's HTML problems is definitely a
huge win. I especially like the fact that the printed documentation
would work just fine with Breathe+Sphinx+latexpdf.

It would great if you could provide some pointers where to start
extending, if there are any plans to make the output easier to
customize, or if you are already planning to integrate some of the
stuff I mentioned above. I definitely would like to drop Doxygen's
HTML alltogether in favour of Sphinx+Breathe, but I'm not sure how
much time I'll be able to invest into adapting Breathe's output (but
I'm really interested in fixing the issues above, which would bring
the Breathe output more or less in sync with what Qt has.)

Cheers,
  -t

Am 18. April 2012 18:54 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
> Hi,
>
> with lots of more wrestling, I finally made sure that each and every
> file is properly UTF-8 encoded. It seems that doxygen passes through
> whatever comes in, and if it is not UTF8, it gets mangled in the
> process.
>
> Ok, so I've built now my documentation using breathe. It worked fine
> for my several-hundred-classes project, albeit it took some time :)
> The documentation looks pretty decent, but -- it's all on one page,
> and the layout is not quite how I would like to it be. For instance, I
> would like to have a short list of all functions/members for each
> class serving as an index to the longer function descriptions. There's
> also very little CSS markup I can use for styling. And finally, how
> the heck can I put each class into its own file?
>
> My biggest problems right now:
> * Each class into one file: How can I change this, or where do I have
> to start extending Breathe to fix this?
> * Each page into one page: Same as above.
> * Where to extend breathe to start tweaking the output?
> * The search ain't working, if I search for a class using the
> Sphinx-built-in-search, it doesn't find anything
>
> Besides this, getting Breathe to work at all is damn cool, and the
> output looks like it could be much better than Doxygen with less
> tweaking.
>
> Cheers,
>  --t
>
> Am 18. April 2012 17:58 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>> Hi Michael,
>>
>> thanks. I'll try to get a bug report to the Doxygen guys as well if I
>> can make a small repro case. Actually, the character is not even
>> unicode per se. It's already in the ANSI set (ANSI Character 176 or
>> 0xB0) but this is no longer in the lower 7-bit UTF-8 plane. My guess
>> is the character gets passed through as-is instead of being encoded
>> properly as UTF-8. I'll try to figure it out, thanks for trying on
>> your end as well.
>>
>> Cheers,
>>  -t
>>
>> Am 18. April 2012 09:10 schrieb Michael Jones <m.pricejones@gmail.com>:
>>> Hi,
>>>
>>> I'm afraid I'm not sure myself.  I'd agree that with everything in the
>>> correct encoding you would hope that it would work.  I'll try to take
>>> a look and maybe raise it with the doxygen guys if I can (though I
>>> know them no better than anyone else.)
>>>
>>> Thanks for debugging it passed the original error and for pointing out
>>> the issue in the quick start guide, I'll fix that.  Also much respect
>>> for using unicode characters in source code comments... I'm still
>>> stuck in the past in seems.
>>>
>>> Cheers,
>>> Michael
>>>
>>>
>>> On Tue, Apr 17, 2012 at 6:15 AM, Timothy J. Arrington
>>> <sse3ml@googlemail.com> wrote:
>>>> Ah, some debugging revealed it chocked at this line:
>>>>
>>>> <para>This works for arbitary matrices in
>>>> O(n³).</para><para><simplesect kind="return"><para>False if a 0-pivot
>>>> was encountered. </para></simplesect>
>>>>
>>>> After fixing this, it still breaks on another "special" character, the
>>>> degree symbol (°)
>>>>
>>>> <codeline lineno="148"><highlight
>>>> class="normal"><sp/><sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight
>>>> 
class="comment">//<sp/>right<sp/>we<sp/>have<sp/>to<sp/>rotate<sp/>by<sp/>30°</highlight><highlight
>>>> class="normal"></highlight></codeline>
>>>>
>>>> Is there any way to get Expat to parse this? I definitely don't want
>>>> to fix in-source comments to use different characters, and the source
>>>> file in question should be correctly UTF-8 encoded. This does not seem
>>>> to carry through doxygen though, and I'm not sure where it's going
>>>> wrong. It seems that the XML is already incorrectly encoded. Any idea
>>>> how I to work around this?
>>>>
>>>> Cheers,
>>>>  -t
>>>>
>>>> Am 16. April 2012 20:03 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>>>> Ok, so I managed to include breathe into my build process. I've
>>>>> generated XML documentation using doxygen 1.8.0. What's happening now
>>>>> is that the build fails. First of all, the documentation
>>>>> (http://michaeljones.github.com/breathe/quickstart.html) is missing
>>>>> spaces between the '..' and the command name. In order to work,
>>>>> commands must be written like this:
>>>>>
>>>>> .. doxygenindex::
>>>>>
>>>>> Anyway, the build fails with some parse error (see the attachment.)
>>>>> Any idea how I can debug this further? In my case, there's quite a
>>>>> bunch of files with more than 1000 lines, and an initial scan didn't
>>>>> bring up anything suspicious.
>>>>>
>>>>> Cheers,
>>>>>  -t
>>>>>
>>>>> Am 15. April 2012 21:42 schrieb Timothy J. Arrington 
<sse3ml@googlemail.com>:
>>>>>> Works now, thanks a lot.
>>>>>>
>>>>>> I'll try to get it working during the next week. I'm highly interested
>>>>>> in combining a lot of doxygen documentation into the Sphinx-based
>>>>>> documentation; my only concern right now is formatting and usability.
>>>>>> I have spent quite some time tweaking the Doxygen-Documentation to
>>>>>> look good, and I definitely like the fact that stuff can be found
>>>>>> easily with the built-in doxygen search. I have also lots of content
>>>>>> stored in doxygen pages, which I want to keep if possible.
>>>>>>
>>>>>> I'll let you know how I get on.
>>>>>>
>>>>>> Cheers,
>>>>>>  -t
>>>>>>
>>>>>> Am 15. April 2012 21:28 schrieb Michael Jones <m.pricejones@gmail.com>:
>>>>>>> Hi,
>>>>>>>
>>>>>>> Thanks for letting me know.  I hadn't realised it had gone wrong.
>>>>>>> There seemed to be something wrong with the "dumb binary" entry.  I've
>>>>>>> deleted that and tried to install and I think it picks up the "source"
>>>>>>> entry instead and works.
>>>>>>>
>>>>>>> Please give it another go and see if you have any luck, otherwise I'll
>>>>>>> investigate further.
>>>>>>>
>>>>>>> Thanks for trying breathe, or for trying to try it :)
>>>>>>> Michael
>>>>>>>
>>>>>>>
>>>>>>> On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
>>>>>>> <sse3ml@googlemail.com> wrote:
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> I just tried to install breathe using easy_install on cygwin, but it
>>>>>>>> fails with:
>>>>>>>>
>>>>>>>> Searching for breathe
>>>>>>>> Reading http://pypi.python.org/simple/breathe/
>>>>>>>> Reading https://github.com/michaeljones/breathe
>>>>>>>> Best match: breathe 0.7.5.linux-x86-64
>>>>>>>> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
>>>>>>>> Processing breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>>> error: Couldn't find a setup script in
>>>>>>>> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>>>
>>>>>>>> Is this a known problem? How can I workaround this issue?
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>>  -t

