| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
xsltproc is picky. We already do this for Mallard, but missed it for DocBook.
https://gitlab.gnome.org/GNOME/yelp-tools/-/issues/22
|
|
|
|
| |
https://gitlab.gnome.org/GNOME/yelp-tools/-/merge_requests/5
|
|
|
|
|
|
|
| |
I don't think any distro has python3 installed in /bin.
Some distros have completed UsrMerge, which makes /bin a symlink to
/usr/bin, so /bin/python3 is available on those distros. But it breaks
on all other distros.
|
|
|
|
|
|
| |
This is mostly useful in config files, where you have specific pages that
you might want exempted from some custom checker, but you don't want to
force people to explicitly list which filenames to test.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This commit allows you to specify your own checkers in .yelp-tools.cfg.
This concept of how these work is very similar to Schematron, but very
simplified in a basic config file syntax. For example, here's how to
write a check to make sure you always have a non-empty desc element:
[namespaces]
mal = http://projectmallard.org/1.0/
[check:gnome-desc]
select = /mal:page/mal:info
assert = normalize-space(mal:desc) != ''
message = Must have non-empty desc
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
You can now add a .yelp-tools.cfg file to provide default options.
All the command-line args now have long forms, and you can use that
long form in the cfg file under a section called [check:*tool*],
where *tool* is the specific subcommand. For example, let's say
you always want strict validation, but to allow the ITS namespace:
[check:validate]
strict = true
allow = http://www.w3.org/2005/11/its
You can also use the [check] secton for options that apply to all
subcommands. This is most useful for Mallard sites:
[check]
site = true
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| |
| | |
Previous fix had a mistake. I added a check in two places, did some copy pasta,
and failed to change a variable name.
|
| |
| |
| |
| |
| |
| |
| |
| | |
At some point, docbook.org starting doing 301 redirects to send all
requests thru https. Unfortunately, neither jing nor xmllint likes
these redirects, and xmllint doesn't do https at all. So, we have
to do some special casing depending on the validator and on whether
we have local copies in the catalogs. Fun stuff.
|
| |
| |
| |
| |
| | |
Bare sh doesn't have nullglob, so we have to check the existence of
filenames we got from iterating over a glob.
|
| | |
|
|/
|
|
|
|
|
|
|
|
| |
The meson is a build system focused on speed an ease of use,
which helps speeding up the software development.
https://wiki.gnome.org/Initiatives/GnomeGoals/MesonPorting
https://gitlab.gnome.org/GNOME/yelp-tools/issues/17
https://bugzilla.gnome.org/show_bug.cgi?id=794279
|
|
|
|
|
| |
https://bugzilla.gnome.org/show_bug.cgi?id=772334
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=839549
|
| |
|
|
|
|
|
|
| |
libxml2 doesn't always give the best error messages for validation failures.
With this option, you can choose to use jing instead of xmllint, which might
give you better error messages in some cases.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
Getting return values for commands in a pipeline is hard. I did it
wrong. I'm giving up for now for license, since error exits aren't
normal for the license subcommand.
|
| |
|
|
|
|
|
|
| |
Probably won't gain Sites support for building HTML. Use a site tool
like Pintail for that. But being able to generate site-aware cache
means more of the yelp-check commands can be site-aware.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Both `expr index` and `expr substr` are non-portable. What's more,
my usage of `expr index` was wrong, and there was a bug that nobody
ever seemed to trip up. Fix it and use portable stuff.
Bug #743911 partially addressed.
|
|
|
|
|
|
|
| |
When I changed the way yelp-build finds media files for Mallard
documents, I forgot to change a variable name in one place. So
the -i option to ignore missing media files actually causes it
to ignore all media files.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* This patch fixes bug #701959 which describes
and issue that prevents make install from being
successfully run if make has not been run first.
* The problem is that make install requires some
generated files to be present that are only generated
by make. The solution was to add $(_HELP_LC_FILES) as
a dependency to install-help so that these
files are always generated when make install is run.
|
|
|
|
|
|
|
|
|
|
|
| |
* This patch fixes bug #738142 which describes
and issue that prevents make distcheck from being
successfully run if make has not been run first.
* The problem is that make distcheck requires some
generated files to be present that are only generated
by make. The solution was to add $(_HELP_LC_FILES) as
a dependency to distdir-help-files so that these
files are always generated when make distcheck is run.
|
| |
|