1724178 - remove the cpp LDAP implementation

Status: RESOLVED FIXED

(MailNews Core :: LDAP Integration, task)

Description

[Magnus Melin [:mkmelin]]

Bug 1696625 is fixed, and 91 has branched. We can soon remove the old c++ implementation. We may want to hold on a slight bit yet, to allow easier performance comparisons if needed. OTOH, those can be tried on the branch as well...

Comment 1

[Ping Chen [:rnons] (away)]

Attachment: Bug 1724178 - Make LDAPDirectory.jsm the only implementation of nsIAbLDAPDirectory. r=darktrojan

• Remove LDAPModuleLoader.jsm and mailnews.ldap.jsmodule
• Unregister C++ LDAP xpcoms

Comment 2

[Ping Chen [:rnons] (away)]

Attachment: Bug 1724178 - Remove unused ldap xpcoms implemented in C++. r=darktrojan

Depends on D125037.

Comment 3

[Geoff Lankow (:darktrojan)]

Hooray, deleting large amounts of ancient code!

There's a lot more code left than I was expecting. Which makes me wonder:

• Do we actually make requests to ldap:// URLs or are they just conceptual? I can't think of anywhere that would make a request to one.
• Does anybody use nsLDAPSyncQuery? I vaguely recall some people use our LDAP code for other things and maybe that's why it's #ifdefed, but I've no idea what those other things are or if it involves nsLDAPSyncQuery. Can't think why it would. Do you know, @mkmelin? Can we just destroy these bits?
• Can we get rid of the c-sdk directory, the README and TODO files and ldap-sdk.mozbuild?

Comment 4

[Geoff Lankow (:darktrojan)]

Oh, now I see bug 1729862!

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/comm-central/rev/230fd2220139
Make LDAPDirectory.jsm the only implementation of nsIAbLDAPDirectory. r=darktrojan
https://hg.mozilla.org/comm-central/rev/5bbf561a5577
Remove unused ldap xpcoms implemented in C++. r=darktrojan

Comment 6

[Geoff Lankow (:darktrojan)]

Are we done here?

Comment 7

[Ping Chen [:rnons] (away)]

I think one goal of this bug is to remove libldap, but it's blocked by bug 1729862. Of course we can close this one and do it in 1729862, which do you prefer.

Comment 8

[Geoff Lankow (:darktrojan)]

Let's close this one. If we do remove the rest it'll be because we did bug 1729862, so it might as well happen there. Also it prevents a big gap between patches with the same bug number.