Re: [breathe] Problem installing using easy_install

From:
Michael Jones
Date:
2012-05-07 @ 09:31
Hi Tim,

Sorry for the slow reply, somehow I didn't register your additional
emails in this thread.

Thanks for giving it a go and for providing such a comprehensive run
down.  I can't answer all your questions right now but I can look a
little closer if you're still interested.  Breathe is kind of in this
half-way house of doing quite a bit of cool stuff but not quite being
a full solution yet.  Development stumbles along as different people
show interest and request fixes and features but I'm not actively
developing it otherwise.

From what I can see the single page vs multiple pages issue is the
biggest. Breathe follows the Sphinx model of requiring you (as far as
I know) to create stub files for all the pages you're interested in.
I think the output of a restructured text directive is limited to a
single page so you can't easily create multiple pages from say a
single "doxygenindex" statement.  If the situation has changed with
Sphinx then I'd be interested to know how it is done.  Until then I
had plans to create a script that would help make those stub files but
that is about as far as I had thought.

Thanks again, if you're really interested in hacking on it a bit and
extending it you'd be more than welcome.  I'm happy to contribute and
help as much as I can too.

All the best,
Michael


On Thu, Apr 19, 2012 at 6:15 AM, Timothy J. Arrington
<sse3ml@googlemail.com> wrote:
> I just did some more testing, here's the findings:
>
> * The output seems to contain everything that Doxygen produces as
> well, so I'm fine with that. However, some classes get documented
> which are not included in the Doxygen documentation, in my case,
> private Impl classes basically produce noise. There doesn't seem a way
> to remove them.
> * No inheritance graph: I can live with that. It's rarely useful anyway.
> * Enumeration formatting is a bit off: If there is no description, it
> still produces value - -- the trailing hyphen is irritating. I could
> hack something to resolve this issue relatively quickly by modifying
> the renderer in renderer/.../compound.py, in the
> EnumvalueTypeSubRenderer class. However, I'm not sure if this class is
> generated or open for modification at all. There is definitely a lot
> of pre-generated stuff, and I would be more happy to just "inject"
> modifications instead of modifying existing classes.
> * When using .. doxygenclass:: I have to provide the full class name
> including the namespace, but for the documentation, I would love to
> strip the common namespace prefix. That should be easy to do if I can
> hook the title output, but I couldn't find where exactly this is
> happening.
> * Function parameters get reformatted somewhere; I would like to
> manipulate it so I can change const Id & id to <bold>const</bold> Id&
> id for instance. Same for various instances of typedefs, where typedef
> is not highlighted.
> * The output is packed in <div> instances for Sphinx 1.1.3; I wonder
> if this can be augmented a bit so I can figure out which parts have
> been generated by breathe, so I can apply a different formatting. For
> enums, I have seen that there is a breatheenumvalues, but there is
> certainly a lack of CSS selectors for formatting everything.
> * The printed documentation has a lot of empty space.
> * No easy way to get a list of all classes, functions, etc.
> * Groups seem to be ignored? I have multiple pages nested in groups,
> but I couldn't find a way how to extract those. The most interesting
> feature here is to have a list of all members in each group, and nest
> the groups properly, so I can retain a hierarchical index (this works
> painlessly in doxygen.)
>
> However, that's it actually, otherwise, Breathe seems like the perfect
> solution for projects which use Sphinx for documentation. I like the
> output and being indepedent of Doxygen's HTML problems is definitely a
> huge win. I especially like the fact that the printed documentation
> would work just fine with Breathe+Sphinx+latexpdf.
>
> It would great if you could provide some pointers where to start
> extending, if there are any plans to make the output easier to
> customize, or if you are already planning to integrate some of the
> stuff I mentioned above. I definitely would like to drop Doxygen's
> HTML alltogether in favour of Sphinx+Breathe, but I'm not sure how
> much time I'll be able to invest into adapting Breathe's output (but
> I'm really interested in fixing the issues above, which would bring
> the Breathe output more or less in sync with what Qt has.)
>
> Cheers,
>  -t
>
> Am 18. April 2012 18:54 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>> Hi,
>>
>> with lots of more wrestling, I finally made sure that each and every
>> file is properly UTF-8 encoded. It seems that doxygen passes through
>> whatever comes in, and if it is not UTF8, it gets mangled in the
>> process.
>>
>> Ok, so I've built now my documentation using breathe. It worked fine
>> for my several-hundred-classes project, albeit it took some time :)
>> The documentation looks pretty decent, but -- it's all on one page,
>> and the layout is not quite how I would like to it be. For instance, I
>> would like to have a short list of all functions/members for each
>> class serving as an index to the longer function descriptions. There's
>> also very little CSS markup I can use for styling. And finally, how
>> the heck can I put each class into its own file?
>>
>> My biggest problems right now:
>> * Each class into one file: How can I change this, or where do I have
>> to start extending Breathe to fix this?
>> * Each page into one page: Same as above.
