Patrick Ohly wrote:
Hello!

I've done a full pass over the docs and merged them with the Connection
API.

Some comments about formatting: I suggest to use () after method names
in comments. Not sure what the convention here is elsewhere, both was
used in our docs.

Without word wrapping and indention the .xml files are hard to read. My
expectation is that extra white space is ignored by formatters, so I'd
prefer to reflow long paragraphs (in Emacs: TAB + ALT-Q). When updating
text, limit the amount of reformatting to keep diffs small.

The Docbook syntax might be wrong in some places. We'll sort this out
later.

This is quite possible, I'm not an expert on this.

About error reporting: we haven't specified any particular error
response when the API is used incorrectly or runs into errors, apart
from the sync status codes. I think that is sufficient, because other
errors are either internal or rare (like permission issues when trying
to write config files). Currently the implementation raises a
std::runtime_error exception with a string that explains the error, then
this is translated into a org.syncevolution.Exception error with that
string as detailed explanation.

Previously I used these dbus errors
"org.Moblin.SyncEvolution.GenericError"
"org.Moblin.SyncEvolution.NoSuchServer" -specified config does not exist
"org.Moblin.SyncEvolution.MissingArgs" -should probably be InvalidArgs
"org.Moblin.SyncEvolution.InvalidCall" -e.g. calling Abort when not syncing

Of those I actually check for NoSuchServer and InvalidCall in the client.

      * merge with Connection API

Done.

Do we want to maintain one XML spec file per interface or one spec file
for all of them? I have no idea what is easier to handle in libdbus-glib
or for merging docs later on.

Jussi?

It was useful for the server binding/glue generation, but we're not doing them anymore and client bindings do not actually benefit from the separation... So one file is fine for me.

 - Jussi
_______________________________________________
SyncEvolution mailing list
[email protected]
http://lists.syncevolution.org/listinfo/syncevolution

Reply via email to