Hello Jens, Something is wrong with the automatic update of http://docs.zope.org/zopetoolkit/ Recent changes don't seem to show up. -- Best regards, Adam GROSZER mailto:agroszer@gmail.com -- Quote of the day: For happiness one needs security, but joy can spring like a flower even from the cliffs of despair. - Anne Morrow Lindbergh
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 8/4/10 10:57 , Adam GROSZER wrote:
Hello Jens,
Something is wrong with the automatic update of http://docs.zope.org/zopetoolkit/
Recent changes don't seem to show up.
I rebuilt it now. There's a check in a shell script to see if the Subversion code is newer than the checked-out code, but it doesn't seem to be reliable. I'll have to watch it. If someone could code up something more reliable I'd be very happy. jens -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.8 (Darwin) iEYEARECAAYFAkxZLmYACgkQRAx5nvEhZLLwmgCeMIl0fZf4Qu91XQ4BtmyspoCk hFUAn3CjfLXtbtJVjrQ8qcKxLMRkDOTU =O5Cl -----END PGP SIGNATURE-----
On Wed, Aug 04, 2010 at 11:09:58AM +0200, Jens Vagelpohl wrote:
On 8/4/10 10:57 , Adam GROSZER wrote:
Hello Jens,
Something is wrong with the automatic update of http://docs.zope.org/zopetoolkit/
Recent changes don't seem to show up.
I rebuilt it now. There's a check in a shell script to see if the Subversion code is newer than the checked-out code, but it doesn't seem to be reliable. I'll have to watch it.
If someone could code up something more reliable I'd be very happy.
Can we see the shell script somewhere? Marius Gedminas -- http://pov.lt/ -- Zope 3/BlueBream consulting and development
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 8/4/10 14:55 , Marius Gedminas wrote:
On Wed, Aug 04, 2010 at 11:09:58AM +0200, Jens Vagelpohl wrote:
On 8/4/10 10:57 , Adam GROSZER wrote:
Hello Jens,
Something is wrong with the automatic update of http://docs.zope.org/zopetoolkit/
Recent changes don't seem to show up. I rebuilt it now. There's a check in a shell script to see if the Subversion code is newer than the checked-out code, but it doesn't seem to be reliable. I'll have to watch it.
If someone could code up something more reliable I'd be very happy.
Can we see the shell script somewhere?
The script (see below) sits at the root of the documentation buildout (such as the buildout at http://svn.zope.org/zopetoolkit/), which precreates the necessary sphinx-build script (that's why I am modifying $PATH). It's run every 15 minutes to detect and incorporate documentation changes in SVN. I'm not sure why it did stop firing the way it should. jens #!/bin/sh ROOT="/home/zope/zopetoolkit" export PATH="${PATH}:${ROOT}/bin" before=$(svn info $ROOT | grep "^Last Changed Rev:" | cut -d " " -f 4) svn up -q $ROOT after=$(svn info $ROOT | grep "^Last Changed Rev:" | cut -d " " -f 4) if [ "$before" != "$after" ]; then echo "Updated from revision $before to $after; rebuilding HTML docs." cd $ROOT/doc python bootstrap.py ${ROOT}/doc/bin/buildout -q -q make -s releaseinfo >/dev/null make -s html >/dev/null fi -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.8 (Darwin) iEYEARECAAYFAkxZeVMACgkQRAx5nvEhZLJJzwCcCHstIBdBOiPMch+MQw7AYokD 27IAn0lK9AD+O9mMNALTtbxsVjczyANL =V54Y -----END PGP SIGNATURE-----
On 8/4/10 16:29 , Jens Vagelpohl wrote:
The script (see below) sits at the root of the documentation buildout (such as the buildout at http://svn.zope.org/zopetoolkit/), which precreates the necessary sphinx-build script (that's why I am modifying $PATH). It's run every 15 minutes to detect and incorporate documentation changes in SVN. I'm not sure why it did stop firing the way it should.
I'm wondering: can it be made to skip useless docs such as http://docs.zope.org/zope.hookable/index.html, perhaps by checking for linecount? There are also a lot of packages that only have a changelog and no documentation, for example http://docs.zope.org/zope.contenttype/index.html . Should those be considered good enough to be linked? Wichert.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 8/4/10 17:26 , Wichert Akkerman wrote:
I'm wondering: can it be made to skip useless docs such as http://docs.zope.org/zope.hookable/index.html, perhaps by checking for linecount?
There are also a lot of packages that only have a changelog and no documentation, for example http://docs.zope.org/zope.contenttype/index.html . Should those be considered good enough to be linked?
IMHO trying to come up with some kind of programmatic algorithm to judge a bit of documentation as "good enough to show" is insane. I'm not going to attempt it. Maybe "shitty" (for some definition of "shitty", as someone else would say) package docs will compel the maintainers to improve them. jens -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.8 (Darwin) iEYEARECAAYFAkxZiWcACgkQRAx5nvEhZLLOTACfWhQ7OQGQW1noXiT1MbTATbhi SAwAoK/a6i+E2qbmfbNUxTalTNO78MoA =T1Kj -----END PGP SIGNATURE-----
On 8/4/10 17:38 , Jens Vagelpohl wrote:
IMHO trying to come up with some kind of programmatic algorithm to judge a bit of documentation as "good enough to show" is insane. I'm not going to attempt it.
Fair enough.
Maybe "shitty" (for some definition of "shitty", as someone else would say) package docs will compel the maintainers to improve them.
I doubt it, considering the exposure of "shitty" docs on pypi does not seem to have had any impact (aside from scaring away potential users). Wichert.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 8/4/10 17:41 , Wichert Akkerman wrote:
On 8/4/10 17:38 , Jens Vagelpohl wrote:
Maybe "shitty" (for some definition of "shitty", as someone else would say) package docs will compel the maintainers to improve them.
I doubt it, considering the exposure of "shitty" docs on pypi does not seem to have had any impact (aside from scaring away potential users).
You're probably right, sad as it is. jens -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.8 (Darwin) iEYEARECAAYFAkxZioYACgkQRAx5nvEhZLKTMACgmvcoBs5xV7sjkw4chXs0I54F F84AoIkBOjYT8jOzYkX9CpOHbWh446MI =T2Zd -----END PGP SIGNATURE-----
On Wed, Aug 04, 2010 at 05:41:08PM +0200, Wichert Akkerman wrote:
On 8/4/10 17:38 , Jens Vagelpohl wrote:
IMHO trying to come up with some kind of programmatic algorithm to judge a bit of documentation as "good enough to show" is insane. I'm not going to attempt it.
Fair enough.
Maybe "shitty" (for some definition of "shitty", as someone else would say) package docs will compel the maintainers to improve them.
I doubt it, considering the exposure of "shitty" docs on pypi does not seem to have had any impact (aside from scaring away potential users).
Crappy docs on PyPI make me want to run away (with one exception: ReST formatting errors on PyPI make me want to check out the source and fix them). Crappy docs on docs.zope.org make me want to check out the source package and commit fixes. I'm not prepared to speculate why I'm having these reactions -- is it a prettier CSS? The splitting of the docs into multiple reasonably-sized pages instead of one huge undigestable lump? I guess what I'm saying is: there's hope yet, don't give up! Marius Gedminas -- http://pov.lt/ -- Zope 3/BlueBream consulting and development
participants (4)
-
Adam GROSZER -
Jens Vagelpohl -
Marius Gedminas -
Wichert Akkerman