>> * Where to extend breathe to start tweaking the output?
>> * The search ain't working, if I search for a class using the
>> Sphinx-built-in-search, it doesn't find anything
>>
>> Besides this, getting Breathe to work at all is damn cool, and the
>> output looks like it could be much better than Doxygen with less
>> tweaking.
>>
>> Cheers,
>>  --t
>>
>> Am 18. April 2012 17:58 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>> Hi Michael,
>>>
>>> thanks. I'll try to get a bug report to the Doxygen guys as well if I
>>> can make a small repro case. Actually, the character is not even
>>> unicode per se. It's already in the ANSI set (ANSI Character 176 or
>>> 0xB0) but this is no longer in the lower 7-bit UTF-8 plane. My guess
>>> is the character gets passed through as-is instead of being encoded
>>> properly as UTF-8. I'll try to figure it out, thanks for trying on
>>> your end as well.
>>>
>>> Cheers,
>>>  -t
>>>
>>> Am 18. April 2012 09:10 schrieb Michael Jones <m.pricejones@gmail.com>:
>>>> Hi,
>>>>
>>>> I'm afraid I'm not sure myself.  I'd agree that with everything in the
>>>> correct encoding you would hope that it would work.  I'll try to take
>>>> a look and maybe raise it with the doxygen guys if I can (though I
>>>> know them no better than anyone else.)
>>>>
>>>> Thanks for debugging it passed the original error and for pointing out
>>>> the issue in the quick start guide, I'll fix that.  Also much respect
>>>> for using unicode characters in source code comments... I'm still
>>>> stuck in the past in seems.
>>>>
>>>> Cheers,
>>>> Michael
>>>>
>>>>
>>>> On Tue, Apr 17, 2012 at 6:15 AM, Timothy J. Arrington
>>>> <sse3ml@googlemail.com> wrote:
>>>>> Ah, some debugging revealed it chocked at this line:
>>>>>
>>>>> <para>This works for arbitary matrices in
>>>>> O(n³).</para><para><simplesect kind="return"><para>False if a 0-pivot
>>>>> was encountered. </para></simplesect>
>>>>>
>>>>> After fixing this, it still breaks on another "special" character, the
>>>>> degree symbol (°)
>>>>>
>>>>> <codeline lineno="148"><highlight
>>>>> 
class="normal"><sp/><sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight
>>>>> 
class="comment">//<sp/>right<sp/>we<sp/>have<sp/>to<sp/>rotate<sp/>by<sp/>30°</highlight><highlight
>>>>> class="normal"></highlight></codeline>
>>>>>
>>>>> Is there any way to get Expat to parse this? I definitely don't want
>>>>> to fix in-source comments to use different characters, and the source
>>>>> file in question should be correctly UTF-8 encoded. This does not seem
>>>>> to carry through doxygen though, and I'm not sure where it's going
>>>>> wrong. It seems that the XML is already incorrectly encoded. Any idea
>>>>> how I to work around this?
>>>>>
>>>>> Cheers,
>>>>>  -t
>>>>>
>>>>> Am 16. April 2012 20:03 schrieb Timothy J. Arrington 
<sse3ml@googlemail.com>:
>>>>>> Ok, so I managed to include breathe into my build process. I've
>>>>>> generated XML documentation using doxygen 1.8.0. What's happening now
>>>>>> is that the build fails. First of all, the documentation
>>>>>> (http://michaeljones.github.com/breathe/quickstart.html) is missing
>>>>>> spaces between the '..' and the command name. In order to work,
>>>>>> commands must be written like this:
>>>>>>
>>>>>> .. doxygenindex::
>>>>>>
>>>>>> Anyway, the build fails with some parse error (see the attachment.)
>>>>>> Any idea how I can debug this further? In my case, there's quite a
>>>>>> bunch of files with more than 1000 lines, and an initial scan didn't
>>>>>> bring up anything suspicious.
>>>>>>
>>>>>> Cheers,
>>>>>>  -t
>>>>>>
>>>>>> Am 15. April 2012 21:42 schrieb Timothy J. Arrington 
<sse3ml@googlemail.com>:
>>>>>>> Works now, thanks a lot.
>>>>>>>
>>>>>>> I'll try to get it working during the next week. I'm highly interested
>>>>>>> in combining a lot of doxygen documentation into the Sphinx-based
>>>>>>> documentation; my only concern right now is formatting and usability.
>>>>>>> I have spent quite some time tweaking the Doxygen-Documentation to
>>>>>>> look good, and I definitely like the fact that stuff can be found
>>>>>>> easily with the built-in doxygen search. I have also lots of content
>>>>>>> stored in doxygen pages, which I want to keep if possible.
>>>>>>>
>>>>>>> I'll let you know how I get on.
>>>>>>>
>>>>>>> Cheers,
>>>>>>>  -t
>>>>>>>
>>>>>>> Am 15. April 2012 21:28 schrieb Michael Jones <m.pricejones@gmail.com>:
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> Thanks for letting me know.  I hadn't realised it had gone wrong.
>>>>>>>> There seemed to be something wrong with the "dumb binary" entry.  I've
>>>>>>>> deleted that and tried to install and I think it picks up the "source"
>>>>>>>> entry instead and works.
>>>>>>>>
>>>>>>>> Please give it another go and see if you have any luck, otherwise I'll
>>>>>>>> investigate further.
>>>>>>>>
>>>>>>>> Thanks for trying breathe, or for trying to try it :)
>>>>>>>> Michael
>>>>>>>>
>>>>>>>>
>>>>>>>> On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
>>>>>>>> <sse3ml@googlemail.com> wrote:
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>> I just tried to install breathe using easy_install on cygwin, but it
>>>>>>>>> fails with:
>>>>>>>>>
>>>>>>>>> Searching for breathe
>>>>>>>>> Reading http://pypi.python.org/simple/breathe/
>>>>>>>>> Reading https://github.com/michaeljones/breathe
>>>>>>>>> Best match: breathe 0.7.5.linux-x86-64
>>>>>>>>> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
>>>>>>>>> Processing breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>>>> error: Couldn't find a setup script in
>>>>>>>>> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>>>>
>>>>>>>>> Is this a known problem? How can I workaround this issue?
>>>>>>>>>
>>>>>>>>> Cheers,
>>>>>>>>>  -t

Re: [breathe] Problem installing using easy_install

From:
Timothy J. Arrington
Date:
2012-05-23 @ 19:35
Hi Michael,

we've tried to hack on Breathe to get the features in we need, but
hacking on Breathe turned out to be really complicated even though it
is pretty well documented. The main problem we were facing is that the
xml structure is carried along too far, and we have a hard time to
"just add" new features without breaking the in-place structure. As
time is money, we're currently evaluating a custom solution from
doxygen->sphinx which is pretty minimal but extremely easy to hack if
necessary; if it turns out to work fine, we're going to open source it
as well and there might be parts from Breathe that we could use for
formatting and output. Right now, our main concern is to get something
working quickly.

Cheers,
  --t

2012/5/7 Michael Jones <m.pricejones@gmail.com>:
> Hi Tim,
>
> Sorry for the slow reply, somehow I didn't register your additional
> emails in this thread.
>
> Thanks for giving it a go and for providing such a comprehensive run
> down.  I can't answer all your questions right now but I can look a
> little closer if you're still interested.  Breathe is kind of in this
> half-way house of doing quite a bit of cool stuff but not quite being
> a full solution yet.  Development stumbles along as different people
> show interest and request fixes and features but I'm not actively
> developing it otherwise.
>
> From what I can see the single page vs multiple pages issue is the
> biggest. Breathe follows the Sphinx model of requiring you (as far as
> I know) to create stub files for all the pages you're interested in.
> I think the output of a restructured text directive is limited to a
> single page so you can't easily create multiple pages from say a
> single "doxygenindex" statement.  If the situation has changed with
> Sphinx then I'd be interested to know how it is done.  Until then I
> had plans to create a script that would help make those stub files but
> that is about as far as I had thought.
>
> Thanks again, if you're really interested in hacking on it a bit and
> extending it you'd be more than welcome.  I'm happy to contribute and
> help as much as I can too.
>
> All the best,
> Michael
>
>
> On Thu, Apr 19, 2012 at 6:15 AM, Timothy J. Arrington
> <sse3ml@googlemail.com> wrote:
>> I just did some more testing, here's the findings:
>>
>> * The output seems to contain everything that Doxygen produces as
>> well, so I'm fine with that. However, some classes get documented
>> which are not included in the Doxygen documentation, in my case,
>> private Impl classes basically produce noise. There doesn't seem a way
>> to remove them.
>> * No inheritance graph: I can live with that. It's rarely useful anyway.
>> * Enumeration formatting is a bit off: If there is no description, it
>> still produces value - -- the trailing hyphen is irritating. I could
>> hack something to resolve this issue relatively quickly by modifying
>> the renderer in renderer/.../compound.py, in the
>> EnumvalueTypeSubRenderer class. However, I'm not sure if this class is
>> generated or open for modification at all. There is definitely a lot
>> of pre-generated stuff, and I would be more happy to just "inject"
>> modifications instead of modifying existing classes.
>> * When using .. doxygenclass:: I have to provide the full class name
>> including the namespace, but for the documentation, I would love to
>> strip the common namespace prefix. That should be easy to do if I can
>> hook the title output, but I couldn't find where exactly this is
>> happening.
>> * Function parameters get reformatted somewhere; I would like to
>> manipulate it so I can change const Id & id to <bold>const</bold> Id&
>> id for instance. Same for various instances of typedefs, where typedef
>> is not highlighted.
>> * The output is packed in <div> instances for Sphinx 1.1.3; I wonder
>> if this can be augmented a bit so I can figure out which parts have
>> been generated by breathe, so I can apply a different formatting. For
>> enums, I have seen that there is a breatheenumvalues, but there is
>> certainly a lack of CSS selectors for formatting everything.
>> * The printed documentation has a lot of empty space.
>> * No easy way to get a list of all classes, functions, etc.
>> * Groups seem to be ignored? I have multiple pages nested in groups,
>> but I couldn't find a way how to extract those. The most interesting
>> feature here is to have a list of all members in each group, and nest
>> the groups properly, so I can retain a hierarchical index (this works
>> painlessly in doxygen.)
>>
>> However, that's it actually, otherwise, Breathe seems like the perfect
>> solution for projects which use Sphinx for documentation. I like the
>> output and being indepedent of Doxygen's HTML problems is definitely a
>> huge win. I especially like the fact that the printed documentation
>> would work just fine with Breathe+Sphinx+latexpdf.
>>
>> It would great if you could provide some pointers where to start
>> extending, if there are any plans to make the output easier to
>> customize, or if you are already planning to integrate some of the
>> stuff I mentioned above. I definitely would like to drop Doxygen's
>> HTML alltogether in favour of Sphinx+Breathe, but I'm not sure how
>> much time I'll be able to invest into adapting Breathe's output (but
>> I'm really interested in fixing the issues above, which would bring
>> the Breathe output more or less in sync with what Qt has.)
>>
>> Cheers,
>>  -t
>>
>> Am 18. April 2012 18:54 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>> Hi,
>>>
>>> with lots of more wrestling, I finally made sure that each and every
>>> file is properly UTF-8 encoded. It seems that doxygen passes through
>>> whatever comes in, and if it is not UTF8, it gets mangled in the
>>> process.
>>>
>>> Ok, so I've built now my documentation using breathe. It worked fine
>>> for my several-hundred-classes project, albeit it took some time :)
>>> The documentation looks pretty decent, but -- it's all on one page,
>>> and the layout is not quite how I would like to it be. For instance, I
>>> would like to have a short list of all functions/members for each
>>> class serving as an index to the longer function descriptions. There's
>>> also very little CSS markup I can use for styling. And finally, how
>>> the heck can I put each class into its own file?
>>>
>>> My biggest problems right now:
>>> * Each class into one file: How can I change this, or where do I have
>>> to start extending Breathe to fix this?
>>> * Each page into one page: Same as above.
>>> * Where to extend breathe to start tweaking the output?
>>> * The search ain't working, if I search for a class using the
>>> Sphinx-built-in-search, it doesn't find anything
>>>
>>> Besides this, getting Breathe to work at all is damn cool, and the
>>> output looks like it could be much better than Doxygen with less
>>> tweaking.
>>>
>>> Cheers,
>>>  --t
>>>
>>> Am 18. April 2012 17:58 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>>> Hi Michael,
>>>>
>>>> thanks. I'll try to get a bug report to the Doxygen guys as well if I
>>>> can make a small repro case. Actually, the character is not even
>>>> unicode per se. It's already in the ANSI set (ANSI Character 176 or
>>>> 0xB0) but this is no longer in the lower 7-bit UTF-8 plane. My guess
>>>> is the character gets passed through as-is instead of being encoded
>>>> properly as UTF-8. I'll try to figure it out, thanks for trying on
>>>> your end as well.
>>>>
>>>> Cheers,
>>>>  -t
>>>>
>>>> Am 18. April 2012 09:10 schrieb Michael Jones <m.pricejones@gmail.com>:
>>>>> Hi,
>>>>>
>>>>> I'm afraid I'm not sure myself.  I'd agree that with everything in the
>>>>> correct encoding you would hope that it would work.  I'll try to take
>>>>> a look and maybe raise it with the doxygen guys if I can (though I
>>>>> know them no better than anyone else.)
>>>>>
>>>>> Thanks for debugging it passed the original error and for pointing out
>>>>> the issue in the quick start guide, I'll fix that.  Also much respect
>>>>> for using unicode characters in source code comments... I'm still
>>>>> stuck in the past in seems.
>>>>>
>>>>> Cheers,
>>>>> Michael
>>>>>
>>>>>
>>>>> On Tue, Apr 17, 2012 at 6:15 AM, Timothy J. Arrington
>>>>> <sse3ml@googlemail.com> wrote:
>>>>>> Ah, some debugging revealed it chocked at this line:
>>>>>>
>>>>>> <para>This works for arbitary matrices in
>>>>>> O(n³).</para><para><simplesect kind="return"><para>False if a 0-pivot
>>>>>> was encountered. </para></simplesect>
>>>>>>
>>>>>> After fixing this, it still breaks on another "special" character, the
>>>>>> degree symbol (°)
>>>>>>
>>>>>> <codeline lineno="148"><highlight
>>>>>> 
class="normal"><sp/><sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight
>>>>>> 
class="comment">//<sp/>right<sp/>we<sp/>have<sp/>to<sp/>rotate<sp/>by<sp/>30°</highlight><highlight
>>>>>> class="normal"></highlight></codeline>
>>>>>>
>>>>>> Is there any way to get Expat to parse this? I definitely don't want
>>>>>> to fix in-source comments to use different characters, and the source
>>>>>> file in question should be correctly UTF-8 encoded. This does not seem
>>>>>> to carry through doxygen though, and I'm not sure where it's going
>>>>>> wrong. It seems that the XML is already incorrectly encoded. Any idea
>>>>>> how I to work around this?
>>>>>>
>>>>>> Cheers,
>>>>>>  -t
>>>>>>
>>>>>> Am 16. April 2012 20:03 schrieb Timothy J. Arrington 
<sse3ml@googlemail.com>:
>>>>>>> Ok, so I managed to include breathe into my build process. I've
>>>>>>> generated XML documentation using doxygen 1.8.0. What's happening now
>>>>>>> is that the build fails. First of all, the documentation
>>>>>>> (http://michaeljones.github.com/breathe/quickstart.html) is missing
>>>>>>> spaces between the '..' and the command name. In order to work,
>>>>>>> commands must be written like this:
>>>>>>>
>>>>>>> .. doxygenindex::
>>>>>>>
>>>>>>> Anyway, the build fails with some parse error (see the attachment.)
>>>>>>> Any idea how I can debug this further? In my case, there's quite a
>>>>>>> bunch of files with more than 1000 lines, and an initial scan didn't
>>>>>>> bring up anything suspicious.
>>>>>>>
>>>>>>> Cheers,
>>>>>>>  -t
>>>>>>>
>>>>>>> Am 15. April 2012 21:42 schrieb Timothy J. Arrington 
<sse3ml@googlemail.com>:
>>>>>>>> Works now, thanks a lot.
>>>>>>>>
>>>>>>>> I'll try to get it working during the next week. I'm highly interested
>>>>>>>> in combining a lot of doxygen documentation into the Sphinx-based
>>>>>>>> documentation; my only concern right now is formatting and usability.
>>>>>>>> I have spent quite some time tweaking the Doxygen-Documentation to
>>>>>>>> look good, and I definitely like the fact that stuff can be found
>>>>>>>> easily with the built-in doxygen search. I have also lots of content
>>>>>>>> stored in doxygen pages, which I want to keep if possible.
>>>>>>>>
>>>>>>>> I'll let you know how I get on.
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>>  -t
>>>>>>>>
>>>>>>>> Am 15. April 2012 21:28 schrieb Michael Jones <m.pricejones@gmail.com>:
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>> Thanks for letting me know.  I hadn't realised it had gone wrong.
>>>>>>>>> There seemed to be something wrong with the "dumb binary" entry.  I've
>>>>>>>>> deleted that and tried to install and I think it picks up the "source"
>>>>>>>>> entry instead and works.
>>>>>>>>>
>>>>>>>>> Please give it another go and see if you have any luck, otherwise I'll
>>>>>>>>> investigate further.
>>>>>>>>>
>>>>>>>>> Thanks for trying breathe, or for trying to try it :)
>>>>>>>>> Michael
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
>>>>>>>>> <sse3ml@googlemail.com> wrote:
>>>>>>>>>> Hi,
>>>>>>>>>>
>>>>>>>>>> I just tried to install breathe using easy_install on cygwin, but it
>>>>>>>>>> fails with:
>>>>>>>>>>
>>>>>>>>>> Searching for breathe
>>>>>>>>>> Reading http://pypi.python.org/simple/breathe/
>>>>>>>>>> Reading https://github.com/michaeljones/breathe
>>>>>>>>>> Best match: breathe 0.7.5.linux-x86-64
>>>>>>>>>> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
>>>>>>>>>> Processing breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>>>>> error: Couldn't find a setup script in
>>>>>>>>>> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>>>>>
>>>>>>>>>> Is this a known problem? How can I workaround this issue?
>>>>>>>>>>
>>>>>>>>>> Cheers,
>>>>>>>>>>  -t

Re: [breathe] Problem installing using easy_install

From:
Michael Jones
Date:
2012-05-29 @ 20:12
Hi Tim,

Thanks for keeping me updated.  Breathe certainly does keep the xml
structure around quite a bit.  I can't say it feels like a good thing
but I felt it was a bigger task to design an alternative storage
structure for all the information doxygen can throw at you. But then I
was trying to stay general and trying not to do too much work :), if
you want something to suit your purposes I'm sure it is pretty
reasonable and as you say far easier to work with.

Interested to see what you come up with.  Please open source it, it'd
be great to see some more options for this.

Cheers,
Michael

On Thu, May 24, 2012 at 7:35 AM, Timothy J. Arrington
<sse3ml@googlemail.com> wrote:
> Hi Michael,
>
> we've tried to hack on Breathe to get the features in we need, but
> hacking on Breathe turned out to be really complicated even though it
> is pretty well documented. The main problem we were facing is that the
> xml structure is carried along too far, and we have a hard time to
> "just add" new features without breaking the in-place structure. As
> time is money, we're currently evaluating a custom solution from
> doxygen->sphinx which is pretty minimal but extremely easy to hack if
> necessary; if it turns out to work fine, we're going to open source it
> as well and there might be parts from Breathe that we could use for
> formatting and output. Right now, our main concern is to get something
> working quickly.
>
> Cheers,
>  --t
>
> 2012/5/7 Michael Jones <m.pricejones@gmail.com>:
>> Hi Tim,
>>
>> Sorry for the slow reply, somehow I didn't register your additional
>> emails in this thread.
>>
>> Thanks for giving it a go and for providing such a comprehensive run
>> down.  I can't answer all your questions right now but I can look a
>> little closer if you're still interested.  Breathe is kind of in this
>> half-way house of doing quite a bit of cool stuff but not quite being
>> a full solution yet.  Development stumbles along as different people
>> show interest and request fixes and features but I'm not actively
>> developing it otherwise.
>>
>> From what I can see the single page vs multiple pages issue is the
>> biggest. Breathe follows the Sphinx model of requiring you (as far as
>> I know) to create stub files for all the pages you're interested in.
>> I think the output of a restructured text directive is limited to a
>> single page so you can't easily create multiple pages from say a
>> single "doxygenindex" statement.  If the situation has changed with
>> Sphinx then I'd be interested to know how it is done.  Until then I
>> had plans to create a script that would help make those stub files but
>> that is about as far as I had thought.
>>
>> Thanks again, if you're really interested in hacking on it a bit and
>> extending it you'd be more than welcome.  I'm happy to contribute and
>> help as much as I can too.
>>
>> All the best,
>> Michael
>>
>>
>> On Thu, Apr 19, 2012 at 6:15 AM, Timothy J. Arrington
>> <sse3ml@googlemail.com> wrote:
>>> I just did some more testing, here's the findings:
>>>
>>> * The output seems to contain everything that Doxygen produces as
>>> well, so I'm fine with that. However, some classes get documented
>>> which are not included in the Doxygen documentation, in my case,
>>> private Impl classes basically produce noise. There doesn't seem a way
>>> to remove them.
>>> * No inheritance graph: I can live with that. It's rarely useful anyway.
>>> * Enumeration formatting is a bit off: If there is no description, it
>>> still produces value - -- the trailing hyphen is irritating. I could
>>> hack something to resolve this issue relatively quickly by modifying
>>> the renderer in renderer/.../compound.py, in the
>>> EnumvalueTypeSubRenderer class. However, I'm not sure if this class is
>>> generated or open for modification at all. There is definitely a lot
>>> of pre-generated stuff, and I would be more happy to just "inject"
>>> modifications instead of modifying existing classes.
>>> * When using .. doxygenclass:: I have to provide the full class name
>>> including the namespace, but for the documentation, I would love to
>>> strip the common namespace prefix. That should be easy to do if I can
>>> hook the title output, but I couldn't find where exactly this is
>>> happening.
>>> * Function parameters get reformatted somewhere; I would like to
>>> manipulate it so I can change const Id & id to <bold>const</bold> Id&
>>> id for instance. Same for various instances of typedefs, where typedef
>>> is not highlighted.
>>> * The output is packed in <div> instances for Sphinx 1.1.3; I wonder
>>> if this can be augmented a bit so I can figure out which parts have
>>> been generated by breathe, so I can apply a different formatting. For
>>> enums, I have seen that there is a breatheenumvalues, but there is
>>> certainly a lack of CSS selectors for formatting everything.
>>> * The printed documentation has a lot of empty space.
>>> * No easy way to get a list of all classes, functions, etc.
>>> * Groups seem to be ignored? I have multiple pages nested in groups,
>>> but I couldn't find a way how to extract those. The most interesting
>>> feature here is to have a list of all members in each group, and nest
>>> the groups properly, so I can retain a hierarchical index (this works
>>> painlessly in doxygen.)
>>>
>>> However, that's it actually, otherwise, Breathe seems like the perfect
>>> solution for projects which use Sphinx for documentation. I like the
>>> output and being indepedent of Doxygen's HTML problems is definitely a
>>> huge win. I especially like the fact that the printed documentation
>>> would work just fine with Breathe+Sphinx+latexpdf.
>>>
>>> It would great if you could provide some pointers where to start
>>> extending, if there are any plans to make the output easier to
>>> customize, or if you are already planning to integrate some of the
>>> stuff I mentioned above. I definitely would like to drop Doxygen's
>>> HTML alltogether in favour of Sphinx+Breathe, but I'm not sure how
>>> much time I'll be able to invest into adapting Breathe's output (but
>>> I'm really interested in fixing the issues above, which would bring
>>> the Breathe output more or less in sync with what Qt has.)
>>>
>>> Cheers,
>>>  -t
>>>
>>> Am 18. April 2012 18:54 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>>> Hi,
>>>>
>>>> with lots of more wrestling, I finally made sure that each and every
>>>> file is properly UTF-8 encoded. It seems that doxygen passes through
>>>> whatever comes in, and if it is not UTF8, it gets mangled in the
>>>> process.
>>>>
>>>> Ok, so I've built now my documentation using breathe. It worked fine
>>>> for my several-hundred-classes project, albeit it took some time :)
>>>> The documentation looks pretty decent, but -- it's all on one page,
>>>> and the layout is not quite how I would like to it be. For instance, I
>>>> would like to have a short list of all functions/members for each
>>>> class serving as an index to the longer function descriptions. There's
>>>> also very little CSS markup I can use for styling. And finally, how
>>>> the heck can I put each class into its own file?
>>>>
>>>> My biggest problems right now:
>>>> * Each class into one file: How can I change this, or where do I have
>>>> to start extending Breathe to fix this?
>>>> * Each page into one page: Same as above.
>>>> * Where to extend breathe to start tweaking the output?
>>>> * The search ain't working, if I search for a class using the
>>>> Sphinx-built-in-search, it doesn't find anything
>>>>
>>>> Besides this, getting Breathe to work at all is damn cool, and the
>>>> output looks like it could be much better than Doxygen with less
>>>> tweaking.
>>>>
>>>> Cheers,
>>>>  --t
>>>>
>>>> Am 18. April 2012 17:58 schrieb Timothy J. Arrington <sse3ml@googlemail.com>:
>>>>> Hi Michael,
>>>>>
>>>>> thanks. I'll try to get a bug report to the Doxygen guys as well if I
>>>>> can make a small repro case. Actually, the character is not even
>>>>> unicode per se. It's already in the ANSI set (ANSI Character 176 or
>>>>> 0xB0) but this is no longer in the lower 7-bit UTF-8 plane. My guess
>>>>> is the character gets passed through as-is instead of being encoded
>>>>> properly as UTF-8. I'll try to figure it out, thanks for trying on
>>>>> your end as well.
>>>>>
>>>>> Cheers,
>>>>>  -t
>>>>>
>>>>> Am 18. April 2012 09:10 schrieb Michael Jones <m.pricejones@gmail.com>:
>>>>>> Hi,
>>>>>>
>>>>>> I'm afraid I'm not sure myself.  I'd agree that with everything in the
>>>>>> correct encoding you would hope that it would work.  I'll try to take
>>>>>> a look and maybe raise it with the doxygen guys if I can (though I
>>>>>> know them no better than anyone else.)
>>>>>>
>>>>>> Thanks for debugging it passed the original error and for pointing out
>>>>>> the issue in the quick start guide, I'll fix that.  Also much respect
>>>>>> for using unicode characters in source code comments... I'm still
>>>>>> stuck in the past in seems.
>>>>>>
>>>>>> Cheers,
>>>>>> Michael
>>>>>>
>>>>>>
>>>>>> On Tue, Apr 17, 2012 at 6:15 AM, Timothy J. Arrington
>>>>>> <sse3ml@googlemail.com> wrote:
>>>>>>> Ah, some debugging revealed it chocked at this line:
>>>>>>>
>>>>>>> <para>This works for arbitary matrices in
>>>>>>> O(n³).</para><para><simplesect kind="return"><para>False if a 0-pivot
>>>>>>> was encountered. </para></simplesect>
>>>>>>>
>>>>>>> After fixing this, it still breaks on another "special" character, the
>>>>>>> degree symbol (°)
>>>>>>>
>>>>>>> <codeline lineno="148"><highlight
>>>>>>> 
class="normal"><sp/><sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight
>>>>>>> 
class="comment">//<sp/>right<sp/>we<sp/>have<sp/>to<sp/>rotate<sp/>by<sp/>30°</highlight><highlight
>>>>>>> class="normal"></highlight></codeline>
>>>>>>>
>>>>>>> Is there any way to get Expat to parse this? I definitely don't want
>>>>>>> to fix in-source comments to use different characters, and the source
>>>>>>> file in question should be correctly UTF-8 encoded. This does not seem
>>>>>>> to carry through doxygen though, and I'm not sure where it's going
>>>>>>> wrong. It seems that the XML is already incorrectly encoded. Any idea
>>>>>>> how I to work around this?
>>>>>>>
>>>>>>> Cheers,
>>>>>>>  -t
>>>>>>>
>>>>>>> Am 16. April 2012 20:03 schrieb Timothy J. Arrington 
<sse3ml@googlemail.com>:
>>>>>>>> Ok, so I managed to include breathe into my build process. I've
>>>>>>>> generated XML documentation using doxygen 1.8.0. What's happening now
>>>>>>>> is that the build fails. First of all, the documentation
>>>>>>>> (http://michaeljones.github.com/breathe/quickstart.html) is missing
>>>>>>>> spaces between the '..' and the command name. In order to work,
>>>>>>>> commands must be written like this:
>>>>>>>>
>>>>>>>> .. doxygenindex::
>>>>>>>>
>>>>>>>> Anyway, the build fails with some parse error (see the attachment.)
>>>>>>>> Any idea how I can debug this further? In my case, there's quite a
>>>>>>>> bunch of files with more than 1000 lines, and an initial scan didn't
>>>>>>>> bring up anything suspicious.
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>>  -t
>>>>>>>>
>>>>>>>> Am 15. April 2012 21:42 schrieb Timothy J. Arrington 
<sse3ml@googlemail.com>:
>>>>>>>>> Works now, thanks a lot.
>>>>>>>>>
>>>>>>>>> I'll try to get it working during the next week. I'm highly interested
>>>>>>>>> in combining a lot of doxygen documentation into the Sphinx-based
>>>>>>>>> documentation; my only concern right now is formatting and usability.
>>>>>>>>> I have spent quite some time tweaking the Doxygen-Documentation to
>>>>>>>>> look good, and I definitely like the fact that stuff can be found
>>>>>>>>> easily with the built-in doxygen search. I have also lots of content
>>>>>>>>> stored in doxygen pages, which I want to keep if possible.
>>>>>>>>>
>>>>>>>>> I'll let you know how I get on.
>>>>>>>>>
>>>>>>>>> Cheers,
>>>>>>>>>  -t
>>>>>>>>>
>>>>>>>>> Am 15. April 2012 21:28 schrieb Michael Jones <m.pricejones@gmail.com>:
>>>>>>>>>> Hi,
>>>>>>>>>>
>>>>>>>>>> Thanks for letting me know.  I hadn't realised it had gone wrong.
>>>>>>>>>> There seemed to be something wrong with the "dumb binary" entry.  I've
>>>>>>>>>> deleted that and tried to install and I think it picks up the "source"
>>>>>>>>>> entry instead and works.
>>>>>>>>>>
>>>>>>>>>> Please give it another go and see if you have any luck, otherwise I'll
>>>>>>>>>> investigate further.
>>>>>>>>>>
>>>>>>>>>> Thanks for trying breathe, or for trying to try it :)
>>>>>>>>>> Michael
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On Mon, Apr 16, 2012 at 6:52 AM, Timothy J. Arrington
>>>>>>>>>> <sse3ml@googlemail.com> wrote:
>>>>>>>>>>> Hi,
>>>>>>>>>>>
>>>>>>>>>>> I just tried to install breathe using easy_install on cygwin, but it
>>>>>>>>>>> fails with:
>>>>>>>>>>>
>>>>>>>>>>> Searching for breathe
>>>>>>>>>>> Reading http://pypi.python.org/simple/breathe/
>>>>>>>>>>> Reading https://github.com/michaeljones/breathe
>>>>>>>>>>> Best match: breathe 0.7.5.linux-x86-64
>>>>>>>>>>> Downloading 
http://pypi.python.org/packages/any/b/breathe/breathe-0.7.5.linux-x86_64.tar.gz#md5=2c4fed09acedf3364d3ccb9627a4be55
>>>>>>>>>>> Processing breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>>>>>> error: Couldn't find a setup script in
>>>>>>>>>>> /tmp/easy_install-yaQcRH/breathe-0.7.5.linux-x86_64.tar.gz
>>>>>>>>>>>
>>>>>>>>>>> Is this a known problem? How can I workaround this issue?
>>>>>>>>>>>
>>>>>>>>>>> Cheers,
>>>>>>>>>>>  -t